Keywords: Neural interfaces - Microelectrode technology, Neural interfaces - Implantable systems, Neural interfaces - Tissue-electrode interface
Abstract: The exact localization of signal recording probes or deep stimulation probes by magnetic resonance imaging (MRI), has significant importance in studying and understanding how the brain functions. But the magnetic susceptibility of the probes itself distorts the MRI image and creates error to the position measurement. In this paper we propose an MRI compatible flexible probe with magnetic susceptibility that is well matched with the brain model. The well-matched magnetic susceptibility of the probe enables high resolution structural and functional MRI even at ultra-high B-field strengths. The MRI images shows almost zero artifacts around the implanted probe in the phantom tissue.

Keywords: Neural interfaces - Microelectrode technology
Abstract: Understanding the relationships between different cortical neurons under excitatory or inhibitory modulation is important for researches of many neurological disorders. In order to monitor the neural activities in cortex under gamma-aminobutyric acid (GABA) and glutamate (Glu) modulation, an implantable microelectrode array (MEA) was combined with microinjection capillary. The neurons in motor cortex of rat were modulated by GABA and Glu injection, and multichannel neural electrophysiological signals were simultaneously recorded. Spike analysis showed that the interneurons recorded by the MEA were inhibited after GABA injection and excited after Glu injection, but one pyramidal neuron was found to be not sensitive to the drug. The local field potentials (LFP) were most affected in the frequency band of 5~10 Hz after Glu injection, which greatly increased the amplitudes of the spindle-like waveforms. The MEA combined with microinjection provided a low-cost and effective tool for neurological drug modulation and evaluation in brain tissue.

Keywords: Neural interfaces - Microelectrode technology, Neural interfaces - Biomaterials, Neural interfaces - Tissue-electrode interface
Abstract: Recent neural interfaces are characterized by high functionality and good adaptation to the target tissue. Still, the underlying manufacturing process is mainly planar and so are the device and contact surface. Therefore, three-dimensional structures to contact neuronal tissue are desired to gain higher selectivity. In the present study, local bending structures integrated in flexible electrode arrays based on polyimide are investigated. The bending is achieved by the contraction of a second polyimide (Durimide) that is embedded into grooves with a width of a few micrometers. The angle of the bending can be controlled with a high accuracy from 3 to 20 degrees by changing the geometry of the grooves and the imidization temperature These bending structures can be combined to achieve any desired angle for specific applications.

Keywords: Neural interfaces - Microelectrode technology, Neural interfaces - Tissue-electrode interface, Neural interfaces - Implantable systems
Abstract: Chronic studies of flexible µECoG electrodes and the electrode-brain interface have been limited by the inability to assess tissue response over time. The electrophysiological system presented here combines epidural micro-electrocorticographic (µECoG) recording capabilities with the ability to visualize tissue response over time through light microscopy and optical coherence tomography (OCT). With the ability to interchange both the electrode and the electronics, and a flushing port for injection of flushing saline and/or drugs, this 3D printed system has future applications in chronic electrophysiology, optogenetics, and advanced imaging methods.

Keywords: Neural interfaces - Microelectrode technology, Neural interfaces - Biomaterials, Neural interfaces - Implantable systems
Abstract: An innovative fabrication process of glass waveguides on silicon substrates for miniaturized implants is presented. Thin glass was bonded on oxidized silicon wafers and patterned using wet etching. Multimode waveguides with different shapes and a low surface roughness as well as low scattering of light were successfully fabricated. For efficient coupling of light and accurate alignment, KOH-grooves were etched in the silicon with respect to the glass waveguides to attach optical fibers from external light sources. Towards higher biostability, several coating materials were evaluated in accelerated in vitro tests in 60°C PBS for the first time over a long period of time regarding their optical properties. TiO2, SiC, polyimide, Parylene C and SU-8 showed a very stable optical transmittance after 320 days in accelerated aging while PECVD Si3N4 showed significant changes within the first days.

Keywords: Neural interfaces - Microelectrode technology, Brain physiology and modeling - Cognition, memory, perception
Abstract: Obtaining multiple single-unit recordings in particular neural networks from behaving animals is crucial for the understanding of cognitive functions of the brain. Attaining stable, chronic recordings from the brain is also the foundation to develop effective cortical prosthetic devices. However, severe immune response caused by micromotion between stiff implants and surrounding brain tissue often limits the lifetime of penetrating, neural recording devices. To reduce the stiffness mismatch between recording devices and brain tissue, we developed a flexible, polymer based multi-electrode array for recording single neuron activities from the rat hippocampus, a major subcortical structure of the rat brain. Parylene C, a biocompatible polymer, was used as the structural and insulation material of the multi-electrode array. 64 platinum (Pt) recording electrodes were placed in groups along each shank to conform to the anatomical distribution of hippocampal principle neurons. The multi-electrode array was chronically implanted in three animals. After recovery, neural activity together with movement traces were collected from the behaving animals.

Keywords: Physiological systems modeling - Multivariate signal processing, Physiological systems modeling - Signal processing in physiological systems, Signal pattern classification
Abstract: In this paper, we propose a novel and intuitively pleasing graph-based spatio-temporal feature extraction framework for classification of motor imagery tasks from electroencephalography (EEG) signals for brain-computer interface systems (BCIs). In particular, to account for the observation that measurements obtained from the EEG channels form a non-uniformly distributed sensor field, a representation graph is constructed using geographical distances between sensors to form connectivity neighborhoods. By capitalizing on the fact that functionality of different connectivity neighborhoods varies based on the intensity of the performed activity and concentration level of the subject, we formed an initial functional clustering of EEG electrodes by designing a separate adjacency matrix for each identified functional cluster. Using a collapsing methodology based on total variation measures on graphs, the overall model will eventually be reduced (collapsed) into two functional clusters. The proposed framework offers two main superiorities over its state-of-the-art counterparts: (i) First, the resulting dimensionality reduction is subject-adaptive and respects the brain plasticity of subjects, and; (ii) Second, the proposed methodology identifies active regions of the brain during the motor imagery task, which can be used to re-align EEG electrodes to improve accuracy during consecutive data collection sessions. The experimental results based on Dataset IVa from BCI Competition III show that the proposed method can provide higher classification accuracy as compared to the other existing methods.

Keywords: Physiological systems modeling - Multivariate signal processing, Causality, Physiological systems modeling - Signal processing in physiological systems
Abstract: The analysis of short-term cardiovascular and cardiorespiratory regulation during altered conscious states, such as those induced by anesthesia, requires to employ time series analysis methods able to deal with the multivariate and multiscale nature of the observed dynamics. To meet this requirement, the present study exploits the extension to multiscale analysis of recently proposed information decomposition methods which allow to quantify, from short realizations, the amounts of joint, unique, redundant and synergistic information transferred within multivariate time series. These methods were applied to the spontaneous variability of heart period (HP), systolic arterial pressure (SAP) and respiration (RESP) in patients undergoing coronary artery bypass graft monitored before and after the induction of general anesthesia. We found that, after anesthesia induction, information is processed within the cardiovascular network in a scale-dependent way: at short time scales, a shift from synergistic to redundant information transferred from SAP and RESP to HP occurs, which is associated with enhanced baroreflex-mediated respiratory effects on arterial pressure; at longer time scales, the increased information transfer from SAP to HP denotes an enhancement of the baroreflex coupling related to slow cardiovascular oscillations.

Keywords: Physiological systems modeling - Multivariate signal processing, Coupling and synchronization - Coherence in biomedical signal processing, Time-frequency and time-scale analysis - Time-frequency analysis
Abstract: We propose a novel modelling framework to study non-stationary, directional brain-heart interplay in a timevarying fashion. Considering electroencephalographic (EEG) signals and Heart Rate Variability (HRV) series as inputs, a new multivariate formulation is derived from proper coupling functions linking cortical electrical activity and heartbeat dynamics generation models. These neural-autonomic coupling rules are formalised according to the current knowledge on the central autonomic network and fully parametrised in adaptive coefficients quantifying the information outflow from-brain-to-heart as well as from-heart-to-brain. Such coefficients can be effectively estimated by solving the model inverse problem, and profitably exploited for a novel assessment of brain-heart interactions. Here we show preliminary experimental results gathered from 27 healthy volunteers undergoing significant sympatho-vagal perturbations through cold-pressor test and discuss prospective uses of this novel methodological framework. Specifically, we highlight how the directional brain-heart coupling significantly increases during prolonged baroreflex elicitation with specific time delays and throughout specific brain areas, especially including fronto-parietal regions and lateralisation mechanisms in the temporal cortices.

Keywords: Physiological systems modeling - Multivariate signal processing, Physiological systems modeling - Signal processing in physiological systems, Time-frequency and time-scale analysis - Nonstationary processing
Abstract: In this work we are interested in analyzing any correlations between physiological parameters, extracted from signals such as Electrocardiogram, respiratory signal and Skin Conductance, and self-reported indices related to emotional or cognitive stimulations. For this purpose, an experiment involving twenty participants with a mean age of 25 ± 5 years of both sexes (13 males and 7 females) was carried out. The protocol included the navigation in simulated web-sites and the vision of two different commercial products (utilitarian and hedonistic). At the end of the navigation, a questionnaire was submitted to the subject in order to measure his/her feelings and emotions in a qualitative and subjective way. Quantitative features were extracted from the physiological signals recorded during the execution of the protocol. We performed a correlation analysis between self-reported and physiological responses related to Arousal, Pleasure, Expectancy and Situational Involvement. Findings showed that when a consumer is exposed to a utilitarian product, the physiological emotional responses are disassociated from the self-reported ones. For the hedonistic product, instead, self-reported measures significantly correlate with physiological arousal features like the combined effect of cardiac and respiratory activity and the Heart Rate.

Keywords: Physiological systems modeling - Signals and systems, Physiological systems modeling - Multivariate signal processing, Physiological systems modeling - Signal processing in physiological systems
Abstract: This study aims at investigating the possibility to employ neurophysiological measures to assess the humanmachine interaction effectiveness. Such a measure can be used to compare new technologies or solutions, with the final purpose to enhance operator’s experience and increase safety. In the present work, two different interaction modalities (Normal and Augmented) related to Air Traffic Management field have been compared, by involving 10 professional air traffic controllers in a control tower simulated environment. Experimental task consisted in locating aircrafts in different airspace positions by using the sense of hearing. In one modality (i.e. “Normal”), all the sound sources (aircrafts) had the same amplification factor. In the “Augmented” modality, the amplification factor of the sound sources located along the participant head sagittal axis was increased, while the intensity of sound sources located outside this axis decreased. In other words, when the user oriented his head toward the aircraft position, the related sound was amplified. Performance data, subjective questionnaires (i.e. NASA-TLX) and neurophysiological measures (i.e. EEG-based) related to the experienced workload have been collected. Results showed higher significant performance achieved by the users during the “Augmented” modality with respect to the “Normal” one, supported by a significant decreasing in experienced workload, evaluated by using EEG-based index. In addition, Performance and EEG-based workload index showed a significant negative correlation. On the contrary, subjective workload analysis did not show any significant trend. This result is a demonstration of the higher effectiveness of neurophysiological measures with respect to subjective ones for Human-Computer Interaction assessment.

Keywords: Physiological systems modeling - Multivariate signal processing, Physiological systems modeling - Signal processing in physiological systems, Signal pattern classification
Abstract: Physiological responses are essential for health monitoring. However, modeling the complex interactions between them across activity and environmental factors can be challenging. In this paper, we introduce a framework that identifies the state of an individual based on their activity, trains predictive models for their physiological response within these states, and jointly optimizes for the states and the models. We apply this framework to respiratory rate prediction based on heart rate and physical activity, and test it on a dataset of 9 individuals performing various activities of daily life.

Keywords: Brain imaging and image analysis, Image feature extraction, Image classification
Abstract: Alzheimer's disease (AD), a progressive brain disorder, is the most common neurodegenerative disease in older adults. There is a need for brain structural magnetic resonance imaging (MRI) biomarkers to help assess AD progression and intervention effects. Prior research showed that surface based brain imaging features hold great promise as efficient AD biomarkers. However, the complex geometry of cortical surfaces poses a major challenge to defining such a feature that is sensitive in qualification, robust in analysis, and intuitive in visualization. Here we propose a novel isometry invariant shape descriptor for brain morphometry analysis. First, we calculate a global area-preserving mapping from cortical surface to the unit sphere. Based on the mapping, the Beltrami coefficient shape descriptor is calculated. An analysis of average shape descriptors reveals that our detected features are consistent with some previous AD studies where medial temporal lobe volume was identified as an important AD imaging biomarker. We further apply a novel patch-based spherical sparse coding scheme for feature dimension reduction. Later, a support vector machine (SVM) classifier is applied to discriminate 135 amyloid-beta positive persons with the clinical diagnosis of Mild Cognitive Impairment (MCI) from 248 amyloid-beta-negative normal control subjects. The 5-folder cross-validation accuracy is about 81.82% on the dataset, outperforming some traditional, Freesurfer based, brain surface features. The results show that our shape descriptor is effective in distinguishing dementia due to AD from age-matched normal aging individuals. Our isometry invariant shape descriptors may provide a unique and intuitive way to inspect cortical surface and its morphometry changes.

Keywords: Image feature extraction, Image analysis and classification - Machine learning / Deep learning approaches, Image classification
Abstract: Attention-deficit/hyperactivity disorder (ADHD) is a childhood-onset neurodevelopmental disorder that often persists into adulthood, resulting in adverse effects on work performance and social function. The current diagnosis of ADHD primarily depends on the judgment of clinical symptoms, which highlights the need for objective imaging biomarkers. In this study, we aim to classify ADHD (both children and adults [34/112]) from age-matched healthy controls (HCs [28/77]) with functional connectivity (FCs) pattern derived from resting-state functional magnetic resonance imaging (rs-fMRI) data. However, the neuroimaging classification of brain disorders often meets a situation of high dimensional features were presented with limited sample size. Thus an efficient method that is able to reduce original feature dimension into a much more refined subspace is highly desired. Here we proposed a novel Feature Selection method based on Relative Importance and Ensemble learning (FS_RIEL). Compared with traditional feature selection methods, FS_RIEL algorithm improved the ADHD classification by about 15% in both child and adult ADHD classification, achieving 80-86% accuracy. Moreover, we found the most frequently selected FCs were mainly involved in frontoparietal network, default network, salience network, basal ganglia network and cerebellum network in both child and adult ADHD cohorts, which indicates that ADHD is characterized by a widely-impaired brain connectivity profile that may serve as potential biomarkers for its early diagnosis.

Keywords: Image feature extraction
Abstract: In this paper, we propose an end-to-end approach to addressing QRS complex detection and measurement of Electrocardiograph (ECG) paper using convolutional neural networks (CNNs). Unlike conventional detection solutions that convert images to digital data, our method can directly detect QRS complex in images using Faster-RCNN, then the R-peak can be located and measured through a CNN. Validated by clinical ECG data in the St.-Petersburg Institute of Cardiological Technics 12-lead Arrhythmia Database and real ECG paper from Peking University People's Hospital, the proposed method can achieve the recall of 98.32%, the precision of 99.01% in detecting and 0.012 mv of mean absolute error in measuring. Experimental results demonstrate the superior performance of our method over conventional solutions, which would pave the way to detect and measure ECG paper using CNNs.

Keywords: Image feature extraction, Image analysis and classification - Digital Pathology, Image classification
Abstract: Tumor localization, especially in case of minimally invasive lung tumor resection surgery, is extremely challenging due to the continuous motion of the organ. This motion can be troublesome as it results in spatial discrepancy corresponding to preoperative and intraoperative tumor location. In order to characterize lung tissue stiffness for the purpose of lung tumor localization, in this paper, we present a novel characterization approach based on variability in resistance of the healthy region vs. the tumorous region resulting from lung motion. The proposed approach is numerically validated on a Finite Element (FE) model of the lung with varying surface stiffnesses, where higher stiffness represents tumor and lower stiffness corresponds to healthy lung tissue. The numerical simulation validates the sensitivity of our mechanism for different grades of tumors by demonstrating that the strain on the healthy tissue is 31.8 and 67.1 times higher than that on the tumor surface for a selected relative stiffness variation of 3.6x and 24.4x respectively, at a pressure of 1.6 KPa. Additionally, a framework is developed to validate the proposed approach in a video of a video-assisted thoracoscopic surgery (VATS), where multiple landmarks on the lung surface are tracked. This enables us to quantify the motion of points residing on healthy surface and tumorous surface. The motion data is further analyzed to study the relative surface strain, and it is shown that the proposed approach differentiates a tumor from healthy surface.

Keywords: Image feature extraction, Brain imaging and image analysis, Optical imaging and microscopy - Microscopy
Abstract: Cellular phenotypic features derived from histopathology images are the basis of pathologic diagnosis and are thought to be related to underlying molecular profiles. Due to overwhelming cell numbers and population heterogeneity, it remains challenging to quantitatively compute and compare features of cells with distinct molecular signatures. In this study, we propose a self-reliant and efficient analysis framework that supports quantitative analysis of cellular phenotypic difference across distinct molecular groups. To demonstrate efficacy, we quantitatively analyze astrocytomas that are molecularly characterized as either Isocitrate Dehydrogenase (IDH) mutant (MUT) or wildtype (WT) using imaging data from The Cancer Genome Atlas database. Representative cell instances that are phenotypically different between these two groups are retrieved after segmentation, feature computation, data pruning, dimensionality reduction, and unsupervised clustering. Our analysis is generic and can be applied to a wide set of cell-based biomedical research.

Keywords: Implantable systems, Wearable power and on-body energy harvesting
Abstract: We propose a contactless energy harvesting system driven by the contraction of an electrically-stimulated skeletal muscle to be used to supply electrical energy to implantable medical devices. In order to realize a durable generator, the one proposed here has a contactless clutch mechanism with parallel leaf springs, with which the generator can be driven without friction. In this system, the muscle connected to the parallel leaf spring is intentionally contracted by electrical stimulation. The generator can be driven not only in the contraction phase of the muscle, but also relaxation phase. The result an evaluation showed that the prototype could generate 26.1 μW with an efficiency of 13.7%. Finally, we conducted an animal experiment using the gastrocnemius muscle of a toad with a weighing of 200 g. The generator was driven in the contraction phase generating 1.37 μW of power from the energy supplied by the muscle.

Keywords: Wearable low power, wireless sensing methods, Magnetic sensors and systems, Physiological monitoring - Instrumentation
Abstract: Body-worn battery-less Wireless Resistive Analog Passive (WRAP) sensor can be unobtrusive while collecting physiological data continuously. Inductive connection between a pair of Printed Spiral Coils (PSC) eliminates the intrusive wires. Inductive connection of primary and secondary PSC enabled us to probe the body signals using the inductive link. The primary side voltage is modulated by the sensed body signal at the secondary PSC. The coil physical characteristics influence the sensitivity which is defined as observed voltage changes over the sensor variation. We have previously reported an iterative method to optimize the coil specifications for maximum sensitivity with constrained coil profile size by maximizing the power transfer efficiency from primary to secondary. In this study sensitivity is maximized by first, driving an analytical multivariable equation of circuit components and physical characteristics, and then using Genetic Algorithm (GA) to maximize it with considering the size and fabrication constraints. The results are compared to the other methods that shows a higher result in the range of 102 comparing to the best alternate methods (sqp). It helps us to detect smaller physiological signals in the noisy environment.

Keywords: Implantable systems, Wearable power and on-body energy harvesting, Wearable low power, wireless sensing methods
Abstract: This paper presents the study of subcutaneous solar energy harvesting for implantable sensor systems. The characteristics of a flexible solar panel under a 3 mm thick porcine skin are measured under different ambient light conditions. The output power of the solar panel when covered by the skin varies from tens of micro Watts to a few milli Watts depending on the light source. A low-power implantable sensor prototype is proposed to evaluate the performance of the subcutaneous solar energy harvester. It consists of a power management circuit, a temperature sensor and a Bluetooth low energy (BLE) module. The average working current of the prototype is 400 muA (transient BLE transmission current is 8 mA), while its sleep current is only 7 muA. Experimental results show that the subcutaneous solar energy harvester illuminated by both sunlight and artificial light sources can power the implantable prototype.

Keywords: Implantable technologies, Implantable systems, Portable miniaturized systems
Abstract: Coagulation of blood inside the implanted medical device is quite a critical problem to limit the lifetime. In this paper, we propose a microfluidic blood separating device using curved and branched channels. It utilizes centrifugal force on curved flow and separates blood flow into blood cell rich and blood cell poor ones at the bifurcation. Though it cannot separate the plasma from blood cells completely, the blood with small concentrations of blood cells will have low coagulatibity and extend the lifetime of the implant medical device. The device does not require any external pumps or valves, i.e., the system does not need any power sources but the blood pressure. We conducted experiments with a titanium foil which contacted to human whole blood with different hematocrit values for 7 days. The device was experimentally characterized with respect to the channel design. The former experiments suggested that lower concentration of blood cells helps avoiding blood coagulations, and the latter showed that the separation by our device is mainly affected by the flow rate and channel curvature.

Keywords: New sensing techniques, Bio-electric sensors - Sensing methods, Bio-electric sensors - Sensor systems
Abstract: Human limb movement intent recognition fundamentally provides the control mechanism for assistive devices such as exoskeleton and limb prosthesis. While different biopotential signals have been utilized for limb movement intent decoding, they seldom could account for spatial information associated with changes in muscle shape that could also be used to characterize the limb motor intent. Therefore, this study developed a novel nano gold stretchable-flexible sensor that captures spatial information associated with the muscle shape change signal (MSCS) during different muscle activation patterns. The novel sensor consists of 2-channels to acquire MSCS at a sampling rate of 125 Hz, corresponding to multiple classes of upper limb movements acquired across six able-bodied subjects. By utilizing the linear discriminant analysis algorithm on the acquired data with a single extracted feature, an overall average motion decoding accuracy of 90.9% was achieved. In addition, the waveform analysis results show that the novel sensor’s recordings were less affected by external interferences, thus yielding high quality signals. This study is the first to utilize nano gold stretchable-flexible material for sensor fabrication in pattern recognition of upper limb movement intent, which may facilitate the development of effective assistive devices.

Keywords: Chemo/bio-sensing - Biological sensors and systems, Bio-electric sensors - Sensor systems, Bio-electric sensors - Sensing methods
Abstract: Carbon nanotube (CNT)-cellulose pellicle was developed to create a conductive CNT network on 20 μm nanostructured cellulose film. The flexible and electrically conductive film was prepared by the modification of bacterial nanocellulose pellicle with multi-walled carbon nanotubes (MWCNTs). The composite film was further modified with redox enzymes including pyroquinoline quinone glucose dehydrogenase (PQQ-GDH) and bilirubin oxidase (BODx) functioning as the anodic and cathodic catalyst, respectively with glucose as the biofuel source. The enzyme functionalized MWCNT-cellulose based glucose/O2 biofuel cell system harnessed the biochemical energy of glucose via the oxidation of glucose and reduction of molecular oxygen to generate electrical power in the microwatt range. The biofuel cell system exhibited an open circuit voltage and power density of 470 mV and 46.25 μW/cm2, respectively, with a current density of 381 μA/cm2 in the presence of 25 mM glucose. At physiological glucose concentration, the biofuel cell exhibited an open circuit voltage and power density of 418 mV and 24.975 μW/cm2 respectively, with a current density of 293.75 μA/cm2. As a result, we expect that this facile strategy to prepare flexible conductive bioelectrodes for the development of glucose biofuel cell system using synthesized bacterial nanocellulose crosslinked with MWCNTs and enzyme can be readily extended to diverse applications in enzymatic biofuel cell and biosensor technology.

Keywords: EEG imaging, Electrical source brain imaging, Brain imaging and image analysis
Abstract: Electroencephalography (EEG) source localization aims at reconstructing the current density on the brain cortex from scalp EEG recordings. It often starts with a generative model that maps brain activity to the EEG recording, and then solves the inverse problem. Previously proposed method graph fractional-order total variation (gFOTV) is based on spatial regularization, and was shown superior to some other existing spatial-regularized methods in simulation tests. However, the gFOTV addresses inverse problem for one time point at a time. The resultant estimated times series of brain activity is a simple concatenation of reconstructions independently performed at each time instance, and risks spurious temporal discontinuity due to overfitting noise in EEG recordings. In addition, the performance is subject to low signal-to-noise ratio (SNR) and small number of electrodes, which happens in realistic EEG recordings. To account for the generally continuous temporal variation in brain activity, but also allow for properly triggering abrupt changes, we propose a novel formulation that incorporates spatiotemporal regularization. Specifically, our method, called spatiotemporal graph total variation (STGTV) adopts graph fractional-order total variation (gFOTV) for spatial regularization and total variation (TV) for temporal regularization. The gFOTV encourages spatially smooth source distributions, and the temporal TV enhances temporal consistency in estimated activity maps. The introduction of implicit temporal coupling by temporal TV also helps with noise cancelation and enhances SNR. In a simulation study, the performance of the proposed method was compared against that from the gFOTV regularization alone. The results showed that the proposed STGTV method significantly improved gFOTV, with lower localization errors and less spuriously discovered sources.

Keywords: Electrical source brain imaging, Image reconstruction - Performance evaluation, Brain imaging and image analysis
Abstract: Invasive electrophysiological measurement of brain activity is commonly employed during epilepsy surgery to provide final validation of required resection regions. These data are critical to clinical decision making, but manual expert analysis of these data can be compli- cated by the need to relate individual electrode measure- ments to specific brain regions. To improve analysis of these data with source analysis, accurate bioelectric models are needed. Given the proximity of the measurement locations to the generating cortical sources, modeling of electrode- tissue interactions is particularly important for invasive measurements. Here, we evaluate the effect of a finite difference complete electrode model on the accuracy of leadfield computations for invasive electrocorticography. Our results show that in the vicinity of electrode locations, use of the simpler point electrode model produces large topographic and magnitude differences that will likely impact the accuracy of computed source localizations.

Keywords: EEG imaging
Abstract: To investigate whether a motor attempt EEG paradigm coupled with functional electrical stimulation can detect command following and, therefore, signs of conscious awareness in patients with disorders of consciousness, we recorded nine patients admitted to acute rehabilitation after brain lesion. We extracted peak classification accuracy and peak average discriminant power (PADP) and we assessed their correlation to the established coma recovery scale revised (CRS-R) and the agreement with diagnosis based on the novel motor behavior tool (MBT). Only PADP correlated significantly with mbox{CRS-R} and it also outperformed peak accuracy regarding the MBT. We conclude that PADP is more suitable than accuracy to complement CRS-R and MBT in evaluating ambiguous cases and in detecting cognitive motor dissociation.

Keywords: EEG imaging
Abstract: This paper reports successful measurement of even-related potential (ERP) using candle-like dry microneedle electrodes, which can acquire high-quality electroencephalogram (EEG) from hairy parts without any pretreatment. In our previous work, we successfully measured spontaneous EEG activity and its application to assess the stress state of the subjects. ERPs originate from electrophysiological response to stimulus and are one of the most important indices to capture the cognitive and sensory activities. In this work, using the candle-like dry microelectrodes, we demonstrate successful measurement of ERPs elicited by oddball tasks. Two oddball tasks using pure tone stimuli and speech stimuli were assigned to the subjects, where EEG was acquired from the parietal region (Cz in international 10-20 system). Note that no pretreatment, such as removal of hairs and abrasion of the scalp, was applied. As a result, P300 and mismatch negativity (MMN) were successfully measured in the both oddball tasks from the averaged EEG after the stimuli. Based on these results and given the attractive natures of the candle-like dry microneedle electrodes; they do not need any skin treatment and conductive gels and they can measure EEG from the hairy parts, the developed electrodes will accelerate cognitive neuroscience research using ERPs.

Keywords: EEG imaging, Brain imaging and image analysis
Abstract: This work presents a method for automatic independent component (IC) scalp map analysis of electroencephalogram during motor preparation in visuomotor tasks. The strength of this approach is the analysis of the IC scalp maps based on the apriori given mask. This uses an image processing approach, comparable to visual classification used by experts, to automate the selection of relevant ICs in visuomotor tasks. Thirty iterations of the Infomax ICA algorithm were used to test the reliability of the ICs. ICs above 95% quality index were used for IC scalp topography image analysis. Here, we used a linkage-clustering algorithm for IC clustering and gap statistic to estimate the number of clusters. After classifying the components with our approach, the labels were compared to those from well-known MARA ("Multiple Artifact Rejection Algorithm") – an open-source EEGLAB plug-in. It was found that 334 of the 568 labels were in-agreement. MARA labeled 81 out of the 177 source-related components, and 238 out of the 319 non-source-related components, as artifacts. Here, the strength of our approach lies in using an image-processing algorithm to identify the task-specific ICs whereas MARA focuses on the automatic classification of the artifactual ICs by combining stereotyped artifact-specific spatial and temporal features that depend on the electrode montage. After “artefactual” ICs are removed, task-specific ICs still remains to be identified from the remaining “good” ICs where our scalp topography image analysis approach can be applied. Our IC scalp topography image analysis is focused on task-specific IC selection based on an apriori mask, which is not limited to specific EEG features and/or electrode configurations for high-density EEG.

Keywords: Neuromuscular systems - EMG processing and applications, Neurological disorders - Stroke, Neuromuscular systems - Peripheral mechanisms
Abstract: Botulinum toxin (BT) is widely prescribed by physicians for managing spasticity post stroke. In an ongoing study, we examine the spatial pattern of muscle activity in biceps brachii of stroke survivors before and after receiving BT, examined over the course of 11 weeks (2 weeks before – 9 weeks after). We hypothesize that BT alters muscle electrophysiology by disrupting fiber neuromuscular transmission in an inhomogeneous manner and we seek to detect these changes using grid surface electromyography (sEMG). Also, we obtained B-mode ultrasound images to have an accurate interpretation of sEMG data by looking at the fiber angle and subcutaneous fat thickness distribution across muscle. Here, we are reporting a single case where a chronic stroke survivor received BT injection in the biceps brachii (BB). A 16x8 sEMG electrode grid was used to capture the muscle activity distribution of BB during sustained non-fatiguing isometric contraction at 40% of maximal voluntary (MVC) elbow flexion. We obtained the root mean squared (RMS) maps of the signal recorded at each of the 16x8 electrodes. We observed substantial changes in the RMS pattern of BB muscle after receiving BT. More than 80% decrease in sEMG amplitude (RMS) was observed for the channels around the BT injection site as well as about 74% elbow flexion force reduction at the time point of 3–4 weeks post-injection. We also found significant differences between the spatial voluntary activation pattern of pre and post BT RMS maps. We further observed a non-uniform effect and recovery caused by the BT on the distribution of muscle activity. In conclusion, we observed evidence of alteration of the amplitude and pattern of muscle activity after botulinum toxin injection and can document the capability of grid recordings to detect these pattern changes. Our major goals target further investigation to provide an in-depth understanding of the effect of botulinum toxin injection at motor unit level.

Keywords: Neuromuscular systems - Locomotion, Neuromuscular systems - EMG processing and applications, Neurological disorders
Abstract: The aim of this study was to quantify and compare the inter-limb muscle coordination during crawling between typically developing infants and infants with developmental delay. Typically developing (TD, N=20) infants and infants with at risk of developmental delay (ARDD, N=33) or confirmed developmental delay (CDD, N=14) participated in this study. Surface electromyography of eight muscles from arms and legs and the corresponding joint kinematic data were collected while they were crawling on hands and knees at their self-selected velocity. The number of used inter-limb muscle synergies during crawling was identified by nonnegative matrix factorization algorithm. Our results showed that there was no significant difference in the number of used muscle synergies between ARDD and TD infants during crawling. However, a reduced number of synergies were identified in infants with CDD, as compared to that in TD and ARDD infants, indicating constrained neuromuscular control strategy during crawling in developmental delayed infants. The absence of inter-limb muscle synergies may be one of the mechanisms underlying the impairments of crawling in developmental delayed infants, who are at high risk of cerebral palsy. This result also suggests that the metrics of muscle synergy during infant crawling, such as the number of synergy, may be feasible as a biomarker for early diagnosis of infants with cerebral palsy.

Keywords: Neuromuscular systems - EMG processing and applications, Neurorehabilitation, Brain-computer/machine interface
Abstract: It has been shown that maintaining a neutral arm position during collection of pattern recognition training data for myoelectric prosthesis control results in high offline classification accuracies; however, that precision does not translate to real-time applications, when the arm is used in different positions. Previous studies have shown that collecting training data with the arm in a variety of positions can improve pattern recognition control systems. In this work, we extended these findings to real-time myoelectric control in an immersive testing environment using virtual reality. We show that collecting training data for a pattern recognition algorithm under dynamic conditions, where the user moves their arm, significantly improves control efficiency and achievement of testing metrics.

Keywords: Neuromuscular systems - Learning and adaption, Motor neuroprostheses, Human performance - Sensory-motor
Abstract: A central challenge for myoelectric limb prostheses resides in the fact that, as the level of amputation becomes more proximal, the number of functions to be replaced increases, while the number of muscles available to collect input signals for control decreases. Differential activation of compartments from a single muscle could provide additional control sites. However, such feat is not naturally under voluntary control. In this study, we investigated the feasibility of learning to differentially activate the two heads of the bicep brachii muscle (BBM), by using biofeedback via high-density surface electromyography (HD-sEMG). Using a one degree of freedom Fitts’ law test, we observed that eight subjects could learn to control the center of gravity of BBM’s myoelectric activity. In addition, we examined the activations patterns of BBM that allow for the decoding of distal hand movements. These patterns were found highly individual, but different enough to allow for decoding of motor volition of distal joints. These findings represent promising venues to increase the functionality of myoelectrically controlled upper limb prostheses.

Keywords: Neuromuscular systems - EMG processing and applications, Neuromuscular systems - Learning and adaption, Motor neuroprostheses - Prostheses
Abstract: This work focuses on a system for hand prostheses that can overcome the delay problem introduced by classical approaches while being reliable. The proposed approach based on a recurrent neural network enables us to incorporate the sequential nature of the surface electromyogram data and the proposed system can be used either for classification or early prediction of hand movements. Especially the latter is a key to a latency free steering of a prosthesis. The experiments conducted on the first three Ninapro databases reveal that the prediction up to 200 ms ahead in the future is possible without a significant drop in accuracy. Furthermore, for classification, our proposed approach outperforms the state of the art classifiers even though we used significantly shorter windows for feature extraction.

Keywords: Neuromuscular systems - Learning and adaption, Motor learning, neural control, and neuromuscular systems
Abstract: This study used evidence from trial-by-trial errors to understand how humans can generalize what they learn across different movement directions while reaching. We trained 15 healthy subjects to reach in six directions in the presence of challenging visuomotor distortions. We then tested a number of candidate models suggested by the literature of how the brain might use error to improve performance. Our cross-validated results point to a discrete affine model whose generalization, or influence of practice in one direction to neighboring directions, is reduced nearly to zero by 60 degrees away, and the subjects learned 6.25 times more from the error that was observed at a movement direction than neighboring directions.

Keywords: Physiological systems modeling - Closed loop systems, Signal pattern classification, Time-frequency and time-scale analysis - Time-frequency analysis
Abstract: This paper presents the results of our recent work on studying the effects of deep brain stimulation (DBS) and medication on the dynamics of brain local field potential (LFP) signals used for behavior analysis of patients with Parkinson’s disease (PD). DBS is a technique used to alleviate the severe symptoms of PD when pharmacotherapy is not very effective. Behavior recognition from the LFP signals recorded from the subthalamic nucleus (STN) has application in developing closed-loop DBS systems, where the stimulation pulse is adaptively generated according to subjects’ performing behavior. Most of the existing studies on behavior recognition that use STN-LFPs are based on the DBS being “off”. This paper discovers how the performance and accuracy of automated behavior recognition from the LFP signals are affected under different paradigms of stimulation on/off. We first study the notion of beta power suppression in LFP signals under different scenarios (stimulation on/off and medication on/off). Afterward, we explore the accuracy of support vector machines in predicting human actions (“button press” and “reach”) using the spectrogram of STN-LFP signals. Our experiments on the recorded LFP signals of three subjects confirm that the beta power is suppressed significantly when the patients take medication (p-value<0.002) or stimulation (p-value<0.0003). The results also show that we can classify different behaviors with a reasonable accuracy of 85% even when the high-amplitude stimulation is applied.

Keywords: Physiological systems modeling - Signal processing in physiological systems, Physiological systems modeling - Signal processing in simulation, Physiological systems modeling - Signals and systems
Abstract: We propose a framework to evaluate the information content of different stimulation strategies used in neuroprosthetic implants. We analyze the responses of retinal ganglion cells to electrical stimulation using an information theory framework. This methodology allows us to calculate the information content by looking at the consistency of neural responses generated across multiple repetitions of the same stimulation protocol.

Keywords: Physiological systems modeling - Signal processing in physiological systems, Physiological systems modeling - Signal processing in simulation, Physiological systems modeling - Signals and systems
Abstract: We study how initial conditions of the Hodgkin-Huxley model affect the dynamics of simulated neurons. We systematically vary the amplitudes of depolarization currents in order to bring neuron dynamics to stable equilibrium. Our results demonstrate that simulated neurons can have spontaneous spiking or a silent state, depending on the initial conditions. We propose the methodology to study the circumstances under which Purkinje cells transit between hyperpolarized quiescent state (down state) and a depolarized spiking state (up state). We show that results derived using the Hodgkin-Huxley methodology should be carefully analyzed before suggesting a direct relevance to neuroprosthetic implants.

Keywords: Neural networks and support vector machines in biosignal processing and classification, Nonlinear dynamic analysis - Nonlinear filtering, Physiological systems modeling - Signal processing in physiological systems
Abstract: The emergence of deep learning techniques has provided new tools for the analysis of complex data in the field of neuroscience. In parallel, advanced statistical approaches like point-process modeling provide powerful tools for analyzing the spiking activity of neural populations. How statistical and machine learning techniques compare when applied to neural data remains largely unclear. In this research, we compare the performance of a point-process filter and a long short-term memory (LSTM) network in decoding the 2D movement trajectory of a rat using the neural activity recorded from an ensemble of hippocampal place cells. We compute the least absolute error (LAE), a measure of accuracy of prediction, and the coefficient of determination (R^2), a measure of prediction consistency, to compare the performance of these two methods. We show that the LSTM and point-process filter provide comparable accuracy in predicting the position; however, the point-process provides further information about the prediction which is unavailable for LSTM. Though previous results report better performance using deep learning techniques, our results indicate that this is not universally the case. We also investigate how these techniques encode information carried by place cell activity and compare the computational efficiency of the two methods. While the point-process model is built using the receptive field for each place cell, we show that LSTM does not necessarily encode receptive fields, but instead decodes the movement trajectory using other features of neural activity. Although it is less robust, LSTM runs more than 7 times faster than the fastest point-process filter in this research, providing a strong advantage in computational efficiency. Together, these results suggest that the point-process filters and LSTM approaches each provide distinct advantages; the choice of model should be informed by the specific scientific question of interest.

Keywords: Parametric filtering and estimation
Abstract: Neural oscillations reflect the coordinated activity of neuronal populations across a wide range of temporal and spatial scales, and are thought to play a significant role in mediating many aspects of brain function, including attention, cognition, sensory processing, and consciousness. Brain oscillations are typically analyzed using frequency domain methods such as nonparametric spectral analysis, or time domain methods based on linear bandpass filtering. A typical analysis might seek to estimate the power within an oscillation sitting within a particular frequency band. A common approach to this problem is to estimate the signal power within that band, in frequency domain using the power spectrum, or in time domain by estimating the power or variance in a bandpass filtered signal. A major conceptual flaw in this approach is that neural systems, like many physiological or physical systems, have inherent broad-band "1/f" dynamics, whether or not an oscillation is present. Calculating power-in-band, or power in a bandpass filtered signal, can therefore be misleading, since such calculations do not distinguish between broadband power within the band of interest, and true underlying oscillations. In this paper, we present an approach for analyzing neural oscillations using a combination of linear oscillatory models. We estimate the parameters of these models using an expectation maximization (EM) algorithm, and employ AIC to select the appropriate model and identify the oscillations present in the data. We demonstrate the application of this method to electroencephalogram (EEG) data recorded at quiet rest and during propofol-induced unconsciousness.

Keywords: Neural stimulation, Neurological disorders - Stroke, Neurorehabilitation
Abstract: The role of established contralateral cerebro-cerebellar connections on cerebellar injury during stroke has been increasingly revealed in recent years. An extensive number of studies have investigated alteration in inter-hemispheric correlation in order to find brain regions whose responses are specific to restore functional loss and enhance adaptive neural plasticity after stroke. Although, several non-invasive brain stimulation studies have proven their efficacy in the treatment of stroke recovery, finding more effective brain regions that responsible for stroke rehabilitation as well as optimizing neural stimulation protocol are the main goals of further investigations. In this study, the lateral cerebellar nucleus (LCN) was exposed to Low-Intensity Focused Ultrasound (LIFU) to reduce the cerebellar damage resulting from crossed cerebellar diaschisis (CCD) phenomenon after cerebral ischemia. A mouse brain ischemia was induced by middle cerebral artery occlusion (MCAO). A level of decrease in Purkinje cell (PC) number and a quantity of myeloperoxidase (MPO) positive neutrophils in the cerebral cortex were compared between stroke and stroke+LIFU groups after MCAO. In stroke+LIFU group, the increased ipsilateral water content due to tissue swelling was observed, showing an attenuation of brain edema. Prominently, the reduction of the neuroimmune reactivity at the infarct core and the peri-infarct regions, and the increased rate of survival among PCs clearly demonstrated primary evidence of neuroprotective effect induced by LIFU-mediated cerebellar modulation.

Keywords: Neural stimulation, Neural signal processing
Abstract: Bi-directional brain-computer interfaces (BCIs) require simultaneous stimulation and recording to achieve closed-loop operation. It is therefore important that the interface be able to distinguish between neural signals of interest and stimulation artifacts. Current bi-directional BCIs address this problem by temporally multiplexing stimulation and recording. This approach, however, is suboptimal in many BCI applications. Alternative artifact mitigation methods can be devised by investigating the mechanics of artifact propagation. To characterize stimulation artifact behaviors, we collected and analyzed electrocorticography (ECOG) data from eloquent cortex mapping. Ratcheting and phase-locking of stimulation artifacts were observed, as well as dipole-like properties. Artifacts as large as ±1100 µV appeared as far as 15-37 mm away from the stimulating channel when stimulating at 10 mA. Analysis also showed that the majority of the artifact power was concentrated at the stimulation pulse train frequency (50 Hz) and its super-harmonics (100, 150, 200 Hz). Lower frequencies (0-32 Hz) experienced minimal artifact contamination. These findings could inform the design of future bi-directional ECOG-based BCIs.

Keywords: Neural stimulation, Brain functional imaging, Brain functional imaging - fMRI
Abstract: This article considers a new type of integrated multichannel Transcranial Magnetic Stimulator and Magnetic Resonance Imaging (TMS/MRI) system at 3T that is currently being designed. The system will enable unprecedented spatiotemporal control of the TMS-induced electric fields (E-fields) with simultaneous rapid whole-head MRI acquisition to record the brain activity. A critical design question is how TMS coil elements interact with the transmit field (B1+) of the volume coil integrated in 3T MRI systems. In general, the TMS coils are not designed to have any resonant characteristics at the MRI frequency, they may potentially disturb the RF field due to the eddy currents induced. This is especially a concern with a multichannel TMS setup where the subject’s head will be largely covered with the stimulation coils. Therefore, we investigated this problem by computational simulations with realistic TMS coil geometries and a birdcage transmit coil in conjunction with a human body model. We compared the B1+ interaction effects of a commercially available MR-compatible TMS coil with our coil prototype. In both cases, the results show small local changes in the transmit field B1+of the birdcage coil. Maximal Average Specific Absorption Rate (SAR) values over 1g tissue were found to be slightly lower when the TMS elements were present. We conclude that it should be feasible and safe to use the conventional body transmit coil even when an array of TMS coils is used.

Keywords: Neural stimulation, Neurological disorders, Sensory neuroprostheses - Somatosensory
Abstract: In recent years, the trigeminal nerve (CN V) has become a popular target for neuromodulation therapies to treat of a variety of diseases due to its access to neuromodulatory centers. Despite promising preclinical and clinical data, the mechanism of action trigeminal nerve stimulation (TNS) remains in question. In this work, we describe the development and evaluation of a neural interface targeting the mouse trigeminal nerve with the goal of enabling future mechanistic research on TNS. We performed experiments designed to evaluate the ability of a peripheral nerve interface (i.e. cuff electrode) to stimulate the infraorbital branch of the trigeminal nerve. We found that both artificial and naturalistic stimulation of the trigeminal nerve elicited robust cortical responses in the somatosensory cortex that scaled with increases in stimulus amplitude. These results suggest that an infraorbital nerve interface is a suitable candidate for examining the neural mechanisms of TNS in the mouse.

Keywords: Neural stimulation
Abstract: The vagus nerve (VN) is involved in the autonomic regulation of many physiological systems (cardiovascular, respiratory, gastrointestinal, etc.) and its stimulation is already an approved therapy for refractory epilepsy and depression. Other pathologies are thought to be treatable through vagus nerve stimulation (VNS), such as heart failure, cardiac arrhythmia, inflammation or auto-immune diseases. However, the efficacy of the stimulation is not always optimal, partly due to the materials and the architecture of currently available electrodes. Standard electrodes, composed of metallic rings that stimulate the whole diameter of the nerve, are not adapted to experimentations involving spatial selectivity. Efficient and selective charge injection is usually difficult to achieve simultaneously, especially in experimental setups using rodents, due to the thin diameter of their VN. In this paper, we show that we can take advantage of the high charge injection property of conducting polymers to acutely stimulate the vagus nerve in rodents, using individual active electrodes with dimensions 725 um x 450 um. A particular PEDOT:PSS architecture integrating 12 active electrodes is developed and applied to the VN of one rat. A closed-loop VNS system developed in our previous works is used to stimulate the VN while analyzing the heart rate response. Results show the feasibility of this kind of electrodes for acute VNS on rodents and open the path towards new experimentations focused on selective stimulation and recording.

Keywords: Neural stimulation, Neurorehabilitation
Abstract: Non-invasive brain stimulation such as transcranial direct current stimulation (tDCS) involves passing low currents through the brain and is a promising tool for the modulation of cortical excitability. We computationally investigated the effects of the size of the anode in the conventional montage (contralateral supraorbital cathode) using finite element analysis (FEA) for the targeted leg area of the motor cortex where tDCS is challenging due to the depth and orientation of the leg motor area in the inter-hemispheric fissure. We used FEA to develop two anode sizes (same cathode size) with the same current density but different electric field magnitude at the targeted leg area of the motor cortex. Then, we evaluated the effects of the two anode sizes via neurophysiological testing on twelve healthy subjects, seven males and five females (age: 21-36 years, all right-leg dominant). Here, conventional anodal tDCS electrode montage for the leg area of the motor cortex used a large-anode (5cmx7cm, current strength 2mA) which was compared based on transcranial magnetic stimulation (TMS)-induced motor evoked potentials (MEP) with a small-anode (3.5cmx1cm at 0.2mA) montage of the same current density at the skin-electrode interface and identical contralateral supraorbital cathode placement. Small-anode decreased the electric field magnitude by almost one-tenth but still got a similar statistically significant (P<0.05) increase in the cortical excitability (MEP) at the targeted leg motor area when compared to sham tDCS. Since the electric field magnitude was similar at the scalp (skin-electrode interface) level but differed significantly at the leg motor area in the inter-hemispheric fissure, so a possible contribution of scalp sensory nerve responses to electrocutaneous stimulation is proposed.

Keywords: Neural networks and support vector machines in biosignal processing and classification
Abstract: Automated monitoring and analysis of eating behavior patterns, i.e., ”how one eats”, has recently received much attention by the research community, owing to the association of eating patterns with health-related problems and especially obesity and its comorbidities. In this work, we introduce an improved method for meal micro-structure analysis. Stepping on a previous methodology of ours that combines feature extraction, SVM micro-movement classification and LSTM sequence modelling, we propose a method to adapt a pretrained IMU-based food intake cycle detection model to a new subject, with the purpose of improving model performance for that subject. We split model training into two stages. First, the model is trained using standard supervised learning techniques. Then, an adaptation step is performed, where the model is fine-tuned on unlabeled samples of the target subject via semisupervised learning. Evaluation is performed on a publicly available dataset that was originally created and used in [1] and has been extended here to demonstrate the effect of the semisupervised approach, where the proposed method improves over the baseline method.

Keywords: Neural networks and support vector machines in biosignal processing and classification, Signal pattern classification, Physiological systems modeling - Multivariate signal processing
Abstract: Over- and under-sedation are common in critically ill patients admitted to the Intensive Care Unit. Clinical assessments provide limited time resolution and are based on behavior rather than the brain itself. Existing brain monitors have been developed primarily for non-ICU settings. Here, we use a clinical dataset from 154 ICU patients in whom the Richmond Agitation-Sedation Score is assessed about every 2 hours. We develop a recurrent neural network (RNN) model to discriminate between deep vs. no sedation, trained end-to-end from raw EEG spectrograms without any feature extraction. We obtain an average area under the ROC of 0.8 on 10-fold cross validation across patients. Our RNN is able to provide reliable estimates of sedation levels consistently better compared to a feed-forward model with simple smoothing. Decomposing the prediction error in terms of sedatives reveals that patient-specific calibration for sedatives is expected to further improve sedation monitoring.

Keywords: Neural networks and support vector machines in biosignal processing and classification
Abstract: Given the world-wide prevalence of heart disease, the robust and automatic detection of abnormal heart sounds could have profound effects on patient care and outcomes. In this regard, a comparison of conventional and state-of-the-art deep learning based computer audition paradigms for the audio classification task of normal, mild abnormalities, and moderate/severe abnormalities as present in phonocardiogram recordings, is presented herein. In particular, we explore the suitability of deep feature representations as learnt by sequence to sequence autoencoders based on the auDeep toolkit. Key results, gained on the new Heart Sounds Shenzhen corpus, indicate that a fused combination of deep unsupervised features is well suited to the three-way classification problem, achieving our highest unweighted average recall of 47.9% on the test partition.

Keywords: Signal pattern classification, Neural networks and support vector machines in biosignal processing and classification, Connectivity measurements
Abstract: Identification of the treatment-related responders for adolescent Major Depressive Disorder (MDD) is urgently needed to develop effective treatments. In this paper, machine learning based classifiers are used to reveal anatomical features as responders for distinguishing MDD patients who have received treatment from those who never received any treatment. The features are drawn from two sets of measurements: 1)anatomical connectivity defined by diffusion tensor imaging measurements between a pair of brain regions, and 2) topological measurements from anatomical networks. Feature selection was performed based on p-value and minimum redundancy maximum relevance (mRMR) method to achieve improved classification accuracy. The classification performance is evaluated with a leave-one-out cross-validation method using 37 treated and 15 untreated subjects. The proposed methodology achieves 73% accuracy, 100% specificity, and 100% precision for 52 subjects. The most distinguishing features are the strength of the right hippocampus of the mean diffusivity (MD) network at 18% density and of the track-count (TR) network, the participation coefficient of the left middle temporal gyrus of the radial diffusivity (RD) network at 20% density, the axial diffusivity (AD) connectivity between right middle temporal gyrus and right supramarginal gyrus, the betweenness centrality of the right hippocampus of the TR network at 11% density, the apparent diffusion coefficient (ADC) connectivity between the left pars opercularis and the left rostral anterior cingulate cortex, the clustering coefficient of the middle anterior corpus callosum of the TR network at 11% density, and the AD connectivity between the left pars opercularis and the left rostral anterior cingulate cortex.

Keywords: Neural networks and support vector machines in biosignal processing and classification, Signal pattern classification, Data mining and processing - Pattern recognition
Abstract: Acoustic analysis using signal processing tools can be used to extract voice features to distinguish whether a voice is pathological or healthy. The proposed work uses spectrogram of voice recordings from a voice database as the input to a Convolutional Neural Network (CNN) for automatic feature extraction and classification of disordered and normal voice. The novel classifier achieved 88.5%, 66.2% and 77.0% accuracy on training, validation and testing data set respectively on 482 normal and 482 organic dysphonia speech files. This indicates that the proposed novel algorithm can effectively been used for screening pathological voice recordings.

Keywords: Signal pattern classification, Neural networks and support vector machines in biosignal processing and classification, Physiological systems modeling - Signal processing in physiological systems
Abstract: Inspiratory Flow Limitation (IFL) is a phenomenon associated with narrowing of the upper airway, preventing an increase in inspiratory airflow despite an elevation in intrathoracic pressure. It has been shown that quantification of IFL might complement information provided by standard indices such as the apnea-hypopnea index (AHI) in characterizing sleep disordered breathing and identifying subclinical disease. Defining guidelines for visual scoring of IFL has been of increasing interest, and automated methods are desirable to avoid inter-scorer variability and allow analysis of large datasets. In addition, as recording instrumentation and practices may vary across hospitals and laboratories, it is useful to assess the influence of the recording parameters on the accuracy of the automated classification. We employed nasal pressure signals recorded as part of polysomnography (PSG) studies in 7 patients. Two experts independently classified approximately 2000 breaths per subject as IFL or non-IFL, and we used the consensus scoring as the gold standard. For each breath, we derived features indicative of the shape and frequency content of the signals and used them to train and validate a Support Vector Machine (SVM) to distinguish IFL from non-IFL breaths. We also assessed the effect of signal filtering (down-sampling and baseline-removal) on classification performance. The performance of the classifier was excellent (accuracy ~ 93%) for the raw signals (collected at 125 Hz with no filtering), and decreased for increasing high-pass cut-off frequencies (fc = [0.05, 0.1, 0.15, 0.2] Hz) down to 84% for fc=0.2 Hz and for decreasing sampling rate ( fs = [20, 50, 75, 100] Hz) down to ~85% for fs=20 Hz. Loss of performance was minimized when the classifier was re-trained using data with matched filtering characteristics (accuracy > 89%). We can conclude that the SVM feature-based algorithm provides a reliable and efficient tool for breath-by-breath classification.

Keywords: Ultrasound imaging - Photoacoustic/Optoacoustic/Thermoacoustic
Abstract: Photoacoustic imaging has been intensively studied in recent years, and many of the achievements have already been applied in important biomedical and clinical applications, e.g. spectroscopic photoacoustic imaging to extract functional and molecular information. However, spectroscopic photoacoustic imaging requires expensive and bulky tunable laser source, which severely hinder its further development towards portable device. In this paper, we propose a novel imaging method, named optical spectroscopic ultrasound displacement (OSUD) imaging, which enables optical spectroscopic imaging in deep scattering tissue using multiple low-cost continuous-wave laser sources and ultrasound imaging equipment. The principle of the OSUD imaging method will be introduced, and followed by preliminary experimental results. The OSUD imaging may provide another pathway to provide spectroscopic optical absorption contrast in deep scattering tissue beyond commonly used photoacoustic imaging.

Keywords: Ultrasound imaging - Photoacoustic/Optoacoustic/Thermoacoustic
Abstract: Photoacoustic (PA) tomography enables imaging of optical absorption property in deep scattering tissue by listening to the PA wave. However, it is an open challenge that the conversion efficiency from light to sound based on PA effect is extremely low. The consequence is the poor signal-to-noise ratio (SNR) of PA signal especially in scenarios of low laser power and deep penetration. The conventional way to improve PA signal’s SNR is data averaging, which however severely limits the imaging speed. In this paper, we propose a new adaptive wavelet threshold de-noising (aWTD) algorithm, and apply it in photoacoustic tomography to increase the PA signal’s SNR without sacrificing the signal fidelity and imaging speed. PA image quality in terms of contrast is also significantly improved. The proposed method provides the potential to develop real-time low-cost PA tomography system with low-power laser source.

Keywords: Ultrasound imaging - Photoacoustic/Optoacoustic/Thermoacoustic, Ultrasound imaging - High-frequency technology, Image visualization
Abstract: Photoacoustic (PA) imaging is rapidly progressing imaging modality in which very short pulsed laser causes thermal expansion to generate PA signal. In previous PA microscope systems, relatively long time was required for image acquisition because they required mechanical scan of the PA transducer. The objective of the present study is to develop fast laser scanning optical resolution photoacoustic microscopy (OR-PAM) with fixed PA signal detector to realize very high frame rate PA imaging. Q-switched Nd:YAG laser with the wavelength of 532 nm, pulse width of 5.5 ns and pulse repetition rate of up to 50 kHz was equipped in the system. Low frequency PA detector was comprised of a glass prism configuring acoustic focusing and a polymethyl methacrylate (PMMA) prism with 5 MHz PZT (lead zirconate titanate) transducer. High frequency PA detector was comprised of the same glass prism and a glass prism with ZnO (zinc oxide) thin film transducer. Galvano scanner operating in the air was controlled by a microcomputer to scan laser beam. PA signal was detected with the fixed PA detector thus realized the frame rate of 5 fps for C-mode equivalent to 500 fps for B-mode. The lateral resolution of the low frequency system was found to be 11.6 μm and the blood vessel on the surface of cod roe was clearly visualized. The system may be applicable for imaging blood flow and vascular dynamics of micro vessel.

Keywords: Ultrasound imaging - Photoacoustic/Optoacoustic/Thermoacoustic
Abstract: Multi-wavelength photoacoustic (PA) imaging has been studied extensively to explore the spectroscopic absorption contrast of biological tissues. To generate strong PA signals, a high-power wavelength-tunable pulsed laser source has to be employed, which is bulky and quite expensive. In this paper, we propose a hybrid multi-wavelength PA imaging (hPAI) method based on combination of single-wavelength pulsed and multi-wavelength continuous-wave (CW) laser sources. By carefully controlling laser illumination sequence (pulse-CW-pulse), and extracting the PA signals’ difference before and after heating of CW lasers, the optical absorption property of multi-wavelength CW lasers could be obtained. Compared with conventional PA imaging, the proposed hPAI shows much lower system cost due to the usage of single-wavelength pulsed laser and cheap CW lasers. Theoretical analysis and analytical model are presented in this paper, followed by proof-of-concept experimental results.

Keywords: Ultrasound imaging - Photoacoustic/Optoacoustic/Thermoacoustic
Abstract: As one of the fastest-growing imaging modalities in recent years, photoacoustic (PA) imaging has attracted tremendous research interest for various applications including anatomical, functional and molecular imaging. However, the PA signal’s amplitude is usually quite weak and can be easily distorted by instrumental noise and interference, which can severely degrade the image quality. To improve the PA signal’s signal-to-noise ratio efficiently, this paper introduces a pattern-learning based PA (PLPA) detection method to eliminate the periodically interference noise for PA sensing and imaging. Both simulation and experimental results are demonstrated to prove the validity of the proposed algorithm.

Keywords: Neural signals - Coding, Human performance - Cognition, Brain-computer/machine interface
Abstract: Working Memory (WM) processing is central for human cognitive behavior. Using neurofeedback training to enhance the individual WM capacity is a promising technique but requires careful consideration when choosing the feedback signal. Feedback in terms of univariate spectral power (specifically theta and alpha power) has yielded questionable behavioral effects. However, a promising new direction for WM neurofeedback training is by using a measure of WM that is extracted by multivariate pattern classification. This study recorded EEG oscillatory activity from 15 healthy participants while they were engaged in the n-back task, n∈[1,2]. Univariate measures of the theta, alpha, and theta-over-alpha power ratio and a measure of WM extracted from multivariate pattern classification (of n-back task load conditions) was compared in relation to individual n-back task performance. Results show that classification performance is positively correlated to individual 2-back task performance while theta, alpha and theta-over-alpha power ratio is not. These results suggest that the discriminability of multivariate EEG oscillatory patterns between two WM load conditions reflects individual WM capacity.

Keywords: Neural signals - Coding, Neural signals - Blind source separation (PCA, ICA, etc.), Brain-computer/machine interface
Abstract: In this paper, scalp electroencephalographic (EEG) signals were recorded from 10 subjects during hand grasping. Six objects that span different grasp types were used. Grasp kinematics were recorded using CyberGlove. From a training subset of the data, kinematic synergies were determined and their reconstruction weights in these grasps were calculated. EEG features (power spectral densities in four low and high frequency bands) were trained on kinematic synergy weights using multivariate linear regression. Using this model, kinematics from testing subset of data were decoded from EEG with 3-fold cross validation. Results are compared to chance level to determine if reconstruction weights are related to EEG features. Results indicate that EEG features can decode synergy-based movement generation. Study implications and future implementations were discussed.

Keywords: Neural signals - Coding, Neural signals - Blind source separation (PCA, ICA, etc.), Neuromuscular systems - EMG processing and applications
Abstract: Robust human-machine interactions require accurate and intuitive interfaces. Neural signals associated with muscle activities are widely used as the interface signals. This preliminary study evaluated the feasibility of a novel neural-drive-based interface in estimating the individual finger joint angles. The motor unit pool discharge probability was used to predict the neural drive associated with the fine control of the finger joint angle during individual finger extension movement. To obtain the neural drive information, individual motor unit discharge events were extracted from the decomposition of high-density surface electromyogram (sEMG) signals, and discharge events from different motor units were pooled to from a composite discharge event train. The neural-drive-based estimate was obtained by calculating the probability (normalized frequency) of the populational motor unit discharge. The global EMG signal (root-mean-squared value) was also used to estimate the joint angles as a control condition. Our preliminary results showed that the accuracy and stability of the neural-drive-based approach outperformed the classic EMG-based method. Our findings suggest that the novel neural-drive-based interface could be used as a promising control input for intuitive dynamic control of a robotic hand.

Keywords: Neural signals - Blind source separation (PCA, ICA, etc.), Brain-computer/machine interface, Neural signal processing
Abstract: Muscular artifacts often contaminate electroencephalograms (EEGs) and deteriorate the performance of brain–computer interfaces (BCIs). Although many artifact reduction techniques are available, most of the studies have focused on their reduction ability (i.e. reconstruction errors), and it has been missing to evaluate their effect on the performance of BCIs. This study aims at evaluating the performance of a state-of-the-art muscular artifact reduction technique on a scenario of a steady-state visual evoked potentials (SSVEPs)-based BCI. The performance was evaluated based on a semi-simulation setting using a benchmark dataset of SSVEPs artificially contaminated by muscular artifacts acquired from the trapezius. Our results showed that combining the artifact reduction method and the classification algorithm based on the task-related component analysis gained improved classification accuracy. Interestingly, the artifact reduction setting minimizing the reconstruction errors, i.e. elaborately recovering the true EEG waveforms, was inconsistent to the one maximizing the classification performance. The results suggest that artifact reduction methods should be tuned so as to maximize performance of BCIs.

Keywords: Cardiac electrophysiology - Inverse problems, Cardiac electrophysiology - Simulation for cardiac arrhythmia
Abstract: Invasive cardiac catheterisation is a precursor to ablation therapy for ventricular tachycardia. Invasive cardiac diagnostics are fraught with risks, especially for young children. Decades of research has been conducted on the inverse problem of electrocardiography, which can be used to reconstruct Heart Surface Potentials (HSPs) from Body Surface Potentials (BSPs), for non-invasive cardiac diagnostics. State of the art solutions to the inverse problem are unsatisfactory, since the inverse problem is known to be ill-posed. In this paper we propose a novel approach to reconstructing HSPs from BSPs using a Time- Delay Artificial Neural Network (TDANN). We first design the TDANN architecture, and then develop an iterative search space algorithm to find the parameters of the TDANN, which results in the best overall HSP prediction. We use real-world recorded BSPs and HSPs from individuals suffering from serious cardiac conditions to validate our TDANN. The results are encouraging, in that the predicted and recorded HSPs have an average correlation coefficient of 0.7 under diseased conditions.

Keywords: Cardiac electrophysiology - Defibrillation, ablation, and cardioversion, Cardiac electrophysiology - Ventricular arrhythmia mechanisms
Abstract: One major cause of death in the industrialized world is sudden cardiac death, which so far can be reliably treated only by applying strong electrical shocks. Developing improved methods, aimed at lowering shock intensity and associated side effects has potentially significant clinical implications. Thus, photogenetic optical stimulation using structured illumination has been introduced as a promising experimental tool to investigate mechanisms underlying multi-site pacing and to optimize potential low-energy approaches. Furthermore, aim of this work is to strengthen the application of photogenetic tools for cardiac arrhythmia research, which in return also helps to improve applicable technologies towards tissue-protective defibrillation.

Keywords: Cardiac electrophysiology - Defibrillation, ablation, and cardioversion, Cardiac electrophysiology - Pacemakers, Cardiac mechanics, structure & function - Ventricular mechanics
Abstract: We present the first invasive use of a stiff, multi-LED optical probe for intramural optical stimulation of cardiac tissue. We demonstrate that optical pacing is possible with high spatial and temporal resolution in transgenic mice expressing channelrhodopsin-2. The technical implementation of this study builds on optical probes recently developed and tested ex vivo in cerebral tissue of mice. The probes comprise LEDs integrated on flexible substrates stiffened by silicon-based MEMS structures enabling the successful penetration into the cardiac tissue. The probe technology is extended to allow dual-sided illumination for directional tissue stimulation. Implantation trials affirm the ability to optically pace the isolated perfused heart at stimulation frequencies between 4Hz and 12Hz with experimentally determined emittance levels of 10mW/mm2. Rapid activation of two distant LEDs could reliably be used to induce short runs of ventricular fibrillation, while simultaneous activation of all LEDs allowed termination of re-entrant rhythm disturbances (optical defibrillation). Thus, spatially-resolved intramural pacing and rhythm control of the isolated heart is possible using stiff, multi-LED optical probes.

Keywords: Cardiovascular and respiratory signal processing - Cardiovascular signal processing, Cardiac electrophysiology - Atrial fibrillation, Cardiac electrophysiology - Simulation for cardiac arrhythmia
Abstract: Catheter ablation therapy is an effective approach to treat different arrhythmias. Cardiac conduction velocity (CV), extracted from intracardiac electrograms, shows the speed and direction of the wavefront propagation at different sites and is an insightful feature to guide ablation therapy. To create a propagation map, a small mapping catheter with a high density of electrodes is usually used to sequentially collect electrograms from different sites in a desired chamber of the heart. The CV and isochrone surface estimations are very challenging during complex arrhythmias such as atrial fibrillation, where multiple wavefronts simultaneously excite different cardiac sites. Specifically, the performances of CV estimators significantly degrade at catheter sites where wavefronts collide. This is mainly because during collision, different wavefronts pass the areas under different electrodes of the catheter. Consequently, the activation times of the electrodes are the results of different wavefronts, and there are sharp changes in isochrone line patterns in the vicinity of the collision's border. In this paper, we propose a method that is able to identify the collision sites and improve the estimation of CV and isochrone maps. The proposed method finds the electrodes of the catheter that are excited by a similar wavefront and then estimates the corresponding isochrone lines for that wavefront. Our simulation results confirmed the efficiency of the proposed method during collision.

Keywords: Cardiac mechanics, structure & function - Atrial Fibrillation, Cardiovascular and respiratory signal processing - Cardiovascular signal processing
Abstract: Atrial fibrillation (AF) is the most common type of cardiac arrhythmia. Although not life-threatening itself, AF significantly increases the risk of stroke and myocardial infarction. Current tools available for screening and monitoring of AF are inadequate and an unobtrusive alternative, suitable for long-term use, is needed. This paper evaluates an atrial fibrillation detection algorithm based on wrist photoplethysmographic (PPG) signals. 29 patients, recovering from surgery in the post-anesthesia care unit were monitored. 15 patients had sinus rhythm (SR, 67.5 ± 10.7 years old, 7 female) and 14 patients had AF (74.8 ± 8.3 years old, 8 female) during the recordings. Inter-beat intervals (IBI) were estimated from PPG signals. As IBI estimation is highly sensitive to motion or other types of noise, acceleration signals and PPG waveforms were used to automatically detect and discard unreliable IBI. AF was detected from windows of 20 consecutive IBI with 98.45 ± 6.89% sensitivity and 99.13 ± 1.79% specificity for 76.34 ± 19.54% of the time. For the remaining time, no decision was taken due to the lack of reliable IBI. The results show that wrist PPG is suitable for long term monitoring and AF screening. In addition, this technique provides a more comfortable alternative to ECG devices.

Keywords: Cardiovascular and respiratory signal processing - Complexity in cardiovascular or respiratory signals, Cardiovascular and respiratory signal processing - Heart Rate and Blood Pressure Variability, Cardiovascular regulation - Autonomic nervous system
Abstract: The study compares a recently proposed short-term model-based linear multiscale complexity approach to a single-scale application of the same method and to a model-free nonlinear one based on the computation of conditional entropy with the aim at assessing the complementary information. Comparison was carried out over 24 hours Holter recordings of heart period variability during daytime and nighttime in 12 healthy men (age: 34-55 years). Single-scale methods were able to detect the increased complexity of the cardiac control during nighttime. Multiscale complexity analysis showed that this increase was due to an increase of complexity in the low frequency band (from 0.04 to 0.15 Hz), while complexity in the range of frequencies typical of the respiratory rate was unmodified. Regardless of the method (i.e. linear or nonlinear) single-scale complexity indexes were uncorrelated to the multiscale ones. We conclude that short-term model-based linear multiscale complexity approach provides complementary information to single-scale methods in an application devoted to the analysis of cardiac control from 24 hours Holter recordings.

Keywords: Cardiovascular and respiratory signal processing - Cardiovascular signal processing
Abstract: Fetal T-wave alternans (TWA) is a still little-known marker for severe fetus-heart instabilities and may be related to some currently unjustified fetal deaths. Automatically detecting TWA on direct fetal electrocardiograms (DFECG) means possibility of providing fetuses the right treatment during delivery. Instead, automatically identifying TWA on indirect fetal electrocardiograms (IFECG) means possibility of providing fetuses the right treatment even during pregnancy, when taking actions for outcome improvement is still possible. Moreover, TWA identification from IFECG is noninvasive, and thus safe for both fetuses and mothers. The aim of this work was testing the heart-rate adaptive match filter (HRAMF) for automatic TWA identification in IFECG and comparing HRAMF performance in IFECG against DFECG. To this aim, simultaneously recorded DFECG and IFECG tracings from 5 healthy fetuses were used (“Abdominal and Direct Fetal Electrocardiogram Database” from Physionet). TWA measurements (frequency, mean amplitude, maximum amplitude, and amplitude standard deviation) in IFECG (1.09±0.04 Hz, 11±5 µV, 21±12 µV and 7±3 µV) were of the same order of magnitude of those in DFECG (1.07±0.02 Hz, 9±2 µV, 30±11 µV and 6±2 µV). Moreover, a direct correlation (ρ) was found between maximum TWA and fetal heart rate (IFECG: ρ=0.999; P=0.022; DEFEG: ρ=0.642; P=0.243). Thus, HRAMF was able to detect TWA from IFECG as well as from DFECG.

Keywords: Cardiovascular and respiratory signal processing - Cardiovascular signal processing, Cardiovascular and respiratory system modeling - Blood flow models
Abstract: Abstract— T-Wave Alternans (TWA) in the electro cardiogram (ECG) has been widely investigated as a potential predictor of ventricular arrhythmia. However, large clinical trials show that TWA has a high negative predictive value (NPV) but poor positive predictive value (PPV). Therefore, there is need for exploration of approaches to improve PPV of TWA. More recent studies suggest that whether alternans is spatially concordant or discordant affects arrhythmic potential. Results of our previous animal and simulation studies show that the phase relation between depolarization and repolarization alternans has an effect on the transition of concordant to discordant alternans. Towards the eventual goal of developing indexes that complement TWA and improve prediction of arrhythmia, the objectives in this study were to verify the existence of R wave amplitude alternans (RWAA, a surrogate of depolarization alternans) and investigate the phase relationship between RWAA and TWA in clinical grade ECGs. Results show that RWAA does occur in ECGs and that the phase relationship between RWAA and TWA can be labile. These results support further investigation of the co-occurrence of these alternans for prediction of arrhythmic events.

Keywords: Cardiovascular and respiratory signal processing - Complexity in cardiovascular or respiratory signals, Cardiovascular regulation - Heart rate variability, Cardiovascular regulation - Baroreflex
Abstract: In congestive heart failure (CHF), dilated cardiomyopathy (DCM) and ischemic cardiomyopathy (ICM) are two highly related pathologies that are not fully characterized. The aim of this study is to assess respiratory sinus arrhythmia (RSA) index of the parasympathetic system, in order to discriminate between both pathologies, DCM and ICM. For this, ECG-signals of 49 subjects (12 DCM patients, 21 ICM patients, 6 ICM patients with diabetes mellitus (DM) type II and 10 control subjects) from the database HERIS II and of 173 subjects (50 DCM, 50 ICM, 15 DCM with DM type II, 15 ICM with DM type II and 47 control subjects) from the database MUSIC2 were analyzed. The RSA was quantified using linear and non-linear analysis methods (fractal dimension and entropy). The results showed a significant difference between ICM and DCM subjects (p=0.013) with a sensitivity of 83% and specificity of 90%. Decreasing RSA values were present in CHF patients, especially in ICM patients, in comparison with healthy subjects. Alterations in the parasympathetic system due to DM were also identified.

Keywords: Cardiovascular and respiratory signal processing - Blood pressure measurement, Cardiovascular regulation - Blood pressure variability
Abstract: Physiological records are one of the most relevant elements to obtain objective information from the patients. The presence of artifacts in biomedical signals can give misleading in the analysis of information that these signals give. The blood pressure signal is one of the records clearly affected by different artifacts, especially the ones due from the calibration episodes. We propose a method to reconstruct different episodes of artifacts in these signals. This method is sustained on the detection of the events of the signal, differentiating between to the physiological cycles and the artifacts. The performance of the method is based on the detection of the cycles and artifact’s position, the identification of the number of cycles to reconstruct, and the prediction of the cycle model used to generate the missing cycles. The parameter Theta_E represents the difference between the area under the curve when two events are compared. The value of this parameter is low when two similar events are compared like the physiological cycles, whereas it is high comparing a cycle with an artifact. An adaptive threshold is defined to identify the artifact episodes. The number of cycles to reconstruct is generated considering the same number of their neighbours physiological cycles, to left and right, of the original signal. Finally, the performance of the method has been analyzed comparing the number of events and artifacts detected and their correct reconstruction. According to the results, the reconstruction error was less than 1% in all cases.

Keywords: Diagnostic devices - Physiological monitoring
Abstract: Microbiota analysis is a fundamental element for a better understanding of microbiota role, its relationship with the human body and its impact on different pathologies. There is today no non-invasive tool for easy collection of the microbiota in the small intestine. In this paper, we describe the development of such a device that opens the way to new diagnostic techniques. The device is based on a capsule designed as a passive system to maximize the safety during its use. Originality of the design relies in the use of a bistable mechanism obtained using additive manufacturing in order to provide a compact design with integration of opening, sampling and closing functions within the capsule. Design, implementation and initial lab evaluation of sampling are presented.

Keywords: Wearable or portable devices for vital signal monitoring, Diagnostic devices - Physiological monitoring, Plethysmography
Abstract: The goal of this study is to examine the influence of various imitated and video invoked emotions on the vital signs (respiratory and pulse rates). We also perform an analysis of the possibility to extract signals from sequences acquired with cost-effective cameras. The preliminary results show that the respiratory rate allows for better separation of some emotions than the pulse rate, yet this relation highly depends on a subject. The invoked positive emotion resulted in a respiratory rate difference > 1.8bpm, comparing to the average respiration rate of all neutral results (in 89% of observations). Visual facial expression in many cases was insufficient for emotion recognition (in video based experiment only 11.4% of visual responses were classified as an expected emotion).

Keywords: Wearable or portable devices for vital signal monitoring, Health technology - Verification and validation
Abstract: Sleep disturbances in modern-day life lead to cognitive and motor performance impairments in everyday tasks such as gait. The most common symptom of these disturbances is daytime sleepiness, which can be assessed by questionnaires such as the Epworth Sleep Scale (ESS). The ESS evaluates sleep health and daytime dysfunction. The goal of this study is to assess the influence of sleepiness on a motor-auditory synchrony task, rhythmed gait. High and low sleepiness clusters were formed based on the participants ESS scores. Walking a treadmill, two different rhythmic auditory stimulus conditions were set with a metronome: isochronous and non-isochronous. Reflexive markers on both heels with seven infrared cameras were used to assess the difference between footfall and metronome beep, a synchronization error. There was a tendency to anticipate the beep in the HS group when compared to the LS group only in the non-isochronous stimulus condition that was statistically significant. Sleep disturbances that generate daytime sleepiness may bring detrimental effects on brain areas that could be responsible for the real-time adjustment of gait and sustained attention. These impairments may be responsible for the larger synchronization error with larger relative phase of the group with high sleepiness. More studies are necessary involving other parameters of sleep and gait to identify sleep disturbances through gait analysis.

Keywords: Diagnostic devices - Physiological monitoring, Wearable or portable devices for vital signal monitoring, Clinical engineering - Device alarm, alert, and communication systems
Abstract: This paper proposes and implements an intuitive and pervasive solution for neonatal EEG monitoring assisted by sonification and deep learning AI that provides information about neonatal brain health to all neonatal healthcare professionals, particularly those without EEG interpretation expertise. The system aims to increase the demographic of clinicians capable of diagnosing abnormalities in neonatal EEG. The proposed system uses a low-cost and low-power EEG acquisition system. An Android app provides single-channel EEG visualization, traffic-light indication of the presence of neonatal seizures provided by a trained, deep convolutional neural network and an algorithm for EEG sonification, designed to facilitate the perception of changes in EEG morphology specific to neonatal seizures. The multifaceted EEG interpretation framework is presented and the implemented mobile platform architecture is analyzed with respect to its power consumption and accuracy.

Keywords: Diagnostic devices - Physiological monitoring, Health technology - Verification and validation
Abstract: Quantitative assessment of the muscle tone is important when studying patients with neurological disorders such as Parkinson’s disease (PD). For the assessment of therapeutic progress, quantitative and objective outcome measures are needed. This article presents a novel electromechanical device to monitor the quantitative rigidity of the wrist joint against passive movement. The novel device is equipped with an electrical motor to move the wrist joint in a flexion-extension manner with different velocity profiles. The accuracy of the device was measured in terms of position, velocity and torque accuracy. The feasibility of the measurement procedure was tested in a pilot study with four PD patients and 12 healthy controls (HC), at velocities of 10°/s, 50°/s, and 100°/s. The position and velocity of the developed device were (0.005 ± 0.105)° and (0.734 ± 0.276)°/s, unloaded, and (0.003 ± 0.113)° and (0.013 ± 0.038)°/s, loaded with a relaxed arm, respectively. The torque accuracy was (15.029 ± 2.235) mNm. The comparison of the median rigidity between the PD patients and HC showed significant differences at all tested velocities, during both flexion and extension movements. This device proved to have sufficient accuracy and sensitivity to precisely measure the interaction torque at the wrist joint and to differentiate PD rigidity from normal muscle tone. The device, thus provides a quantitative and objective measure of rigidity in PD.

Keywords: Diagnostic devices - Physiological monitoring, Clinical laboratory, assay and pathology technologies, Neural stimulation (including deep brain stimulation)
Abstract: Anorectal manometry is a diagnostic technique used to investigate the correct mechanical performance of the internal anal sphincter (IAS). By distending the rectal ampulla while recording changes in the luminal pressure, this method allows for characterizing the anorectal reflex. It can also provide, indirectly, information about the electrical activity of the IAS. In this study, seventeen neonates having 24-hour delayed passage of meconium or presenting distal intestinal obstruction symptoms underwent anorectal manometry to discard Hirschsprung’s disease. All patients had normal anorectal reflex. The time delay between stimulation of the rectal ampulla and the relaxation of the anal canal was studied. The average period of the pressure fluctuations was 5.44 ± 0.13 s. The overall duration of the relaxation time was 9.71 ± 0.21 s. The maximum lag between the onset of the stimulus and the relaxation of the IAS was 2.90 s, and was achieved when the stimulus was applied following a local maximum of the pressure wave. The existence of a refractory period during the suprathreshold depolarization of smooth muscle cells can explain the evidence of a temporal delay between the stimulus and the mechanical response. In occasions, relaxation appeared first distally. This phenomenon can be explained by the arrangement and morphology of bipolar cells, which may evidence the anisotropic propagation of the mechanical activity. These data may contribute to depict the alterations in excitability underlying the relaxation reflex by means of manometric recording of the anal canal.

Keywords: Computer modeling for treatment planning, Cardiovascular assessment and diagnostic technologies, Clinical engineering
Abstract: Coronary bypass grafting (CABG) is a surgical procedure for anastomosing small grafts to the coronary vessels. The bypass graft bridges the occluded or diseased coronary artery, allowing sufficient blood flow to deliver oxygen and nutrients to the heart muscles. Patient-specific (PS) anatomy obtained from coronary computed tomography angiography (CCTA) was used to generate a 3D aorto-coronary model (pre-surgery). Additionally, three more models with idealized grafts (individual and sequential grafts), were created using Boolean operations to represent post-surgery configuration. Fractional flow reserve (FFR) and wall shear stress (WSS) were estimated from the computational fluid dynamics (CFD). The pre-surgical FFR values for all the three left coronary arteries were significant (FFR<0.80). The flow was restored (FFR>0.80) distal to stenosis in all the three post-surgical idealized graft models. Peak WSS values of 468, 336 and 295 dynes/cm2 were observed at the toe of the individual end-to-side anastomosis for the three graft models. More importantly, low WSS (<100 dynes/cm2) prevails at the heel and the walls opposite to the anastomosis in the sequential graft models. The prevailing low WSS at the heel and the wall bed opposite to anastomosis, in a sequential graft model, reduces restenosis rates and promotes a uniform hemodynamic environment for a better long-term patency of the graft. PS-CFD simulations based on CCTA can be helpful in assessing the hemodynamic parameters of graft models for optimal surgical planning.

Keywords: Computer modeling for treatment planning, Models of therapeutic devices and systems, Clinical engineering
Abstract: Abstract - Tumor treating fields (TTFields) is an anticancer treatment that inhibits tumor growth with alternating electrical fields. Finite element (FE) methods have been used to estimate the TTFields intensity as a measure of treatment “dose”. However, TTFields efficacy also depends on field direction and exposure time. Here we propose a new FE based approach, which uses all these parameters to quantify the average field intensity and the amount of unwanted directional field correlation (fractional anisotropy, FA). The method is based on principal component decomposition of the sequential TTFields over one duty cycle. Using a realistic head model of a glioblastoma patient, we observed significant unwanted FA in many regions of the brain, which may potentially affect therapeutic efficacy. FA varied between different array layouts and indicated a different order of array performance than predicted from the field intensity. Tumor resection nullified differences in field distributions between layouts and increased FA considerably. Our results question the rationale for the use of macroscopically orthogonal array layouts to reduce field correlation and rather indicate that arrays should be placed to maximize pathology coverage and field intensity. The proposed calculation framework has several potential applications, incl. improved treatment planning, technology development, and accurate prognostication models. Future studies are required to validate the method.

Keywords: Computer modeling for treatment planning, Image-guided drug delivery
Abstract: Patients affected by glioblastomas have a very low survival rate. Emerging techniques, such as convection enhanced delivery (CED), need complex numerical models to be effective; furthermore, the estimation of the main parameters to be used to instruct constitutive laws in simulations represents a major challenge. This work proposes a new method to compute tortuosity, a key parameter for drug diffusion in fibrous tissue, starting from a model which incorporates the main white matter geometrical features. It is shown that tortuosity increases from 1.35 to 1.85 as the extracellular space width decreases. The results are in good agreement with experimental data reported in the literature.

Keywords: Computer modeling for treatment planning, Models of therapeutic devices and systems, Computer model-based assessments for regulatory submissions
Abstract: The University of Virginia/Padova Type 1 Diabetes (T1D) simulator has been widely used for testing artificial pancreas controllers, and, recently, novel insulin formulations and glucose sensors. However, a module describing the pharmacokinetics of the new long-acting insulin analogues is not available. The aim of this contribution is to reproduce multiple daily insulin injection (MDI) therapy, with insulin glargine 100 U/mL (Gla-100) as basal insulin, using the T1D simulator. This was achieved by developing a model of Gla-100 and by incorporating it into the simulator. The methodology described here can be extended to other insulins, allowing an extensive in silico testing of different long-acting insulin analogues under various settings before starting human trials.

Keywords: Computer modeling for treatment planning, Health technology - Verification and validation, Diagnostic devices - Physiological monitoring
Abstract: Pediatric speech sound disorders (SSD) encompass a wide range of speech production deficits that can interfere with children’s educational growth, social engagement and employment opportunities. Early detection of SSDs can facilitate timely intervention and minimize the potential for life-long adverse effects, but distinguishing between typical and atypical speech production in preschoolers is challenging due to developmental and individual variability in speech acquisition. In this study we apply Gaussian Mixture Models to speech samples from 3- to 6-year-old children, recorded by parents using an iOS app. Speech-language pathologists previously classified the samples as positive (’at risk’ speech, warranting a referral for a speech-language evaluation) or negative (’no risk’ speech). In a series of exploratory analyses, novel distance measures and group scoring techniques are developed which show good subject-level prediction accuracy. Our results provide evidence that it may be feasible to use Speech Processing and Speaker Verification techniques to model and screen speech samples from children for possible speech sound disorders.

Keywords: Models of therapeutic devices and systems, Clinical engineering, Computer modeling for treatment planning
Abstract: Design of an upper-arm exoskeleton requires knowledge of human operational ranges and workspace distributions. Motion capture recordings of right-arm motion during common tasks, known as activities of daily living (ADLs), are taken to represent a plausible workspace for an exoskeleton. An inverse kinematic model of BLUE SABINO (BiLateral Upper-extremity Exoskeleton for Simultaneous Assessment of Biomechanical and Neuromuscular Output), driven by ADL data is established to map right-arm joint locations to exoskeleton motor joint space. A kinematic representation of a human right-arm driven by ADL data is implemented via a vector analysis utilizing quaternion rotation/translation and used to visualize ADL recordings. A model of the BLUE SABINO exoskeleton whose motion is driven by the mapped motor joint-space data is used to validate the mapping graphically. The available ADL database is mapped to motor joint space. Motor position distributions are generated from the resulting dataset and estimates of robot range of motion, (ROM) and statistics for shoulder motor positions are established. A kinematically and inertially accurate model of the BLUE SABINO is developed by exporting SolidWorks® part models into SimScape Multibody (MathWorks™). The model is used to produce operational torque estimates for shoulder motors. Initial simulations indicate that the motors of interest have been properly sized.

Keywords: General and theoretical informatics - Supervised learning method, General and theoretical informatics - Computational disease profiling, General and theoretical informatics - Predictive analytics
Abstract: Rapid eye movement (REM) sleep behavior disorder is considered the prodromal stage of alpha-synucleinopathies. Its diagnosis requires careful detection of REM sleep and the gold standard manual sleep staging is inconsistent and expensive. This work proposes a new automatic sleep staging model to add robust automation to such applications, using only electroencephalography (EEG) and electrooculography (EOG) recordings. The publicly available ISRUC-Sleep database was used to optimize the design of the proposed model. The model was trained and tested on subgroup-I consisting of 100 subjects with evidence of having different sleep disorders and the polysomnographic data were manually scored by two individual experts. We divided the EOG and EEG recordings in overlapping moving 33-s epochs with step of 3s and for each of them we computed several time and frequency-domain features. The features were used to train a random forest classifier that was able to label each 33-s epoch with the probabilities of being wakefulness, REM and non-REM. The mean of the probability values of ten 33-s epochs were calculated, and the sleep stage with the highest probability was chosen to classify a 30-s epoch and matched with the manual staged hypnogram. The performance of the model was tested using 20-fold cross validation scheme. When the epochs where the scorers agreed were used, the classification achieved an overall accuracy of 92.6% and a Cohen's kappa of 0.856. Future validation on RBD patients is needed, but these performances are promising as first step of development of an automated diagnosis of RBD.

Keywords: General and theoretical informatics - Machine learning, General and theoretical informatics - Data mining, Health Informatics - Decision support methods and systems
Abstract: There is considerable interest in developing methods for early detection of medical advances which are likely to lead to effective treatments. Because the number of potentially significant discoveries published each year is very large, practically-useful detection must be accomplished using computational models. This paper considers a key component of this detection process – early identification of patentable innovations – and presents a novel machine learning-based prediction model to perform the identification. Experiments demonstrate the efficacy of the proposed approach.

Keywords: General and theoretical informatics - Machine learning, General and theoretical informatics - Artificial Intelligence
Abstract: Medication dosing in a critical care environment is a complex task that involves close monitoring of relevant biomarkers and sequential adjustments. Misdosing of medications with narrow therapeutic windows may result in preventable adverse events, negatively influencing quality and cost of care. Therefore, a robust recommendation and surveillance system can be helpful to clinicians by providing dosing suggestions or corrections to existing protocols. We present a clinician-in-the-loop framework for dose adjustment with deep reinforcement learning algorithm. There are two main objectives in this work, the first one is providing continuous dosing suggestions based on the multi-dimensional features of patients. The second one is evaluating the individualized dosing, in the presence of confounding factors. The data used in the experiments included the publicly available MIMIC-II database and Emory hospital intensive care unit. There are two important processes with respect to our objectives. In the training process, the system learned from the actual dosing executed by clinicians in the ICU. In the evaluating process, we compared the therapeutic effect among different treatments and focused on the causality between variables and outcomes. The experimental results suggested that given the states of patients, our medication dosing support system is able to provide a reasonable recommendation.

Keywords: General and theoretical informatics - Machine learning, Health Informatics - Outcome research, Health Informatics - Electronic health records
Abstract: Unplanned readmissions to the ICU contribute to high health care costs and poor patient outcomes. 6-7% of all ICU cases see a readmission within 72 hours. Machine learning models on electronic health record data can help identify these cases, providing more information about short and long term risks to clinicians at the time of ICU discharge. While time-to-event models have been used in clinical care, models that identify risks over time using higher-dimensional, non-linear machine learning models need to be developed to present changes in risk while using additionally available data. This work identifies risks of ICU readmissions at 24 hours, 72 hours, 7 days, 30 days, and bounceback readmissions in the same hospital admission with an AUROC of 0.76 (72 hours) and 0.84 (bounceback).

Keywords: General and theoretical informatics - Machine learning, Public Health Informatics - Health risk evaluation and modeling, Public Health Informatics - Public health management solutions
Abstract: Combat veterans; especially those with mental health conditions are an at risk group for suicidal ideation and behaviour. This study attempts to use machine learning algorithm to predict suicidal ideation (SI) in a treatment seeking veteran population. Questionnaire data from 738 patients consisting of veterans, still serving members of the Canadian Forces (CF) and Royal Canadian Mountain Police (RCMP) were examined to determine the likelihood of suicide ideation and to identify key variables for tracking the risk of suicide. Unlike conventional approaches we use pattern recognition methods, known collectively as machine learning (ML), to examine multivariate data and identify patterns associate with suicidal ideation. Our findings show that accurate prediction of SI of over 84.4% can be obtained with 25 variables, and 81% using as little as 10 variables primarily obtained from the patient health questionnaire (PHQ). Surprisingly the best identifiers for SI did not come from occupational experiences but rather the patient quality of health, signifying that these findings could be applied to the general population. Our results suggest that ML could assist clinicians to develop a better screening aid for suicidal ideation and behaviour.

Keywords: General and theoretical informatics - Machine learning, General and theoretical informatics - Statistical data analysis, General and theoretical informatics - Pattern recognition
Abstract: Extremely preterm infants often require endotracheal intubation and mechanical ventilation during the first days of life. Due to the detrimental effects of prolonged invasive mechanical ventilation (IMV), clinicians aim to extubate infants as soon as they deem them ready. Unfortunately, existing strategies for prediction of extubation readiness vary across clinicians and institutions, and lead to high reintubation rates. We present an approach using Random Forest classifiers for the analysis of cardiorespiratory variability to predict extubation readiness. We address the issue of data imbalance by employing random undersampling of majority samples before training each Decision Tree in a bag. By incorporating clinical domain knowledge, we further demonstrate that our classifier could have identified 71% of infants who failed extubation, while maintaining a success detection rate of 78%.

Keywords: Wearable sensor systems - User centered design and applications, Modeling and analysis, Novel methods
Abstract: Poor medication adherence threatens an individual's health and is responsible for substantial medical costs in the United States annually. In order to improve medication adherence rates and provide timely reminders, we developed a smartwatch application that collects data from embedded inertial sensors, which include an accelerometer and gyroscope, to monitor a series of actions happening during an individual's medication intake. After the collected data was delivered to a server, Apache Spark was used to distribute the data and apply machine learning algorithms in order to predict several discrete actions including medication intake. By utilizing these tools, we were able to preprocess high frequency sensor data and apply a random forest algorithm, yielding high frequency and recall of the aforementioned actions.

Keywords: Sensor systems and Instrumentation, Physiological monitoring - Instrumentation, Physiological monitoring - Novel methods
Abstract: This paper presents preliminary results on the effect of contact pressure on bioimpedance measurements in an excised section of human colon tissue. The impedance measurements were performed with a small diameter probe suitable for in-vivo use, which is capable of measuring contact force. Force measurements are performed by fiber optic sensor which consisted of a Fiber Bragg Grating. The obtained results highlight the importance on limiting the applied pressure during bioimpedance measurements.

Keywords: Wearable body sensor networks and telemetric systems, Wearable sensor systems - User centered design and applications, Sensor systems and Instrumentation
Abstract: Virtual Reality (VR) provides a highly immersive medium for assessment of cognitive capacity.In this paper we use this to examine the feasibility of an arm muscle-motion based (EMG) interaction technique for interacting with a VR cognitive performance diagnostic and training environment. Therefore, we compared the state-of-the-art controller input to our EMG-based approach regarding presence and user experience. Results show significant differences in terms of textit{Novelty} and textit{Dependability}.Since there are only minor differences in terms of presence and user experience, the advantages of using more demanding physical interactions and hence reinforcing the connection between decision making and action execution, the developed interaction approach seems to be a promising method with a high potential having a positive effect on the cognitive training progress.

Keywords: Wearable body-compliant, flexible and printed electronics, Bio-electric sensors - Sensing methods, Wearable wireless sensors, motes and systems
Abstract: We have recently developed a conformable solid state material solution (carbon nanofiber filled polydimethylsilisoxane, CNF-PDMS) for electroencephalography (EEG) electrodes. In this study, we tested the efficacy of electrodes molded from this material to record well studied neural phenomena using a battery of standard laboratory tasks. Event related potential (ERP) and eyes open/closed results show performance matching that of commercially available metal-pin based dry EEG electrode, while summary statistics (correlation and RMSE) show matched and even improved ability to track local and global fluctuations in EEG. We present baseline data that demonstrates CNF-PDMS is a viable solution for conformable, safe, dry EEG electrodes.

Keywords: Wearable low power, wireless sensing methods, Physiological monitoring - Novel methods, New sensing techniques
Abstract: Pulse arrival time (PAT) and pulse transit time (PTT) derived from the finger have been widely investigated for noninvasive blood pressure (BP) measurement. The study investigates the feasibility of BP measurement using a chest-worn patch sensor derived systolic timing intervals and pulse timing measurements. Healthy volunteers (N=14, 38±13 years) carried out a protocol including deep breathing test, sustained hand grip test and modified Valsalva test with continuous physiological measurements from a patch sensor attached on left chest and intermittent BP measurements from an automated oscillometric monitor as a reference. The efficacy of chest derived PAT and PTT for univariate BP prediction is assessed using correlation and regression slope. The cross validation performance of predicting BP using multivariate regression model with chest derived systolic timing intervals and pulse timing features were also evaluated. The results suggest that the chest derived PAT and PTT had modest correlations (−0.52 and −0.31) and regression slopes (−0.21 and −0.14) with automated oscillometric systolic and diastolic BP references, respectively. On the other hand, a multivariate regression approach for prediction of mean blood pressure (MBP) using patch sensor measurements showed a correlation of 0.72, mean error of 0.1 mmHg and RMSE error of 5.1 mmHg compared to the oscillometric MBP values. The study demonstrated the feasibility of BP measurement using a wearable chest-worn patch sensor in healthy control subjects.

Keywords: Bio-electric sensors - Sensing methods, Sensor systems and Instrumentation
Abstract: Recent advancement in technology has brought about increase in the application areas of wearable electroencephalographic devices. In that, new types of electrodes take place, and particular attention is needed to ensure the required quality of obtained signals. In this study, we evaluate electrode-skin impedance and signal quality for several kinds of electrodes when used in conditions typical for wearable devices. Results suggest that active dry electrode coated with gold alloy is superior while it was challenging to obtain appropriate signal quality when using passive dry electrodes. We also demonstrate electrode-skin impedance measurement using the analog frontend ADS1299, which is suitable for implementation in wearable devices

Keywords: Point of care - Detection and monitoring, Point of care - Technologies in resource limited settings, Point of care - Home-based applications
Abstract: The Smartphone-based Compression-induced Scope (SCIS) is a mobile device designed to sense the mechanical properties of tumors. Here, an SCIS system with an infrared temperature (SCIS-T) sensor is developed. The color and texture information of target skin are extracted from the SCIS-T images using a color-based edge detection technique and a texture filter. This new system provides mechanical properties (size, elasticity) of the inclusion as well as the skin surface (color, temperature, texture) characteristics. The application of this system is in the identification of inflammatory breast cancer, which is characterized by color, texture, and temperature change. The device is tested using chicken breast phantoms with embedded silicone inclusion.

Keywords: Point of care - Detection and monitoring, Empowering individual healthcare decisions through technology, Point of care - Technologies in resource limited settings
Abstract: Refractive errors are the most common visual defects in humans. They are corrected using lenses whose power is determined using expensive and bulky devices operated by trained professionals. This limits the outreach of eye-health care. We exploit commercial virtual reality (VR) setup to create a portable and inexpensive system for subjective estimation of spherical refractive errors. In doing so, we aim to keep hardware additions simple and to a minimum. We add a plain reflecting mirror in a VR headset to project optotypes on programmable focal planes at varying distances from the subject's eye. An interactive interface uses feedback from the user to estimate accommodation range and spherical refractive errors automatically. We compute the range and precision of our system, and validate them in a user trial study. The proposed setup strongly agrees with clinical subjective refraction.

Keywords: Empowering individual healthcare decisions through technology, Point of care - Detection and monitoring, Medical technology - Design and development
Abstract: We present a practical electronic nose (e-nose) system, NOS.E, for the rapid detection and identification of human health conditions. By detecting the changes in the composition of an individual's respiratory gases, which have been shown to be linked to changes in metabolism, e-nose systems can be used to characterize the physical health condition. We demonstrated our system's viability with a simple data set consists of breath collected under three different scenarios from one volunteer. Our preliminary results show the popular classifier SVM can discriminate NOS.E's responses under the three scenarios with high performance. In future work, we will aim to gather a more varied data set to test NOS.E's abilities.

Keywords: Point of care - Detection and monitoring, Point of care - Innovations, Precision medicine
Abstract: The presence of abnormal amounts of bilirubin in the blood stream and skin, usually referred to as hyperbilirubinemia, is associated with a wide range of pathologies that can pose considerable risks for human health. The early and effective screening of the severity degrees of this medical condition can play an important role on the selection of the appropriate treatment for the associated pathologies. This, in turn, can minimize the need for more aggressive and costly therapeutic interventions which can themselves pose considerable risks for morbidity and mortality. The current noninvasive protocols used to differentiate these severity degrees, however, are hindered by the relatively limited knowledge about the impact of different amounts of extravascular bilirubin on skin spectral responses and on the onset of jaundice, the resulting yellow-tinted skin appearance. In this paper, we address this open problem through controlled in silico experiments supported by measured data provided in the related literature. Our experimental findings bring biophysically-based insights to bear on the clarification of this biomedical entanglement, and unveil optical features that can potentially lead to more effective screening protocols for the noninvasive differentiation and monitoring of variable degrees of hyperbilirubinemia severity.

Keywords: Point of care - Detection and monitoring, Point of care - Home-based applications, Empowering individual healthcare decisions through technology
Abstract: Reliably detecting emotions is a topic of current research in understanding mental health. Among the many modes of detecting emotion, audio has a prominent place. In this paper, we propose a two-level, multi-way classifier applied to classification of seven emotions from the standard Emo-DB database. It is an automated methodology of analyzing a confusion matrix of a first-level classifier to build more classifiers at successive levels. A random forest classifier is used on state-of-the-art features for analyzing affective speech. The confusion matrix from this classification level is analyzed to decide, for each class, which other classes are most confused by using a threshold on the misclassification rate. For the chosen pairs, second level classifiers are built and trained on the same data. Its performance on the training-set (73.3%) as well as a non-intersecting training set (72.9%) are both better than state-of-the-art performance. We initiate a possible explanation of the performance improvement by considering the confusion among emotions placed on Russel’s circumplex model.

Keywords: Point of care - Detection and monitoring, Medical technology - Innovation, Point of care - Innovations
Abstract: Bedridden patients always need more attention in order to prevent unexpected falling, bedsores and other dangerous situations in daily care. This work proposes a data-driven classification method to recognize different bed-related human activities including exiting the bed, turning over, stretching out for something on the bedside table, sitting up and lying down by analyzing the real-time signals that are acquired from four load cells installed around the hospital bed. Considering the dynamic characteristics of the signals, dynamic principal component analysis (DPCA), here serving as a pre-processing step, is firstly utilized to extract both static and dynamic relations from the variables. Then the final statistical model for each class is established by Gaussian mixture model (GMM) with Figueiredo-Jain algorithm that can optimally select the number of components. An alarm will be triggered when a noteworthy action is detected. The proposed method has achieved superior performance using the experimental data from 10 adult volunteers. The results move a step forward towards the design of an intelligent hospital bed for practical applications.

Keywords: Sensor Informatics - Low power, wireless sensing methods and systems, Sensor Informatics - Smart home technology, Health Informatics - Emerging IT for efficient/low-cost healthcare delivery
Abstract: This paper describes a method for the detection of changes in sleeping conditions using multipoint ambient sensing for safety and amenity in the sleep environment. This method continuously detects changes during sleep, such as the movements of the person and bedding based on the measurement of acceleration, temperature and humidity of the comforter. Through the measurement of these ambient conditions, this system improves sleep quality and prevents accidents, such as falling off the bed. In this study, we build a basic system using internet of things (IoT) devices and performance verification. The experimental results show the feasibility of the proposed method.

Keywords: Sensor Informatics - Sensor-based mHealth applications, Sensor Informatics - Physiological monitoring, Health Informatics - Mobile health
Abstract: Cuff less Blood Pressure (BP) monitoring has gained interest of the research community in recent years, due to its importance in continuous and non-invasive monitoring of BP for early detection of hypertension, thereby reducing mortality. Several approaches that involve photoplethysmography (PPG) and Pulse Transit Time (PTT) have been explored with promising results; however the requirement of two sensors makes them obtrusive for continuous use. Single PPG sensor approaches using machine learning have also been attempted, but there are certain deficiencies in these methods as they go for a one-size-fits-all approach. In this work, we develop an ensemble of BP prediction models based on demographic and physiological partitioning. Also, we incorporate a set of unique PPG features into our models, which results in test accuracies of 5 mmHg Mean Absolute Error (MAE) for Diastolic BP, and 6.9 mmHg MAE for Systolic BP. Given our marked improvement over ubiquitous models (18% for Diastolic BP and 11.5% for Systolic BP), this approach opens up avenues where single PPG sensor based methods can predict BP with a high degree of accuracy. This is a big step towards developing continuous BP monitoring systems, and can help in better management of cardiac health.

Keywords: Sensor Informatics - Sensors and sensor systems, Imaging Informatics - Image analysis, processing and classification, Sensor Informatics - Intelligent medical devices and sensors
Abstract: In recent years, as a way to prevent patient fall-down, the study has been conducted to detect patient behavior that is leaving from a bed by a patient room camera. It is very important to specify a patient bed location for process of detecting patient behavior by camera images. In this study, we propose a method to specify the patient bed location using a monocular camera. In this paper, we convert a camera image view point as preprocessing to produce a bird’s eye view image. By using planer perspective transformation, it is possible to display the bed as a rectangular shape with fixed ratio even the bed location or the camera position was changed, therefore it is possible to detect the bed location with a high degree of accuracy by means of machine learning. In the simulation experiment results, we confirmed the average error of the bed coordinates was 7.9 pixels and the standard deviation was 5.0 pixels, and in practical scene, we confirmed the average error of the bed coordinates is 12.1 pixels and the standard deviation is 8.2 pixels.

Keywords: Health Informatics - Technology and services for assisted-living and elderly, Sensor Informatics - Physiological monitoring, Imaging Informatics - Image analysis, processing and classification
Abstract: This paper presents a physiological monitoring system for assistive robots using a thermal camera. It is based on the detection of subtle changes in temperature observed on different parts of the face. First, we segment and estimate these face regions on thermal images. Then, by applying Fourier analysis on temperature data, we estimate respiration and heartbeat rate. This physiological monitoring system has been integrated in an assistive robot for elderly people at home, as part of the ENRICHME project. Its performance has been evaluated on a new thermal dataset for physiological monitoring, which is made publicly available for research purposes.

Keywords: General and theoretical informatics - Pattern recognition, Sensor Informatics - Sensors and sensor systems, General and theoretical informatics - Machine learning
Abstract: In this new research, we expand on our previous system for vocal fatigue detection by adding five new features in the classifier. We also perform further testing on 37 test subjects. The goals were: 1) to classify subjects performing normal versus simulated pressed vocal gestures; 2) to distinguish vocally healthy from vocally fatigued subjects as determined by VFI score on factor 1; and 3) to determine the validity of the labels vis-a-vis the choice of this same VFI-factor-1 boundary. As the results demonstrated, the choice of classifier and the new features were quite appropriate, while there is margin for better choices of the VFI-factor-1 boundary.

Keywords: Sensor Informatics - Data inference, mining, and trend analysis, Sensor Informatics - Intelligent medical devices and sensors, Sensor Informatics - Physiological monitoring
Abstract: The fluctuation of an RR interval (RRI) on an electrocardiogram (ECG) is called heart rate variability (HRV). HRV reflects the autonomic nerve activity, thus HRV analysis has been used for health monitoring such as stress estimation, drowsiness detection, epileptic seizure prediction, and cardiovascular disease diagnosis. However, RRI and HRV features are easily affected by arrhythmia, which deteriorates the health monitoring performance. Ventricular contraction (PVC) is common arrhythmia that many healthy persons have. Thus, a new methodology for dealing with RRI fluctuation disturbed by PVC needs to be developed for realizing precise health monitoring. To modify RRI data affected by PVC, the present work proposes a new method based on a denoising autoencoder (DAE), which reconstructs original input data from the noisy input data by using a neural network. The proposed method, referred to as DAE-based RRI modification (DAE-RM), aims to correct the disturbed RRI data by regarding PVC as artifacts. The present work demonstrated the usefulness of the proposed DAE-RM through its application to real RRI data with artificial PVC (PVC-RRI). The result showed that DAE-RM successfully modified PVC-RRI data. In fact, the root mean squared error (RMSE) of the modified RRI was improved by 83.5% from the PVC-RRI. The proposed DAE-RM will contribute to realizing precise HRV-based health monitoring in the future.

Keywords: Systems biology and systems medicine - Modeling of signaling networks, Systems biology and systems medicine - Modeling of biomolecular system dynamics, Systems biology and systems medicine - Modeling of biomolecular system pathways
Abstract: Classically, the Wnt/ß-catenin and Notch/Delta signaling pathways were thought to operate through separate mechanisms, performing distinct roles in tissue patterning. However, it has been shown that ß-catenin activates transcription of Hes1, a signaling intermediate in the Notch/Delta pathway that controls its lateral inhibition mechanism. To investigate this non-canonical crosstalk mechanism, a new gene circuit, integrating the two pathways, is proposed and simulated in two-cell and multi-cell environments. This model also captures both Paneth cell-mediated and mesenchymal Wnt production. The simulations verify that the gene circuit is temporally bistable and capable of forming a pattern on a multi-cell grid. Last, the model exhibits a bifurcation based on the steady state concentration of Wnt and the relative amount of control ß-catenin has over the Hes1 promoter, providing a possible mechanism to explain why a homogeneous population of transit amplifying cells is observed directly above the more diverse stem niche.

Keywords: Systems biology and systems medicine - Modeling of gene/epigene regulatory networks, Systems biology and systems medicine - Modeling of biomolecular system dynamics
Abstract: High risk HPV can induce cervical and oropharyngeal cancerous lesions. The initial phase of the infection is characterized by a fine regulation of the viral DNA replication, in order to maintain 10-100 DNA copies per cell. Such regulation is primarily controlled by E1 and E2 proteins produced by the early promoter. The recently discovered E8 promoter is capable to co-regulate the early one in order to maintain a low and constant viral DNA copy number. The aim of this study is to develop a novel stochastic mathematical model of the co-regulation between the E8 and the early promoter, with the main purpose to rigorously show the E8 promoter capability to finely regulate the HPV transcripts which control the DNA replication in the first stages of the infection. The model, condensing the biological knowledge present in literature, describes the interaction between the two promoters and shows how the E8 co-regulation is capable to reject the stochastic noise of E2 gene expression to a higher extent than the early promoter negative auto-feedback. This proves the capability of the E8 promoter to finely control the HPV genomes copy number.

Keywords: Systems biology and systems medicine - Modeling of signaling networks, Computational modeling - Biological networks, Systems biology and systems medicine - Modeling of biomolecular system pathways
Abstract: The number of published results in biology and medicine is growing at an exceeding rate, and thus, extracting relevant information for building useful models is becoming very laborious. Furthermore, with the newly published information, previously built models need to be extended and updated, and with the voluminous literature, it is necessary to automate the model extension process. In this work, we introduce a methodology for extending logical models of cell signaling networks using a Genetic Algorithm (GA). The proposed procedure is developed to optimally search for a subset of biological interactions that extend logical models while preserving their desired behavior. To evaluate the effectiveness of the proposed methodology, we randomly removed a subset of elements from an existing T cell differentiation model, and mixed them with randomly created interactions to mimic the output of literature reading. We then used the GA to search for the extensions that optimally reconstructed the model. The simulation results showed that the GA was able to find a set of extensions that preserved the desired behavior of the model with fewer elements than the original model. The results demonstrate that the GA is an efficient tool for model extension, and suggest that it can be used for model reduction as well.

Keywords: Computational modeling - Biological networks, Systems biology and systems medicine - prediction of disease related regulator, Systems biology and systems medicine - RNA-seq analysis
Abstract: Identifying the gene regulatory networks that control development and disease is one of the most important problems in biology. Here, we introduce a computational approach, called PIPER (ProgressIve network PERturbation), to identify genes that drive differences in the gene regulatory network across different points in a biological progression. PIPER employs algorithms tailor-made for single cell RNA sequencing (scRNA-seq) data to jointly identify gene networks for multiple progressive conditions. It then performs differential network analysis along the identified gene networks to determine master regulators. We demonstrate that PIPER outperforms state-of-the-art alternative methods on simulated data and is able to predict key regulators of differentiation on real scRNA-Seq datasets.

Keywords: Computational modeling - Biological networks
Abstract: Identifying protein complexes within a protein-protein interaction (PPI) networks is a crucial task in computational biology that helps to facilitate a better understanding of the cellular mechanisms it is possible to observe in various organisms. Datasets of predicted PPIs have been determined using high-throughput experimental technology. However, the datasets typically contain many spurious interactions. It is essential that these interactions, observed in the given datasets, are validated before they are employed to predict protein complexes. This paper describes the identification of missing interactome links in the PPI network as a way of improving the detection of protein complexes. The missing links have been identified by extracting several topological features. These are subsequently employed in conjunction with a two-class boosted decision-tree classifier to develop a machine-learning model that is capable of distinguishing between existing and non-existing interactome links. The model was trained on a PPI network that consisted of 1,622 proteins and 9,074 interactions, then tested on another PPI network that consisted of 1,430 proteins and 6,531 interactions. All 6,531 interactions were identified with a precision of 0.994 and a recall of 1. The model was also able to detect 37 novel interactions that were then validated using a STRING database of known and predicted PPIs. The detection of the protein complexes using ClusterONE was improved by the inclusion of the 37 novel interactions.

Keywords: Data-driven modeling, Synthetic biology, Systems modeling - Clinical applications of biological networks
Abstract: Calcium spiking can be used for drug screening studies in pharmaceutical industries. However, performing experiments for multiple drugs and doses are highly expensive. The oscillatory behavior of calcium spiking data demonstrates extreme nonlinearity and phase singularity. This makes it more challenging to construct physics-based models for the experimental observations. In this scenario, data based modelling, such as Artificial Neural Networks (ANN), and thereafter the model based prediction of calcium profiles may offer a cost–effective and time saving solution. Therefore, a novel ANN building algorithm is presented in the current work, where data based simultaneous estimation of ANN architecture and nonlinear activation function stands out as the main highlight. The resultant ANN was then used to learn the oscillatory behavior in calcium ion concentration data, obtained from hippocampal neurons of rats by fluorescent labelling and confocal imaging. The paper shows that the novel technique can be used in general for emulating biochemical oscillations (with or without drug injection) and can be implemented to predict the cell-drug responses for intermediated doses. The proposed algorithm can also be used for obtaining high resolution data from low resolution experimental measurements.

Keywords: Neural interfaces - Bioelectric sensors, Neural interfaces - Microelectrode technology, Neural interfaces - Implantable systems
Abstract: The high complexity of the biological response to implanted materials builds a serious barrier against implanted recording and stimulation electrode arrays to succeed in clinically relevant chronic studies. Some of the cell and molecular interactions and their contribution to inflammation and device failure are still unclear. The interrelated mechanisms leading to tissue damage and electrode array failure during simultaneous faradaic, elec-trochemical reactions and biological response under electrical stimulation are not understood sufficiently. One variable, with which inflammatory and electrode surface processes can be analyzed and assessed, is the pH change in the immediate envi-ronment of the material-tissue interface. Here, the greatest chal-lenges are in the biocompatibility and in-vivo long-term stability of selected sensor materials, the measurement of small transient pH oscillations and positioning of the sensor at a defined and nearest possible distance in the micrometer range, to the site of activity without the pH sensing being affected by the material-tissue interactions itself. This work represents the in-situ meas-urement of local and transient pH changes at a pulsed electrode with an embedded in-vivo compatible pH sensor and therein differentiating from current approaches of pH sensing during electrical stimulation.

Keywords: Neural interfaces - Body interfaces, Neurological disorders, Motor neuroprostheses
Abstract: Spinal cord injury (SCI), stroke and other nervous system conditions can result in partial or total paralysis of in- dividual’s limbs. Numerous technologies have been proposed to assist neurorehabilitation or movement restoration, e.g. robotics or neuroprosthesis. However, individuals with tetraplegia often find difficult to pilot these devices. We developed a system based on a single inertial measurement unit located on the upper limb that is able to classify performed movements using principal component analysis. We analyzed three calibration algorithms: unsupervised learning, supervised learning and adaptive learning. Eight participants with tetraplegia (C4- C7) piloted three different postures in a robotic hand. We achieved 89% accuracy using the supervised learning algorithm. Through offline simulation, we found accuracies of 76% on the unsupervised learning, and 88% on the adaptive one.

Keywords: Neural interfaces - Bioelectric sensors, Neural interfaces - Microelectrode technology, Neural interfaces - Tissue-electrode interface
Abstract: Micro-electrocorticography (µECoG) is a minimally invasive neural interface that allows for recording from the surface of the brain with high spatial and temporal resolution. However, discerning multi-unit and local field potential (LFP) activity with potentially highly-correlated signals across a dense µECoG array can be challenging. Here we describe a novel µECoG design to compare the effect of referencing recordings to a local reference electrode and common average referencing (CAR). The filtering effect and the significant increase in evoked signal to noise ratio (ESNR) can be seen after re-referencing for both types of referencing. In a preliminary analysis, re-referencing the µECoG signals can increase recording performance at high contact densities in the auditory cortex. This also provides promising evidence for a versatile in-house fabricated µECoG electrode.

Keywords: Neural interfaces - Microelectrode technology, Neural interfaces - Implantable systems, Neural interfaces - Tissue-electrode interface
Abstract: A reactive-accelerated-aging (RAA) soak-test has been employed to challenge microfabricated neural interface devices against an aggressive environment that mimics worst-case chronic physiological inflammation. The soak tests were able to determine the ability of different thin-film materials to increase the adhesive strength between the polyimide and platinum-gold-platinum metallization layers in the interface. It was found that a 3-day soaking at 87 ˚C in phosphate buffered with 10 to 20 mM hydrogen peroxide resulted in adhesion failure of the metal-polyimide interface when titanium was used as the primary adhesion layer. The addition of hydrogenated amorphous silicon carbon was able to eliminate the onset of adhesion failure of the metal-polyimide interface for soak tests lasting 7 days. However, residual-film stress resulted in cracking of the silicon carbide layer. These tests have demonstrated the ability of RAA soak tests to provide rapid in vitro assessment of microfabricated neural interfaces and thereby reduce the time needed to develop a process to produce chronically reliable devices.

Keywords: Neural interfaces - Tissue-electrode interface, Neural interfaces - Cellular, Neural interfaces - Microelectrode technology
Abstract: Neuroblastoma cells are often used as a cell model to study Parkinson’s disease, which causes reduced dopamine release in substantia nigra, the midbrain that controls movements. In this paper, we developed a 1024-ch monolithic CMOS sensor array that has the spatiotemporal resolution as well as low-noise performance to monitor single vesicle release of dopamine from neuroblastoma cells. The CMOS device integrates 1024 on-chip electrodes with an individual size of 15 µm × 15 µm and 1024 transimpedance amplifiers for each electrode, which are each capable of measuring sub-pA current. Thus, this device can be used to study the detailed molecular dynamics of dopamine secretion at single vesicle resolution.

Keywords: Neural interfaces - Microelectrode technology, Neural interfaces - Implantable systems, Brain-computer/machine interface
Abstract: Development of micro electrode arrays for neural recording is an active field that thrives on novel materials and fabrication techniques offered by micro fabrication technology. The material and mechanical properties of microelectrode arrays have a critical role on the quality and longevity of neural signals. In this study, carbon fiber microelectrode (CFME) bundles were developed and implanted in the spinal cord of experimental animals for in vivo recording. Neural data analysis revealed that single spikes could successfully be recorded and sorted. Removal of approximately 75 µm of the parylene-C coating at the tips of the fibers increased the signal-to-noise ratio. Connecting multiple (three) carbon fiber filaments to the same recording channel did not deteriorate the signal quality compared to that of undesheathed fibers. Immunohistochemistry showed that electrode tips were splayed in tissue after implantation and CF bundles had a small footprint with mild encapsulation around them. These results are very promising for the use of CFME bundles for recordings of spinal cord signals in behaving animals.

Keywords: Signal pattern classification, Adaptive filtering, Data mining and processing in biosignals
Abstract: Functional near-infrared spectroscopy (fNIRS) is a non-invasive multi-channel imaging tool for assessing brain activities, which has shown its high potential in brain-computer interface (BCI) technique. Most previous researches have focused on constructing high dimensional features from whole channels, adding to the complexity of their classifiers. Another multi-channel source for BCI is electroencephalograph (EEG), which possesses different spatial and temporal features from fNIRS. In EEG field, Common Spatial Pattern (CSP) algorithm is widely used aimed at dimensionality reduction. In our article, we modified it based on the characteristics of fNIRS and evaluated its effectiveness in discriminating Mental Arithmetic (MA) against resting status in an open-access dataset. The Modified Common Spatial Pattern algorithm significantly outperforms CSP algorithm in fNIRS-based BCI and shows its potential in further BCI related explorations.

Keywords: Physiological systems modeling - Signal processing in simulation, Physiological systems modeling - Signals and systems, Physiological systems modeling - Signal processing in physiological systems
Abstract: The vestibule-ocular reflex (VOR) has been one of the most popular model systems to investigate the role of the cerebellum in adaptive motor control. VOR motor learning can be experimentally induced by continuous application of combination of head rotating stimulus and optokinetic stimulus. For instance, in phase application of those stimuli decreases VOR gain defined by eye velocity of VOR in the dark divided by head velocity, while out of phase of those increases VOR gain. It has been known that VOR gain is modifiable context-dependently. Namely, VOR gains for leftward and rightward head rotations can be respectively increased and decreased simultaneously. The cerebellar signal processing underlying the context dependent VOR motor learning, however, is not fully uncovered. In the present study, we simulated direction selective VOR motor learning, using the artificial cerebellar neuronal network model that we developed to understand the origin of the cerebellar motor learning.

Keywords: Neural networks and support vector machines in biosignal processing and classification, Adaptive filtering, Physiological systems modeling - Multivariate signal processing
Abstract: Common Spectral Pattern (CSP) algorithm re-mains predominant for feature extraction from multichannel EEG motor imagery data. However, multiclass classification of from this featureset has been a challenging job. Different approaches have been proposed to be applied on featureset of different EEG subbands to achieve significant classification accuracy. Ensemble learning is very effective in this context to achieve higher accuracy in Brain-Computer Interface (BCI) domain. In this study, we have proposed enhanced classification algorithms to achieve higher classification accuracies. The methods were evaluated against the motor imagery data from Dataset 2a of the publicly available BCI Competition IV (2008). This dataset consists of 22 channels EEG data of 9 subjects for four different movements. A tree based ensemble approachfor supervised classification, Extra-Trees algorithm, has been proposed in this paper and also evaluated for its efficacy on this dataset to classify between left hand and right hand movement imaginations. Moreover, this classifier has its inherent capability to select optimum features. Furthermore, in this paper an extension of the binary classification into multiclass domain is also implemented with error correcting output codes (ECOC) approach using the same dataset. Subject-specific frequency bands α(8-12Hz) and β(12-30Hz) along with HG(70-100Hz) were considered to extract CSP features. We have achieved an individual peak accuracy of 98% and 84% in binary class and multiclass classification respectively. Furthermore, the results yielded a mean kappa value of 0.58 across all the subjects.This kappa value is higher than of the winner of competition and also from the most of the other approaches applied in this dataset.

Keywords: Neural networks and support vector machines in biosignal processing and classification, Data mining and processing - Pattern recognition, Time-frequency and time-scale analysis - Wavelets
Abstract: Motor imagery (MI) based brain-computer interface (BCI) plays a crucial role in various scenarios ranging from post-traumatic rehabilitation to control prosthetics. Computer aided interpretation of MI has augmented prior mentioned scenarios since decades but failed to address interpersonal variability. Such variability further escalates in case of multiclass MI, which is currently a common practice. The failures due to interpersonal variability can be attributed to handcrafted features as they failed to extract more generalized features. The proposed approach employs convolution neural network (CNN) based model with both filtering (through axis shuffling) and feature extraction to avail end-to-end training. Axis shuffling is performed adopted in initial blocks of the model for 1D pre processing and reduce the parameters required. Such practice has avoided the overfitting which resulted in an improved generalized model. Publicly available BCI Competition-IV 2a dataset is considered to evaluate the proposed model. The proposed model has demonstrated the capability to identify subject-specific frequency band with an average and highest accuracy of 70.5% and 83.6% respectively. Proposed CNN model can classify in real time without relying on accelerated computing device like GPU.

Keywords: Neural networks and support vector machines in biosignal processing and classification, Signal pattern classification, Data mining and processing in biosignals
Abstract: Motor imagery (MI) based Brain-Computer Interfaces (BCIs) are a viable option for giving locked-in syndrome patients independence and communicability. BCIs comprising expensive medical-grade EEG systems evaluated in carefully-controlled, artificial environments are impractical for take-home use. Previous studies evaluated low-cost systems; however, performance was suboptimal or inconclusive. Here we evaluated a low-cost EEG system, OpenBCI, in a natural environment and leveraged neurofeedback, deep learning, and wider temporal windows to improve performance. μ-rhythm data collected over the sensorimotor cortex from healthy participants performing relaxation and right-handed MI tasks were used to train a multi-layer perceptron binary classifier using deep learning. We showed that our method outperforms previous OpenBCI MI-based BCIs, thereby extending the BCI capabilities of this low-cost device.

Keywords: Connectivity measurements, Signal pattern classification, Neural networks and support vector machines in biosignal processing and classification
Abstract: Recent progress in the number of studies involving brain connectivity analysis of motor imagery (MI) tasks for brain-computer interface (BCI) systems has warranted the need for pre-processing methods. The objective of this study is to evaluate the impact of current source density (CSD) estimation from raw electroencephalogram (EEG) signals on the classification performance of scalp level brain connectivity feature based MI-BCI. In particular, time-domain partial Granger causality (PGC) method was implemented on the raw EEG signals and CSD signals of a publicly available dataset for the estimation of brain connectivity features. Moreover, pairwise binary classifications of four different MI tasks were performed in inter-session and intra-session conditions using a support vector machine classifier. The results showed that CSD provided a statistically significant increase of the AUC: 20.28% in the inter-session condition; 12.54% and 13.92% with session 01 and session 02, respectively, in the intra-session condition. These results show that pre-processing of EEG signals is crucial for single-trial connectivity features based MI-BCI systems and CSD can enhance their overall performance.

Keywords: Image reconstruction and enhancement - Compressed sensing / Sampling, Image reconstruction and enhancement - Tomographic reconstruction, CT imaging
Abstract: Classical total variation (TV) based iterative reconstruction algorithms assume that the signal is piecewise smooth, which causes reconstruction results to suffer from the over-smoothing effect. To address this problem, this work presents a novel computed tomography (CT) reconstruction method for the sparse-sampling problem called the group sparsity regularization-based simultaneous algebraic reconstruction technique (GSR-SART). Group-based sparse representation, which utilizes the concept of a group as the basic unit of sparse representation instead of a patch, is introduced as the image domain prior regularization term to eliminate the over-smoothing effect. By grouping the nonlocal patches into different clusters with similarity measured by Euclidean distance, the sparsity and nonlocal similarity in a single image are simultaneously explored. The split Bregman iteration algorithm is applied to obtain the numerical scheme. Experimental results demonstrate that our method both qualitatively and quantitatively outperforms several existing reconstruction methods, including filtered back projection, expectation maximization, SART, and TV-based projections onto convex sets.

Keywords: CT imaging applications, Image feature extraction, Image reconstruction and enhancement - Filtering
Abstract: We present the discrete version of heat kernel smoothing on graph data structure. The method is used to smooth data in an irregularly shaped domains in 3D images. New statistical properties of heat kernel smoothing are derived. As an application, we show how to filter out noisy data in the lung blood vessel trees obtained from computed tomography. The method can be further used in representing the complex vessel trees parametrically as a linear combination of basis functions and extracting the skeleton representation of the trees.

Keywords: Image reconstruction - Fast algorithms, Image visualization, Image registration, segmentation, compression and visualization - Volume rendering
Abstract: Recent researches have shown that the relation between mitochondrial function and degenerative disorders is closely related to aging, such as Alzheimer's and Parkinson's diseases. Because these studies expose the need for detailed analysis of high-resolution physical alterations in mitochondria, three dimensional (3D) visualization of mitochondria from electron microscopy (EM) images is coming into prominence. To this end, how to develop suitable segmentation algorithms and connection algorithms has attracted our attentions. Since previous algorithms have shown preferable segmentation performance on mitochondria with different shapes and sizes. In this paper, we propose to utilize the segmentation information instead of detection information in context to obtain the mitochondrial connection relation in adjacent layers. Additionally, different from previous methods, we present a novel and effective connection approach by obtaining sparse matrixes and implementing a forward connection mode. Experiments on automated tape-collecting ultramicrotome scanning electron microscopy (ATUM-SEM) stacks demonstrate that our approach can effectively handle with the case of split and merge, and achieve a comparable connection quality measured by split error and merge error.

Keywords: Iterative image reconstruction, Image reconstruction and enhancement - Compressed sensing / Sampling, Image reconstruction and enhancement - Tomographic reconstruction
Abstract: This work proposes a novel microwave imaging (MWI) multi-frequency technique, which combines compressive sensing (CS) with the well-known distorted Born iterative method (DBIM) to enhance the accuracy in the imaging procedure. CS strategies are emerging as a promising tool in MWI applications, which can also reduce the number of data samples. The proposed approach is based on an iterative shrinkage thresholding algorithm (ISTA), which has been modified to include an automatic and adaptive selection of multi-threshold values. The proposed implementation is applied in reconstruction of two-dimensional numerical heterogeneous breast phantoms, where it outer-performs the standard thresholding implementation and proves to be an interesting tool for medical imaging applications.

Keywords: Image enhancement, Regularized image Reconstruction, Ultrasound imaging - High-frequency technology
Abstract: Scanning acoustic microscopy (SAM) is an imaging modality used to obtain 2D maps of acoustical and mechanical properties of soft tissues and uses ultrasound transducers operating at very high-frequencies. Such transducers are challenging and costly to manufacture, and SAM systems at higher frequencies become more sensitive to experimental issues. Nevertheless, biomedical applications of SAM often require spatial resolutions nearly as good as light microscopy. In addition, stained histology photomicrographs of thin sections of tissues are easily obtained with the necessary resolution and accuracy. Consequently, the aim of this study is to introduce a bilateral approach that enhances the resolution of SAM images by leveraging the co-registered high-resolution histology image. We propose to use bilateral weighted total variation regularization to solve the super-resolution problem. A fast matrix-less solver is developed by utilizing the Alternating Direction Method of Multipliers (ADMM) and solving the least squares problem in one ADMM step in the Fourier domain. Reconstruction results on experimentally recorded SAM and histology data show promising improvement over the classical techniques.

Keywords: CT imaging, Image enhancement - Denoising, Image reconstruction and enhancement - Machine learning / Deep learning approaches
Abstract: Low-dose Computed Tomography (CT) is considered a solution for reducing the risk of X-ray radiation; however, lowering the X-ray current results in a degraded reconstructed image. To improve the quality of the image, different noise removal techniques have been proposed. Convolutional neural networks also have shown promising results in denoising the low-dose CT images. In this paper, a deep residual network with dilated convolution is proposed. The identity mappings pass the signal to the higher layers and improve the performance of the network and its training time. Moreover, employing dilated convolution helps to increase the receptive field faster. Dilated convolution makes it possible to achieve good results with fewer layers and less computational costs. The proposed network learns end to end mapping from low-dose to normal-dose CT images.

Keywords: Bio-electric sensors - Sensor systems, New sensing techniques, Sensor systems and Instrumentation
Abstract: A prototype 24 GHz radar stethoscope has been developed for the diagnosis of heart sounds when direct contact with the skin is contraindicated. It is shown that a vibration sensing, bi-static radar operating in the near-field has a sensitivity maximum at a non-zero range and that maximum is proportional to the square of the radar operating frequency. By placing the instrument in the near-field, close to, but not touching the skin, a 20 dB sensitivity increase can be demonstrated. The transmitter antenna has a hot spot in the near-field which further increases the stethoscope's sensitivity. 0pt0.03in The instrument is a modified, Doppler radar based, commercial RF motion detector that transmits very low RF power. When used as a stethoscope it is shown to pose no radiation hazard to the patient or medical personnel. 0.03in An example is given to illustrate that the non-contact radar stethoscope has an audio output that is comparable in characteristics and quality to a conventional, skin-contact, acoustic stethoscope.

Keywords: Physiological monitoring - Modeling and analysis, Physiological monitoring - Novel methods, Optical and photonic sensors and systems
Abstract: The pressure wave is attenuated as it travels through the vascular bed of tissue. Consequently, reflectance photoplethysmography (PPG) waveforms probed using dual-penetrating wavelengths, such as green (G) and red (R; the deepest) are dissimilar. To unravel the dual-depth aspect of PPG, we modeled the wavelength-dependency of the shape of reflection-PPG signals in G (520-580 nm) and R (625-720 nm). Skin compression perturbs the relative contributions of the dermal and subdermal blood volume variations sources (BVVs) to PPG and was used to verify our model. We acquired reflectance-PPG in G and R on the finger of nine subjects (ages, 26-32 yrs). Two parameters were used for describing dual-depth dissimilarities: the phase shift, phi, between the first harmonics of the subdermal and dermal BVVs, and the observed phase shift (PS) between PPG signals in G and R. The average phi was 37.6, CI 95% [22.0, 53.2] degrees. At uncompressed skin, this corresponds to an average PS of 12.5, [7.8, 17.2] degrees. Our results suggest that phase parameters may enable microvascular characterization and diagnosis.

Keywords: Wearable sensor systems - User centered design and applications, Wearable wireless sensors, motes and systems, Portable miniaturized systems
Abstract: This work introduces a novel wearable olfactory display that provides just-in-time release of scents based on the physiological state of the wearer. The device can release up to three scents and passively captures subtle chest vibrations associated with the beating of the heart and respiration through clothes. BioEssence is controlled via a custom-made smartphone app that allows the creation of physiological rules to trigger different scents (e.g., when the heart rate is above 80 beats per minute, release lavender scent). The device is wireless and lightweight, and it is designed to be used during daily life, clipped on clothes around the sternum area or used as a necklace. We provide a description of the design and implementation of the prototype and potential use cases in the context of mental wellbeing.

Keywords: Integrated sensor systems, Sensor systems and Instrumentation, Wearable sensor systems - User centered design and applications
Abstract: A major hurdle for the widespread use of wearable assistive devices is determining, moment-by-moment, the control mode appropriate for a given terrain when faced with a complex, multi-terrain environment. Current control strategies focus mainly on measurements of user behavior and less on environment information.

Here we demonstrate the application of location estimates from a vision-based localization system to obtain environment awareness by delineating various terrains into regions. Given the current location and region occupied by the user, a controller could be built to select appropriate modes, predict transitions, or to add error correction.

We quantify the positional accuracy of location estimates, how well these estimates translate to classifying current region, and transitions. Performance was evaluated on eight participants without amputation wearing the sensor on the shank of the leg. We investigated the performance of an instantaneous region classifier, which used location estimates alone, and a time-history based region classifier, which used a Neural Network on a time history of location and height estimates to accomplish environment awareness. Four types of regions and six types of transitions were tested. The classifier using height estimates and time history provided accurate region labels at least 96% of the time, and accurately detected region transitions within 110 milliseconds. These results demonstrate the promise of localization for control of robotic assistive technology.

Keywords: New sensing techniques, Bio-electric sensors - Sensing methods, Bio-electric sensors - Sensor systems
Abstract: Unintentional child poisoning represents an increasingly important health issue in the United States and worldwide, partially due to increased use of drugs and food supplements. Biometric authentication is complex for pill bottles, but we propose a new method of user identification using touch capacitance during bottle-opening attempts. A smart pill bottle could generate an immediate warning to deter a child from opening the bottle and send an alert to parents/guardians. In this paper, we present principle of operation and implementation of a prototype “safe bottle”. We present the results of pilot testing with 5 adults and 3 children using support vector machine (SVM) and neural network (NN). From 232 bottle-opening events, our optimized NN generated no false detections of children as adults and four false detections of adults as children. Preliminary results indicate that smart pill bottles can be used to reliably detect children trying to open pill bottles and reduce risk of child-poisoning events.

Keywords: Novel methods, Wearable body sensor networks and telemetric systems
Abstract: Despite the extensive research that has been carried out on automatic fall detection using wearable sensors, falls in the elderly cannot be detected effectively yet. Although recent fall detection algorithms that evaluate the descent, impact and post impact phases of falls, often using vertical velocity, vertical acceleration and trunk angle respectively, tend to be more accurate than the algorithms that do not consider them, they still lack the desired accuracy required to be used among frail older adults. This study aims to improve the accuracy of fall detection algorithms by incorporating average vertical velocity and difference in altitude as additional parameters to the vertical velocity, vertical acceleration and trunk angle parameters. We tested the proposed algorithms on data recorded from a comprehensive set of falling experiments with 12 young participants. Participants wore waist-mounted accelerometer, gyroscope and barometric pressure sensors and simulated the most common types of falls observed in older adults, along with near-falls and activities of daily living (ADLs). Our results showed that, while the base algorithm with the three parameters provided 91.8% specificity, the addition of difference in altitude and average vertical velocity improved the specificity to 98.0% and 99.6%, respectively.

Keywords: Image reconstruction and enhancement - Image synthesis, CT imaging applications, Magnetic resonance imaging - MR neuroimaging
Abstract: Magnetic resonance (MR) simulators have recently gained popularity; it avoids the unnecessary radiation exposure associated with Computed Tomography (CT) when used for radiation therapy planning. We propose a method for pseudo CT estimation from MR images based on joint dictionary learning. Patient-specific anatomical features were extracted from the aligned training images and adopted as signatures for each voxel. The most relevant and informative features were identified to train the joint dictionary learning-based model. The well-trained dictionary was used to predict the pseudo CT of a new patient. This prediction technique was validated with a clinical study of 12 patients with MR and CT images of the brain. The mean absolute error (MAE), peak signal-to-noise ratio (PSNR), normalized cross correlation (NCC) indexes were used to quantify the prediction accuracy. We compared our proposed method with a state-of-the-art dictionary learning method. Overall our proposed method significantly improves the prediction accuracy over the state-of-the-art dictionary learning method. We have investigated a novel joint dictionary learning-based approach to predict CT images from routine MRIs and demonstrated its reliability. This CT prediction technique could be a useful tool for MRI-based radiation treatment planning or attenuation correction for quantifying PET images for PET/MR imaging.

Keywords: Image reconstruction and enhancement - Image synthesis, Image reconstruction - Performance evaluation, Image segmentation
Abstract: We introduce a novel method to generate biologically grounded synthetic cerebrovasculature models in a data-driven fashion. First, the centerlines of vascular filaments embedded in an acquired imaging volume are obtained by a segmentation algorithm. That imaging volume is reconstructed from a graph encoding of the centerline (i.e., generating the model's ground truth) and the segmentation algorithm is applied to the resultant volume. As the location and characteristics of the vasculature embedded in this volume are known, the accuracy of the segmentation algorithm can be assessed. Moreover, because the synthetic volume was reconstructed directly from biological data, an assessment is made on embedded filaments that are representative of the topological and geometrical characteristics of the dataset. We believe that such models will provide the means necessary for the enhanced evaluation of vascular segmentation algorithms.

Keywords: Image compression
Abstract: By increasing the volume of telemedicine information, the need for medical image compression has become more important. In angiographic images, a small ratio of the entire image usually belongs to the vasculature that provides crucial information for diagnosis. Other parts of the image are diagnostically less important and can be compressed with higher compression ratio. However, the quality of those parts affects the overall understanding of the image as well. Existing methods compress foreground and background of angiographic images using different techniques. In this paper, we first utilize a convolutional neural network to segment vessels and then represent a hierarchical block processing algorithm capable of both eliminating the background redundancies and preserving the overall visual quality of angiograms.

Keywords: Image registration, segmentation, compression and visualization - Machine learning / Deep learning approaches, Brain imaging and image analysis
Abstract: Content-based image retrieval (CBIR) is a technology designed to retrieve images from a database based on visual features. While the CBIR is highly desired, it has not been applied to clinical neuroradiology, because clinically relevant neuroradiological features are swamped by a huge number of noisy and unrelated voxel information. Thus, effective dimension reduction is the key to successful CBIR. We propose a novel dimensional compression method based on 3D convolutional autoencoders (3D-CAE), which was applied to the ADNI2 3D brain MRI dataset. Our method succeeded in compressing 5 million voxel information to only 150 dimensions, while preserving clinically relevant neuroradiological features. The RMSE per voxel was as low as 8.4%, suggesting a promise of our method toward the application to the CBIR.

Keywords: Motor learning, neural control, and neuromuscular systems, Neuromuscular systems - Learning and adaption, Neuromuscular systems - Postural and balance
Abstract: Sensorimotor control regulates balance and stability as well as adaptation to the external environment. We introduce the use of a simulated inverted pendulum to study human sensorimotor control, demonstrating that this system introduces similar control challenges to human subjects as a physical inverted pendulum. Participants exhibited longer stabilization of the system as the pendulum length between the hand and the center of mass increased while the required control input varied in a non-monotonic, yet predictable manner. Finally, we show that the experimental results can be modelled as a PD controller with a time delay of τ = 140 ms, matching the human visuomotor delay. Our results provide evidence of the importance of vision in a control of unstable systems and serve as a proof of concept of a simulated inverted pendulum.

Keywords: Motor learning, neural control, and neuromuscular systems, Neuromuscular systems - Learning and adaption, Neuromuscular systems - Postural and balance
Abstract: We examine the visual influence of stabilization in human sensorimotor control using a simulated inverted pendulum. As the inverted pendulum is fully simulated, we are able to manipulate the visual feedback independently from the dynamics during the motor control task. Human subjects performed a balancing task of an upright pendulum on a robotic manipulandum in two different visual feedback conditions. First we examined how subjects perform a task where the visual feedback is congruent with the pendulum dynamics. Second we tested how subjects performed when the physical dynamics were fixed but the visual feedback of the pendulum length was modulated. Subjects exhibited deficits in the control of the pendulum when haptic and visual feedback did not match, even when the visual feedback provided more sensitive information about the state of the pendulum. Overall we demonstrate the importance of accurate feedback regarding task dynamics for stabilization.

Keywords: Motor learning, neural control, and neuromuscular systems, Neural interfaces - Body interfaces, Neuromuscular systems - EMG processing and applications
Abstract: Permanent implantation of electrodes for prosthetic control is now possible using an osseointegrated implant as a long-term stable communication interface (e-OPRA). The number of myoelectric sites to host such electrodes can be increased by Targeted Muscle Reinnervation (TMR). Traditionally, patients need to wait few months before the TMR signals are strong enough to be recorded by electrodes placed over the skin. In this study, we report the evolution of the TMR myoelectric signals recorded from two subjects via implanted electrodes using e-OPRA, and monitored for up to 48 weeks after surgery. The signals were analyzed in regard to amplitude (signal-to-noise ratio), independence (cross-correlation) and myoelectric pattern recognition (classification accuracy). TMR signal appeared at the first follow-up, one month post-surgery, and developed around 20 dB at the last. Cross-correlation between signals decreased over time and converged to few percentage points. Classification accuracies were over 97% at the last follow up. These preliminary results suggest that implanted electrodes via the e-OPRA interface allow for an earlier and more effective use of motor signals from TMR sites compared to conventional skin surface electrodes.

Keywords: Motor learning, neural control, and neuromuscular systems, Neuromuscular systems - EMG processing and applications, Human performance
Abstract: Myoelectric control based on classification of distinct gestures discretizes the output space available to the user, which can make it difficult to react appropriately to novel scenarios such as changing limb position. While proportional myoelectric control is noisy in comparison to pattern recognition control, this noise may be an important component of skill acquisition. Here we implemented a two-dimensional proportional myoelectric controller to investigate the effects of movement direction and mapping uncertainty on adaptation to trial-by-trial perturbations. We found that subjects who practiced hitting targets despite trial-by-trial random modifications of the control mapping adapted to perturbations faster than a control group with low mapping variability. Our findings suggest that exposure to a variable mapping encourages exploratory behavior and underlies a change in adaption rate, which could potentially be used to train myoelectric control users to achieve more robust control.

Keywords: Motor learning, neural control, and neuromuscular systems, Brain physiology and modeling, Neural signal processing
Abstract: There are two popular and largely independent approaches to study the sensorimotor control system (SCS). One is to construct systems-level models of the SCS that characterize dynamics of motor regions in the brain, alpha motor neurons, and the musculoskeletal system to reconstruct motor behavior. These models view the brain as a feedforward and feedback controller that actuates the musculoskeletal system, and have been useful in understanding how the SCS generates movements. Another approach is to measure neural activity and movements simultaneously in primate and human subjects, and then analyze the data to understand how the brain encodes and controls movement. In this paper, we combine these two approaches by fitting parameters of a systems-level model of the SCS to neural activity and behavior measured from a nonhuman primate executing four types of reach-to-grasp tasks. We applied a nonlinear least squares estimation to fit parameters of the model components that characterize cerebrocerebellar processing of movement error and muscles that are actuated by alpha motor neurons receiving commands from primary motor cortex (M1). Our fitted SCS model accurately reconstructs firing rate activity of six populations of M1 neurons and associated reaching trajectories. This study paves the way for the validation of systems-level models of the SCS using experimental data.

Keywords: Motor learning, neural control, and neuromuscular systems, Neurological disorders - Stroke
Abstract: Humans typically move slower if the movement needs to be more accurate. Such a tradeoff between movement speed and accuracy is quantified in Fitts’ Law as a linear relationship between the movement time (MT) and the index of difficulty (ID). For patients with stroke, the detailed pattern of speed-accuracy tradeoff is likely affected due to disrupted neuromuscular control in stroke. In this study, we adapted a previously published iPad software program designed for the test of Fitts’ Law in children with dystonia. Subjects were asked to touch targets with different sizes and distances on the touchscreen. Data from 3 patients with stroke suggest that the post-stroke upper-extremity movements still obey Fitts’ Law, but the affected side showed larger slopes, and higher endpoint errors compared with the unaffected side. Moreover, the success rate in the affected side was significantly higher than healthy controls, but not than the unaffected side. Our preliminary data suggest that Fitts’ Law provides a promising toolkit for quantitatively assessing the movement behavior in stroke.

Keywords: Time-frequency and time-scale analysis - Nonstationary processing, Signal pattern classification
Abstract: The problem of estimating the heart rate (HR) from a facial video is considered. A typical approach for this problem is to use independent component analysis (ICA) on the red, blue, green intensity profiles averaged over the facial region. This provides estimates of the underlying source signals, whose spectral peaks are used to predict HR in every analysis window. In this work, we propose a maximum likelihood formulation to optimally select a source signal in each window such that the predicted HR trajectory not only corresponds to the most likely spectral peaks but also ensures a realistic HR variability (HRV) across analysis windows. The likelihood function is efficiently optimized using dynamic programming in a manner similar to Viterbi decoding. The proposed scheme for HR estimation is denoted by vICA. The performance of vICA is compared with a typical ICA approach as well as a recently proposed sparse spectral peak tracking (SSPT) technique that ensures that the predicted HR does not vary drastically across analysis windows. Experiments are performed in a five fold setup using facial videos of 15 subjects recorded using two types of smartphones (Samsung Galaxy and iPhone) at three different distances (6inches, 1foot, 2feet) between the phone camera and the subject. Mean absolute error (MAE) between the original and predicted HR reveals that the proposed vICA scheme performs better than the best of the baseline schemes, namely SSPT by -8.69%, 52.77% and 8.00% when Samsung Galaxy phone was used at a distance of 6inches, 1foot, and 2feet respectively. These improvements are 12.13%, 13.59% and 18.34% when iPhone was used. This, in turn, suggests that the HR predicted from a facial video becomes more accurate when the smoothness of HRV is utilized in predicting the HR trajectory as done in the proposed vICA.

Keywords: Physiological systems modeling - Signal processing in physiological systems
Abstract: Facial palsy (FP) is a clinical condition resulting from damage to the facial nerve. We hypothesize that activity can be restored in the injured side, by electrical stimulation of its muscle, using the activity of muscles on the healthy side as a control input. To explore this hypothesis, we are using a rat model of FP, which treats blinking and whisking as the features of interest in facial movement. This paper describes the development of a novel methodology for the automatic detection and measurement of eyelid displacement and blinks in video records of the rat. Specifically, the active contour approach was used to localize and track rodent eyes in a head-fixed video. The algorithm is initialized manually marking the eye contour in the first frame of the video; subsequent frames are analyzed automatically based on an energy function that depends on image features in the region of interest. Our results demonstrate that our novel technique detects blinks in video recordings with a success rate of 100% and a high correlation between the algorithm output and the manual validation.

Keywords: Time-frequency and time-scale analysis - Time-frequency analysis, Time-frequency and time-scale analysis - Nonstationary processing
Abstract: Interest in measuring heart rates (HRs) without physical contact has increased in the area of stress checking and health care. In this paper, we propose head-motion robust video-based heart rate estimation using facial feature point fluctuations. The proposed method adaptively estimates and removes such rigid-noise components as noise stemming from horizontal head motion and extracts relatively small heart signals. Rigid-noise components can be accurately estimated and removed by using changes in facial feature points which are not dominant over heart signals and are more dominant over noise signals than are such luminance signals as RGB. In evaluation experiments on a benchmark dataset, our method achieved the highest accuracy among state-of-the-art methods.

Keywords: Time-frequency and time-scale analysis - Nonstationary processing, Signal pattern classification, Parametric filtering and estimation
Abstract: This paper proposes a method of estimating drowsiness from low-frame-rate facial videos by using eyelid variability features. Since eyelid variability involves slow motions, drowsiness can be estimated more accurately with these features at low frame rate than the conventional blink-related features, in which movements may be made in only some hundred milliseconds per blink. The correlation between the ground-truth drowsiness labels and the estimated drowsiness values is compared through facial videos with frame rates ranging from 3 to 30 frames per second (fps). With conventional blink-related features, the correlation drops from 0.59 (at 30 fps) to 0.28 (at 3 fps), while, with the proposed eyelid variability features, the correlation remains nearly constant, from 0.55 (at 30 fps) to 0.53 (at 3 fps). This characteristic makes it useful for drowsiness estimation with a low computational cost.

Keywords: Neural networks and support vector machines in biosignal processing and classification, Time-frequency and time-scale analysis - Empirical mode decomposition in biosignal analysis, Signal pattern classification
Abstract: Many studies reported that eye-related movements, e.g., blank stares, blinking and drooping eyelids, are highly indicative symptoms of drowsiness. However, few researchers have investigated the computational efficacy accounted for drowsiness estimation by these eye-related movements. This paper thus analyzes two typical eye-related movements, i.e., eyelid movements Xel(t) and eyeball movements Xeb(t), and investigates neural-network-based approaches to model temporal correlations. Specifically, we compare the effectiveness of three combinations of eye-related movements, i.e., [Xel(t)], [Xeb(t)], and [Xel(t),Xeb(t)], for drowsiness estimation. Furthermore, we investigate the usefulness of two typical types of neural networks, i.e., CNN-Net and CNNLSTM-Net, for better drowsiness modeling. The experimental results show that [Xel(t),Xeb(t)] can achieve a better performance than [Xel(t)] for short time drowsiness estimation while [Xeb(t)] alone performs worse even than the baseline method (PERCLOS). In addition, we found that CNN-Net are more effective for accurate drowsiness level modeling than CNNLSTM-Net.

Keywords: Time-frequency and time-scale analysis - Time-frequency analysis, Physiological systems modeling - Signals and systems, Principal component analysis
Abstract: Occurrences of unknown words in a conversation can be challenging and often prevent people from engaging in fluent communication with each other. Even worse, currently very little is known about possible bodily responses when a listener comes across unknown words, especially when context information is not available in the conversation to facilitate understanding. In this work, we look at facial expressions and electroencephalography (EEG) as two potential body signals that may convey whether users are having difficulties understanding the words they hear. We performed an experiment to measure the reaction of users during a vocabulary dictation test using meaningful words and pseudowords. Participants were asked to classify words as they heard them into different categories. As a result, we did not see any significant differences in the facial expressions of our participants. However, significant differences were observed in event-related potentials (ERPs) within the time range of 100ms-300ms since the onset of stimuli, with pseudowords showing significantly stronger negative responses than meaningful words. Starting at about 550ms and up to around 750ms, pseudowords elicited significantly stronger negative responses, primarily over the parietal and central brain regions. Analyses for single-electrode sites revealed that pseudowords elicited more negative responses than real words in all investigated regions except the left temporal and lateral frontal regions from 500ms to 700ms since stimuli onset. These results could pave the way for future work that aims to develop real-time solutions for facilitating communication between users with different language backgrounds.

Keywords: Signal pattern classification, Data mining and processing - Pattern recognition
Abstract: This paper proposes an electromyogram (EMG) pattern classification method based on a mixture of variance distribution models. A variance distribution model is a stochastic model of raw surface EMG signals in which the EMG variance is taken as a random variable, allowing the representation of uncertainty in the variance. In this paper, we extend the variance distribution model to the multidimensional case and enhance its flexibility for multichannel and processed EMG signals. The enhanced model enables the accurate classification of EMG patterns while considering the uncertainty in the EMG variance. The robustness and applicability of the proposed method are demonstrated through a simulation experiment using artificially generated data and EMG classification experiments using two real datasets.

Keywords: Signal pattern classification
Abstract: The surface EMG (sEMG) has been used as control source for upper limb prosthetics since decades. Previous studies suggested that intramuscular EMG showed promising results for upper limb prosthetics. This study investigates the strength of combined surface and intramuscular EMG (cEMG) for improved myoelectric control. Five able-bodied subjects and three transradial amputees were evaluated using offline classification error as performance metric. Six surface and intramuscular channels were recorded concurrently from each subject for seven consecutive days and Stacked sparse autoencoders (SSAE) and LDA classifiers were used for classification. As a control source, either sEMG channels were used or combined channels were used with reduced features using PCA. In the within session analysis, cEMG (2.21 ± 1.19%) outperformed the sEMG (4.63 ± 2.07%) for both able-bodied and amputee subjects using SSAE. For between session analysis, cEMG outperformed the sEMG for both able-bodied and amputee subjects with percentage points difference of 7.93. These results imply cEMG can significantly improve the performance of pattern recognition based myoelectric control scheme for amputee subjects too and further improvement can be made by utilizing SSAE which show improved performance as compared to LDA.

Keywords: Signal pattern classification, Neural networks and support vector machines in biosignal processing and classification, Adaptive filtering
Abstract: In this paper, we present an evaluation of an adaptation of the Antonyan Vardan Transform (AVT) used in combination with an Extreme Learning Machines (ELM) classifier to process surface electromyography (sEMG) data used to classify six finger movements and a rest state. A total of 12 assays formed by three repetitions performed by four volunteers is analyzed. Additionally, a sample-by-sample output label comparison was performed to make a more comprehensive analysis of the system which was tested on a PC and embedded on a Rasp.berry Pi platform. Compared to literature papers, our system was capable to match or outperform similar solutions even using a simpler model, reaching mean accuracy rates above 94%.

Keywords: Signal pattern classification, Physiological systems modeling - Signal processing in physiological systems, Time-frequency and time-scale analysis - Time-frequency analysis
Abstract: One of the most important electrophysiological signal is the Electromyography (EMG) signal, which is widely used in medical and engineering studies. This signal contains a wealth of information about muscle functions. Therefore, the EMG signal is becoming increasingly important and has started to be used in many applications like finger movement rehabilitation. However, an advanced EMG signal analysis method is required for efficient usage of such applications. This signal analysis can include signal detection, decomposition, processing, and classification. There are many approaches in studying the EMG signals, however, one of the important factor of analyzing is to get the most efficient and effective features that can be extracted from the raw signal. This paper presents the best feature extraction set compared to previous studies. Where eighteen well-known features algorithm has been tested using the sequential forward searching (SFS) method to get excellent classification accuracy in a minimum processing time. Among these novel features only four combinations have been selected with perfect results, which are; Hjorth Time Domain parameters (HTD), Mean Absolute Value (MAV), Root Mean Square (RMS) and Wavelet Packet Transform (WPT). The superiority of this feature set has been proven experimentally, and the results show that the classification accuracy could reach up to 99% to recognize the individual and combined for ten classes of finger movements using only two EMG channels.

Keywords: Signal pattern classification, Neural networks and support vector machines in biosignal processing and classification, Data mining and processing - Pattern recognition
Abstract: High fidelity myoelectric control of prostheses and orthoses is paramount to restoring lost function to amputees and neuro-muscular disease sufferers. This research proves that spatio-temporal imaging can be used to allow convolutional neural networks to classify sparse channel EMG samples from a consumer-grade device with over 94% accuracy. 10,572 images are generated from 960 samples of simple and complex isometric finger poses recorded from 4 fully intact subjects. Real-time classification of 12 poses is achieved with a 250ms continuous overlapping window.

Keywords: Signal pattern classification, Physiological systems modeling - Signal processing in physiological systems
Abstract: This paper presents a novel set of temporally inspired time domain features for electromyographic (EMG) pattern recognition. The proposed methods employ simple time series measures derived from peak detection, and could better reflect EMG activity over time. Multiple EMG datasets consisting of 68 able-bodied and transradial amputee subjects performing a large variety of hand, wrist, fingers, and grasping movements are used to evaluate the performance of the proposed features and to design robust EMG feature sets. The results show that the average classification accuracy of two novel features, the mean prominence of local peaks and valleys, outperform several commonly used time domain features, autoregressive coefficients, histogram, and zero crossing, by 8%, 11%, and 17%, respectively. The proposed features are also shown to provide additional information as part of a robust feature set when compared to the common Hudgins’ timedomain feature set, as selected by sequential forward selection and through empirical feature set design.

Keywords: X-ray - Interventional radiology, X-ray imaging applications
Abstract: This paper proposes a method to generate grids to render realistic X-ray images of the narrow blood vessels in real-time angiography simulation. The vertexes of the narrow blood vessels are projected onto the image-rendering plane. The grids aligned in the vessel direction are generated using the projected boundary vertexes on the image-rendering plane. The average computation time of the entire simulation is reduced by 80.17% compared to the simulation using the uniform grids. The results of the questionnaire survey show that the rendered X-ray images are realistic and useful to be applied to the angiography simulation.

Keywords: Micro-CT imaging, Dual-energy X-ray imaging, Image reconstruction and enhancement - Tomographic reconstruction
Abstract: X-ray K-Edge Subtraction Computed Tomography (KES-CT) is based on the acquisition of two images at different energies, one below and one above the K-edge of a contrast agent. KES-CT is mainly performed at synchrotron facilities where a tunable monochromatic X-ray beam is available. Thanks to innovative Photon Counting X-ray Detectors (PCXDs), it would be desirable to collect the two images in a single shot with a conventional polychromatic X-ray spectrum. This approach, sometimes called spectral-CT or color-CT eliminates the risk of misregistration due to motion between consecutive acquisitions and it should allow for scans with much lower doses of contrast medium. Spectral CT is considered very promising but its practical application is being hampered by several practical issues, one of these being the charge sharing affecting the energy resolution of PCXDs. However, latest generations of PCXDs implement hardware solutions to cope with the charge sharing effects, thus allowing sharper color sensitivity. This work presents a K-edge spectral CT imaging preliminary protocol based on the Pixirad-I/Pixie-III detector where discrete tomography is used to present the reconstructed slices as color images. Results show that when a solution for the charge sharing issue is considered and refined reconstruction methods are applied, accurate K-edge subtraction imaging with conventional sources can be performed.

Keywords: PET and SPECT imaging, PET and SPECT Imaging applications
Abstract: Respiratory motion during PET/CT imaging is a matter of concern due to degraded image quality and reduced quantitative accuracy caused by motion artifacts. One class of motion correction methods relies on hardware-based respiratory motion tracking systems in order to use respiratory cycles for correcting motion artifacts. Another class of hardware free methods extract motion information from the reconstructed images or sinograms. Hardware-based methods, however, are limited by calibration requirement, patient discomfort, lack of adaptability during scanning, presence of electronic drift during respiratory monitoring etc. Extracting motion information from reconstructed images is also limited by the fact that the original raw information requires significant processing before it can be used. Hence the motivation behind this work is to introduce a software-based approach that can be applied on raw 64-bit listmode data. The basic design of the proposed method is based on the fundamentals of Positron Emission Particle Tracking (PEPT) with additional incorporation of Time of Flight (TOF) information. Respiratory motion of patients has been extracted from the raw PET data by tracking a point source attached to the patient in areas on and near the chest. The key objective of this work is to describe a new process by which this particle tracking based motion correction system can eventually be lesion specific and correct the motion for a particular lesion within the patient. This work thus serves as a framework for lesion specific motion correction.

Keywords: Cardiovascular and respiratory system modeling - Blood flow models, Cardiovascular and respiratory system modeling - Vascular mechanics and hemodynamics, Cardiovascular, respiratory, and sleep devices - Diagnostics
Abstract: Reduced fluid-structure interaction models are the key component of hemodynamic simulation. In this work, a multi-purpose computational model applicable to specific physiological components such as arterial, venous and cerebrospinal fluid circulatory systems has been developed based on the Hamilton’s variational principle. This model encompasses a viscous Newtonian fluid - structure interaction (FSI) framework for the large compliant bifurcated arterial networks and its subsystems. This approach provides the groundworks for a correct formulation of reduced FSI models with an account for arbitrary non-linear viscoelastic properties of a compliant vascular tree. The hyperbolic properties of the derived mathematical model are analyzed and used to construct the Lax-Wendroff finite volume numerical scheme, with second-order accuracy in time and space. The computational algorithm is validated against well-known numerical and in vitro experimental data reported in the literature for the case of human arterial trees, comprising 55 and 37 main arterial vessels. Utilizing the physics based nonlinear constitutive framework, this model can be adequately tested, calibrated and applied for patient-specific clinical diagnosis and prediction.

Keywords: Cardiovascular regulation - Autonomic nervous system, Cardiovascular regulation - Heart rate variability, Cardiovascular and respiratory system modeling - Cardiovascular-Respiratory Interactions
Abstract: The aim of this preliminary study is to investigate if there is any evidence of maternal-fetal heart rate coupling in mice fetuses and how the coupling patterns are regulated by vagal nervous system on beat by beat. Total of 6 pregnant female mice were divided into two groups [control (N=3) and vagal blockade (N=3)]. On 17.5-day beat-to-beat heart rates of mothers and fetuses (MHR and FHR) were simultaneously measured for 20 minutes (10 minutes under normal condition and 10 minutes with saline (to control group) and atropine (to the vagal blockade group) solution by using an invasive maternal and fetal electrocardiogram techniques with needle electrodes. Results show that occasional strong maternal-fetal heart rate coupling (strength was measured by lamda(λ)) appeared and its patterns changed with atropine infusion (no change with saline). Additionally, fisher’s exact test shows that changes (increase/decrease from pre to post injection values) in mean, rmssd and power spectral density (PSD) (2~4 Hz) of MHR, rmssd FHR and PSD (2~4 Hz) of λ were found to be significantly (p<0.05) associated with treatment types (saline/ atropine). The presented results and protocol allow for the first time in the assessment of autonomic regulation of maternal and fetal heart and their interactions, which will further enhance the value of the mouse as a murine model of heritable human pregnancy and perinatal complications due to maternal conditions.

Keywords: Cardiovascular and respiratory system modeling - Cardiovascular control models
Abstract: Arterial hemodynamic assessment has always been essential for clinical Cardiovascular System (CVS) diagnosis. Using Windkessel (WK) lumped parametric model as noninvasive measurement tool provides the potential of achieving a very convenient, computational inexpensive and accurate prediction of the arterial parameters. Many versions of WK models have been proposed and extensively studied, over the last century. In general, they can be classified into two categories: elastic and viscoelastic models. Recently, several studies have discussed the potential of describing the arterial wall viscoelasticity using fractional order models, reducing the number of parameters and exposing a natural response. Hence, a key missing item in the arterial Windkessel modeling is a fractional-order analog component that can provide a reliable, realistic and reduced representation of the fractional viscoelasticity behavior. In this paper, we present, for the first time, a three-element fractional-order viscoelastic Windkessel model. The proposed model incorporates a fractional-order capacitor that substitutes the ideal capacitor of standard three elements WK model. The latter non-ideal element combines both resistive and capacitive properties which displays viscoelastic behavior of the arterial vessel. The contribution of both properties is controlled by the fractional differentiation order (alpha) enabling an accurate and reliable physiological description.

Keywords: Pulmonary and critical care - Pulmonary function testing & instrumentation, Cardiovascular and respiratory system modeling - Compartmental modeling, Cardiovascular and respiratory system modeling - Gas exchange models
Abstract: Capnography records CO2 partial pressure in exhaled breath as a function of time or exhaled volume. Time-based capnography, which is our focus, is a point-of-care, noninvasive, effort-independent and widely available clinical monitoring modality. The generated waveform, or capnogram, reflects the ventilation-perfusion dynamics of the lung, and thus has value in the diagnosis of respiratory conditions such as chronic obstructive pulmonary disease (COPD). Effective discrimination between normal respiration and obstructive lung disease can be performed using capnogram-derived estimates of respiratory parameters in a simple mechanistic model of CO2 exhalation. We propose an enhanced mechanistic model that can capture specific capnogram characteristics in congestive heart failure (CHF) by incorporating a representation of the inertance associated with fluid in the lungs. The 4 associated parameters are estimated on a breath-by-breath basis by fitting the model output to the exhalations in the measured capnogram. Estimated parameters from 40 exhalations of 7 CHF and 7 COPD patients were used as a training set to design a quadratic discriminator in the parameter space, aimed at distinguishing between CHF and COPD patients. The area under the ROC curve for the training set was 0.94, and the corresponding equal-error-rate value of approximately 0.1 suggests classification accuracies of the order of 90% are attainable. Applying this discriminator without modification to 40 exhalations from each CHF and COPD patient in a fresh test set, and deciding on a simple majority basis whether the patient has CHF or COPD, results in correctly labeling all 8 out of the 8 CHF patients and 6 out of the 8 COPD patients in the test set, corresponding to a classification accuracy of 87.5

Keywords: Cardiovascular and respiratory system modeling - Respiratory Control models, Cardiovascular and respiratory system modeling - Compartmental modeling, Cardiovascular and respiratory system modeling - Gas exchange models
Abstract: We have previously introduced a model for closed-loop respiratory control incorporating an explicit conductance-based model of bursting pacemaker cells driven by hypoxia sensitive chemosensory feedback. Numerical solution of the model equations revealed two qualitatively distinct asymptotically stable dynamical behaviors: one analogous to regular breathing (eupnea), and a second analogous to pathologically rapid, shallow breathing (tachypnea). As an experimental test of this model, we created a hybrid in vitro/in silico circuit. We used Real Time eXperimental Interface (RTXI) dynamic clamp to incorporate a living pacemaker cell recorded in vitro into a numerical simulation of the closed-loop control model in real time. Here we show that the hybrid circuit can sustain the same bistable behavior as the purely computational model, and we assess the ability of the hybrid circuit to recover from simulated bouts of transient hypoxia.

Keywords: Cardiovascular and respiratory system modeling - Cardiovascular control models, Coronary blood flow, Cardiovascular, respiratory, and sleep devices - Implantables
Abstract: Bridge to recovery with left ventricular assist device (LVAD) support has been more prominent with volume displacement pumps (VDPs) than with rotary blood pumps (RBPs), which may be due to VDPs providing greater ventricular unloading and coronary artery flow. To compare ventricular unloading and coronary flow in VDPs and RBPs in a repeatable environment, a physiologic coronary circulation was added to a pre-existing mock circulatory loop. In this study, a physiologic coronary circulation, mimicking a healthy or diseased auto-regulatory response was implemented in a mock circulatory loop. Using the mock circulation loop, a VDP with original (Björk-Shiley) and then replacement (jellyfish) valves was operated in clinically recommended modes and compared to full and partial assist RBP operating at constant speed and rapid speed modulated modes. The Björk-Shiley VDP resulted in increased pressure-volume area, which resulted in greater coronary artery flow when compared to the improved jellyfish valves. Full assist RBP support reduced left ventricular stroke, pressure-volume area and coronary flow compared to partial assist, whilst the effect of speed modulation modes was not as significant. Of all LVAD operating modes, the counter-pulsed VDP with jellyfish valves demonstrated the greatest cardiac energetics and coronary flow. This study provides a basis for further investigation into RBP speed modulation profiles to match the improved haemodynamic performance of VDPs.

Keywords: Cardiovascular and respiratory signal processing - Cardiovascular signal processing
Abstract: A novel technique to derive respiratory rate from pulse photoplethysmographic (PPG) signals is presented. It exploits some morphological features of the PPG pulse that are known to be modulated by respiration: amplitude, slope transit time, and width of the main wave, and time to the first reflected wave. A pulse decomposition analysis technique is proposed to measure these features. This technique allows to decompose the PPG pulse into its main wave and its subsequent reflected waves, improving the robustness against noise and morphological changes that usually occur in long-term recordings. Proposed methods were evaluated with a data base containing PPG and plethysmography-based respiratory signals simultaneously recorded during a paced-breathing experiment. Results suggests that normal ranges of spontaneous respiratory rate (0.1-0.5 Hz) can be accurately estimated (median and interquartile range of relative error less than 5%) from PPG signals by using the studied features.

Keywords: Cardiovascular and respiratory signal processing - Cardiovascular signal processing, Cardiovascular and respiratory signal processing - Heart Rate and Blood Pressure Variability, Cardiovascular regulation - Heart rate variability
Abstract: This study explored the autonomic nervous system (ANS) adaptation in relation to exercise and how this correlates with the ratings of perceived exertion (Borg-RPE) over four ranges 6-8; 9-12; 13-16; 17-20, by using the time domain parameters and the multi-lag Tone-Entropy (T-E) of heart rate variability (HRV). ECG signals were collected from ten subjects who were recruited to participate in a graded exercise protocol on a treadmill. Results showed that SDNN and RMSSD decreased from lower to higher Borg-RPE, indicating a decrease in HRV. Entropy significantly decreased along the first 3 Borg-RPE ranges but increased in the recovery phase in which Tone values became negative (high HRV). As Borg-RPE values increased to the 17-20 range, Tone values decreased and Entropy increased compared to the 13-16 interval suggesting vagal predominance as opposed to HRV time domain results. The highest value of Tone was observed in the Borg-RPE 9-12 range indicating paramount sympathetic dominance. The use of multi-lag in T-E 2D space improved the separation of HRV with reference to the Borg-RPE intervals (p < 0.05), except between the 13-16 and 17-20 ranges of the Borg-RPE. Results highlighted the analytical power of T-E in assessing both HRV changes and the sympatho-vagal balance throughout a graded exercise. Potentially, T-E can be applied to assess rehabilitation settings and to get further information on ANS modulation at high exercise intensities.

Keywords: Cardiovascular and respiratory signal processing - Heart Rate and Blood Pressure Variability, Cardiovascular and respiratory signal processing - Time-frequency, time-scale analysis of cardiorespiratory variability
Abstract: Stress management is particularly important for healthcare of modern people. In stress research, heart-rate variability (HRV), indicating the change of time intervals in successive heart beats, significantly contributed due to its close relationship with autonomic nervous system. However, the adaptive response to stress, also known as stress resilience, has not been studied much yet. We collected electrocardiogram during mental and physical stress, experimentally designed by mental arithmetic tasks and physical activities for 14 healthy subjects. As a result, we found that resting HRV parameters, particularly associated with the parasympathetic activity, had significant positive correlations with reactivity and recovery from mental and physical stress. These HRV parameters can be used as a measure of stress resilience quantitatively. Our findings suggest that these parameters can help one’s stress management by enabling to predict the adaptive response to upcoming stressful events.

Keywords: Cardiovascular, respiratory, and sleep devices - Monitors, Sleep - Obstructive sleep apnea, Cardiovascular and respiratory signal processing - Cardiovascular signal processing
Abstract: In this paper, we present a principal component analysis (PCA) method for estimating the respiration from overnight ECG recording. In comparison to other published methods, our method is very fast to compute and has low memory requirements, which makes it suitable for processing long duration ECG recordings. We used our method to derive respiratory features for the ECG which were then used to identify epochs of sleep apnoea from the ECG. Three classifiers including the extreme learning machine (ELM), linear discriminant analysis, and support vector machine were used to detect sleep apnoea. The method was evaluated on the MIT PhysioNet Apnea-ECG database. Apnoea detection was evaluated with leave-one-record-out cross-validation. Our PCA method obtained the highest accuracy of 74% by ELM classifier. We conclude that the fast PCA method is useful to apply PCA to long ECG recordings.

Keywords: Cardiovascular and respiratory signal processing - Non-linear cardiovascular or cardiorespiratory relations, Cardiovascular and respiratory signal processing - Cardiovascular signal processing
Abstract: Several neurological and mechanical non-linear mechanisms relate the respiratory and cardiovascular systems to one another. Besides the well-known modulation of heart rate by respiration, another form of non-linear interaction between both systems is Cardiorespiratory Phase Synchronization (CRPS). In this study we investigated CRPS on a group of 27 healthy individuals subject to a stimulation protocol with five different mental states: a basal state, a videogame, a comedy video, a suspense video and a reading state. A continuous measure of CRPS was calculated from the phase synchrogram between respiratory and electrocardiographic signals. Periods of CRPS were characterized by their average duration (AvDurSync) and by the percentage of synchronized time (%Sync) within each mental state. These measures were studied considering two thresholds: a minimum amplitude and a minimum duration for synchronization. Each subject exhibited a particular pattern of phase locking ratios along the different mental states. We observed that, in all states, %Sync decreased and AvDurSync increased in proportion to the minimum duration threshold. Both measures were inversely proportional to the minimum amplitude threshold. During the videogame, subjects showed a significantly higher %Sync as compared to any other mental stimulus, irrespective of the minimum duration threshold. Mental stimulation can be an alternative approach to enhance cardiorespiratory coupling when subjects have difficulties to perform aerobic exercise, such as in patients with Chronic Obstructive Pulmonary Disease or Chronic Heart failure.

Keywords: Coronary artery disease, Cardiovascular, respiratory, and sleep devices - Diagnostics, Cardiovascular and respiratory signal processing - Cardiovascular signal processing
Abstract: This research presents a novel statistical model for diagnosing acute myocardial infarction (AMI). The model is based on features extracted from a reduced lead system consisting of a subset of three leads from the standard 12-lead ECG. We selected a set of relevant parameters commonly used in the clinical practice for ECG-based AMI diagnosis, namely ST elevation and T-wave maximum. We also selected features, not used in clinical practice, that were derived from vectorcardiography and computed on the reduced three-lead system (pseudo-VCG parameters). To validate the model, we used 104 patients coming from the Physionet STAFF III database which contains 12-lead ECG recordings at baseline and in coronary artery occlusion condition during angioplasty (PTCA). Results show that pseudo-VCG features are able to diagnose AMI slightly better than ST elevation and T-wave maximum features together (area under the ROC curve (AUC) 0.87 vs AUC 0.85). When combining pseudo-VCG features together with ST elevation, and T-wave maximum, the performance improved significantly (AUC 0.95, sensitivity 89.6% and specificity 82.7%). Results indicate a potential for diagnosing AMI using the proposed reduced lead system and the selected set of features. We suggest its possible use for diagnosing AMI in long-term, ambulatory and home monitoring situations, allowing an earlier and faster diagnosis.

Keywords: Diagnostic devices - Physiological monitoring
Abstract: The electrical properties of biological tissues differ depending on their structural characteristics. In literature, a lot of study have been carried out with the intent of taking advantage of bioimpedance analysis. Unfortunately, many apparatuses used during these evaluations were not always designed for measurements on living tissues. As a consequence, data could be affected by electrode polarization. In 2016, we presented a new impedance meter, developed for measurements on living tissues. Initially, it was tested only on ex-vivo rabbit’s tissues with promising results. As a continuation, this device has been tested on in-vivo samples by placing a needle-probe into 3 tissues (liver, spleen, ovary) of 2 female dogs. Furthermore, was evaluated also the bioimpedance signal of the ovary explanted, comparing it with the in-vivo data. Bioimpedance was analyzed in terms of modulus and phase along a broad spectrum of frequencies (10Hz – 10kHz). Data obtained confirm the possibility of discriminating among the 3 tested tissues, at high frequencies for modulus and at low for phase. Confirmation that values on in-vivo and ex-vivo tissues are comparable if detected within few minutes after the explant, is also reported. We conclude that this clinical evaluation confirmed, also in-vivo, the good performance of the device previously tested on ex-vivo tissues, and provide more information about the tissue properties and characteristics.

Keywords: Clinical laboratory, assay and pathology technologies
Abstract: A versatile robot platform is presented that can be used to design low-cost custom made microscopes in do-it-yourself construction. All components like the framework, the linear drives, robot controller and driver, the illumination and the camera are described as well as optional features like fluorescence microscopy and auto-focus. Finally, an application for automated imaging of 3D-cell cultures in 96-well microplates are presented.

Keywords: Diagnostic devices - Physiological monitoring, Ambulatory diagnostic devices - Wellness monitoring technologies, Wearable or portable devices for vital signal monitoring
Abstract: Fever is one significant sign of infection. Hence, infrared thermography systems are important for detecting infected suspects in public places. Reliable temperature measurements by thermography are influenced by several factors, including environmental conditions. This paper proposes a linear regression analysis-based facial temperature optimization method to improve the accuracy of multiple vital signs-based infection screening at various ambient temperatures. To obtain the relationship between ambient temperature and thermography measurements, 20 instances of axillary temperature, thermography measurements of facial temperature, and five different ambient temperature values at the time of measurement were used as a training set for a linear regression model. Temperatures from a total of 30 subjects were recalculated by the model. The screening system was evaluated using the temperature both before and after optimization to demonstrate the accuracy of the optimization method. A k-nearest neighbor algorithm was used to classify potentially infected patients from healthy subjects. Although the system has already been evaluated in restricted environmental conditions, this is the first time it was tested in Ulaanbaatar, Mongolia. The results show that the Pearson’s correlation coefficient between optimum and axillary temperatures increased to r = 0.82. Paired t-tests revealed that the optimized temperature became statistically highly significant (p<0.001) for differentiating potentially infected patients from healthy subjects. Finally, the system achieved a sensitivity score of 93% and a negative predictive value of 92%. These values are higher than those obtained without temperature optimization. The proposed optimization method is feasible and can notably improve screening performance.

Keywords: Diagnostic devices - Physiological monitoring, Cardiovascular assessment and diagnostic technologies, Plethysmography
Abstract: During pregnancy, the pregnant mother undergoes significant physiological changes in order to accommodate the developing fetus. In recent years, arterial pulse wave has been widely used to reflect these physiological changes. The aim of this study was to investigate the changes of radial pulse and photoplethysmography (PPG) pulse waveform characteristic with gestational age in normal pregnant women. 40 pregnant women volunteers were recruited from February 2016 to September 2016 from the Haidian Maternal & Child Health Hospital in Beijing. Both radial pulses and PPG pulses were recorded simultaneously using a PowerLab data collection system at a sampling rate of 1000Hz for offline analysis. Their pulses were measured from each pregnant woman at three trimesters (first trimester between week 11-13; second trimester between week 20-22 and the third trimester between week 37-39). Three waveform characteristics (total pulse area; pulse area1: the area before the notch position; pulse area2: the area after the notch position) were derived. The results showed that the total pulse area and pulse area2 from both radial and PPG pulses decreased significantly between two paired consecutive trimesters (all P<0.01, except the comparisons between the second and third trimesters for PPG pulses). In summary, this study has quantified the pulse waveform characteristic differences in terms of pulse areas between the three trimesters, providing useful scientific evidence to better understand the cardiovascular physiological changes during normal pregnancy.

Keywords: Ambulatory Diagnostic devices - Point of care technologies, Clinical laboratory, assay and pathology technologies, Diagnostic devices - Physiological monitoring
Abstract: Electrophoresis is widely applied in the field of biochemistry and molecular biology. Tetrapolar electrical impedance sensing (TEIS) has been shown capable of replacing the conventional detection technology in order to develop a point of care electrophoretic analyzer. Besides the advantages of reduced influence of electrode polarization, TEIS is affected by sensitivity distribution depending on the electrode design. A well reported practice outside of electrophoresis, systematic investigation of the effects of sensitivity distribution on the TEIS in microfluidic devices has not been conducted. Here we utilize finite element modeling, backed by experimental results, to optimize the sensor design within an electrophoretic separation device. Numerous sensor designs were validated regarding detectability, sensitivity and spatial resolution. The results show, that minimizing the distance between the central/pick-up electrodes increases sensitivity and spatial resolution whereas the distance between the central electrodes and the outer electrode do not influence sensitivity and spatial resolution.

Keywords: Cardiovascular assessment and diagnostic technologies, Diagnostic devices - Physiological monitoring, Plethysmography
Abstract: In the context of global health, telemedicine, and low-resource settings, we present a non-invasive smart-phone based device that can be used to screen for atherosclerosis, which is the leading factor for ischemic heart attacks and strokes. Using a custom Android mobile application, our device computes Pulse Wave Velocity (PWV) using the pulse signals from photo-plethysmographic (PPG) probes, which are simultaneously clipped onto the ear, index finger, and big toe of a human subject. Unlike other designs which require the use of an ECG reference, our mobile device uses only PPG signals and is entirely powered by the mobile phone via the USB port. Using the ear signal as a reference, we derived PWV values from two locations: the right index finger, and the right big toe. We present data from a recent clinical study with 78 participants (age 26 to 74) who were divided into three groups: Coronary Arterial Disease (“CAD”), hypertensive group (“Pre-CAD”), and Healthy controls. The CAD group was clinically diagnosed and confirmed with a CT-scan and calcium scoring. PWV values derived from the finger was found to have too much variance to be clinically useful. However, PWV values derived from the toe location showed significant differences between the groups, even after accounting for age. Measured PWV values were: 10.07 (8.51-12.01) for the older CAD group, 9.39 (7.44-9.75) for the younger CAD group, 8.26 (7.26-9.22) for the older Pre-CAD group, 10.57 m/s (8.5-11.2) for the younger Pre-CAD group, 7.13 m/s (5.97-7.69) for older healthy controls, and 6.71 m/s (4.86-7.26) for the younger healthy control subjects. These results demonstrate good potential value of our mobile PWV device as a simple low-cost screening tool for atherosclerosis and coronary arterial disease.

Keywords: BioMEMS/NEMS - Tissue engineering and biomaterials, Micro- and nano-technology, Nano-bio technology design
Abstract: We demonstrate a hybrid biological photovoltaic device by forming a 3D cooperative biofilm of cyanobacteria and heterotrophic bacteria. 3D bioprinting technique was applied to engineer a cyanobacterial encapsulation in hydrogels over the heterotrophic bacteria. The device continuously generated bioelectricity from the heterotrophic bacterial respiration with the organic biomass supplied by the cyanobacterial photosynthesis. This innovative device platform can be the most suitable power source for unattended sensors, especially for those deployed in remote and resource-limited field locations.

Keywords: Translational issues in tissue engineering and biomaterials, Stem cells - Tissue morphogenesis, Electric fields - Tissue regeneration
Abstract: Transplantation of cells into central nervous system (CNS) shows a potential for treatment of post-traumatic and neurodegenerative diseases. Cadaver-derived neural cells can help reducing deficit of allogeneic material ready for transplantation. In this study we analyze post-mortal survival of spinal cord neural cells. Maximal time when alive neuronal cells can be recovered form spinal cord of the animals was determined as 56hr for human-size animal and 18hr for rat. Cells with surface expression of ganglioside GD2 and antigen CD24 constituted up to one percent of all recovered alive cells in earlier samples with time dependent decline in percentage. GD2-positive cells from rat spinal cord demonstrated spontaneous and drug-induced electrical activity, which reduces with time post mortem.

Keywords: Biomaterial-cell interactions - Engineered vascular tissue, Scaffolds in tissue engineering - Biofabrication
Abstract: Electrospinning has been widely used to fabricate scaffolds and commonly used biodegradable polymers. Cellular cardiomyoplasty is a type of regenerative medicine that has potential use for treatment of myocardial infarction or terminal cardiac failure. The aims of this study are to use electrospinning to create cardiovascular patches and to assess their potential therapeutic use by transplantation into the hearts of rats. Tissue engineering scaffolds were generated by use of electrospinning, in which the fibers consist of nanoscale-to-microscale fibers whose diameters are comparable to those of essential components of the extracellular matrix. A polymer solution was pumped at a constant rate through a syringe with a small-diameter needle that is connected to a high-voltage source, so that an electric field is created between the needle and a metallic collecting plate. The final product is a mat composed of individual continuous nanofibers. Cell survival, cell characteristics, and growth factors of electrospun patches of different thicknesses using bone marrow and human cardiac stem cells were tested. The results demonstrated that the cells can survive in Poly-caprolactone (PCL) patches, even deep within these patches. The PCL patches are nontoxic and do not alter cell properties. Transplantation of these patches into the hearts of a rat model of myocardial infarction led to strong compliance and good survival. The use of PCL cellular patches is feasible method for cellular transplantation. Future studies should attempt to use orientated electrospun cellular patches to improve overall cell survival within deeper layers of these patches.

Keywords: Microfluidic techniques, methods and systems, Micro- and nano-technology, Microfluidic applications
Abstract: This work presents a microfluidic Digital High-Resolution Melt platform for absolute quantitation and sensitive detection of locus-specific sequence variations on a molecule-by-molecule basis. The platform provides a facile means for assessment of hundreds to thousands of single DNA copies by digitizing template molecules in a 4096 1-nL array microfluidic device and observing the sequence-dependent fluorescence changes during temperature ramping. The analytical capability of this platform is demonstrated in several applications, such as digital assay characterization, detection and assessment of DNA methylation heterogeneity, and detection of rare biomarkers at frequencies as low as 0.0005% target to background molecules.

Keywords: Microfluidic applications
Abstract: In diagnosing bacterial infection, rapid bacterial identification (ID) and antimicrobial susceptibility testing (AST) are critical to clinicians in order to provide an effective treatment in a timely manner. The gold standard, culture-based approach provides both ID and antimicrobial susceptibility but requires several days of turnaround time. Especially in polymicrobial infections, where there are more than one organisms interacting collectively that can complicate the treatment. Here, we demonstrate a rapid bacterial diagnostic approach that is capable of bacterial ID/AST in heterogeneous samples within less than 4 hours by using digital PCR (dPCR) and digital high-resolution melt via microfluidic devices. By utilizing dPCR, we are able to quantify amount of nucleic acid, which correlates to phenotypic responses of individual pathogens in a mixed sample and also shorten the required time of antibiotic exposure. In addition, we employ a machine learning algorithm to automatically identify bacterial species based on melt profiles of 16S rRNA gene.

Keywords: Microfluidic applications, Microfluidic techniques, methods and systems, Nano-bio technology design
Abstract: A microfluidic valve-based trap enabling controlled capture, release, and temporary immobilization of droplets together with on-demand merging of selected droplets is presented in this paper. The microfluidic trap technology can merge droplets passively or in active manner via a pneumatically actuated membrane. A microchip is developed with two functional units of droplet generator and merging mechanism to implement the passive or active merging performance of the microfluidic valve-based trap using a low and high surfactant concentrated continuous oil-phase.

Keywords: General and theoretical informatics - Predictive analytics, General and theoretical informatics - Algorithms, General and theoretical informatics - Pattern recognition
Abstract: Algorithms can automatically predict seizures to reduce the occurrences of accidental injury and improve living conditions of patients. This paper proposes a novel patient-specific algorithm based on multi-channel scalp EEG recordings. 26 features for each channel are extracted from each one-second data, including 8 absolute spectral powers, 8 normalized spectral powers, 8 power spectral entropies, the shortest path length and clustering coefficient. Then, a new step to select the most discriminative five minute preictal period is proposed. The features of preictal period are combined with that of five minute non-seizure period to form a training set in order to achieve the maximum linear separability criteria. Then, the effective features of each channel are selected by Elastic Net. At the same time, greedy algorithm is used to select effective channels. The ten minute effective features obtained from effective channels are input to Logistic Regression. The algorithm is tested on 62 seizures from 12 patients in 217 hours of recordings in MIT database. Results are finally given by average of each 1 minute values of Logistic Regression. It is shown that the proposed algorithm can achieve a sensitivity of 91% and an averaged false positive rate of 0.3 per hour.

Keywords: Public Health Informatics - Health risk evaluation and modeling, General and theoretical informatics - Predictive analytics, Health Informatics - Preventive health
Abstract: A timely prediction of type 2 diabetes (T2D) onset is important for early intervention to prevent, or at least postpone, its incidence. Several models to predict T2D onset according to individual risk factors were proposed. However, their practical applicability is limited by the fact that they often perform suboptimally when applied to a different population. A solution to overcome this limitation is model recalibration, which consists in updating the model parameters. The aim of this work is to demonstrate the benefits of T2D predictive model recalibration. For the purpose, we considered as case study the Diabetes Population Risk Tool (DPoRT), originally tuned for the Canadian population, and we applied it to data collected in older Americans in the Health and Retirement Study (HRS). A subset of 30,274 subjects was extracted from HRS and divided into a training (N=24,219) and a test set (N=6,055) stratifying for sex and diabetes incidence. The DPoRT was recalibrated by re-estimating all model coefficients on the training set, and then assessed on the test set by comparing the performance of recalibrated vs original model. Model discriminatory ability and calibration were assessed by the concordance index (C-index) and the expected to observed event probability ratio (E/O), respectively. Results show that the recalibrated DPoRT presents similar discriminatory ability to the original model, with C-index equal to 0.68 vs. 0.67 in men, 0.73 vs. 0.73 in women, and better calibration than the original model, with E/O ratio equal to 0.75 vs. 4.57 in men, 0.81 vs. 2.53 in women. Results confirm that recalibration is a key step to be performed before the application of predictive models to different populations in order to guarantee an accurate prediction of diabetes incidence.

Keywords: General and theoretical informatics - Predictive analytics, General and theoretical informatics - Machine learning, General and theoretical informatics - Artificial Intelligence
Abstract: Electronic medical claim (EMC) database has been successfully used for predicting occurrences of stroke and a variety of other diseases. However, inadequate predictive performances have been observed in cases of rare occurrences due to both insufficient training samples and highly imbalanced class distribution. In this work, our aim is to improve stroke prediction, especially for young age group (25-45 year-old) in a large population-based EMC database (552,898 subjects). We learn a young stroke predictive deep neural network model using a novel active data augmenter. The augmenter selects the most informative EMC data samples from old age stroke patients. This approach achieves 9.3% and 8.2% area under the receiver operating characteristic curve (AUC) value improvements compared to training directly with only young age group data and training all age groups data, respectively. We further provide analyses on the AUC values obtained as a function of the training data size, and the amount and the type of augmented data samples.

Keywords: General and theoretical informatics - Predictive analytics, General and theoretical informatics - Statistical data analysis
Abstract: Acute myocardial infarction (AMI) is a serious cardiovascular disease caused by acute or persistent ischemic and anoxia of the coronary artery. A more practical and effective risk model is still remained to be established for AMI patients. This study aims to investigate the predictive value of prothrombin time (PT) in AMI patients. In this study, 2734 AMI patients available in the public MIMIC III clinical database were investigated, with 629 deaths occurring within 2-years follow-up. More than 20 risk factors including demographics, clinical disease history, laboratory test information, surgery history, and mediation information were analyzed as potential predictors for all-cause mortality in AMI patients. After adjustment for other covariates, PT was showed to be a significant risk factor for all-cause mortality in AMI patients (adjusted hazard ratio, 4.04; 95% confidence interval, 2.83 to 5.75) from Cox regression analysis. We also developed a comprehensive risk model for AMI mor-tality using multivariate Cox proportional hazards model based on the above 20 risk factors. Combined with PT, the model achieved a good accuracy with an AUC (area under ROC curve) of 0.843. Overall, PT is an independent predictor for 2-year mortality in AMI, and it might be useful in identifying AMI patients with a high risk for mortality.

Keywords: Health Informatics - Healthcare modeling and simulation, General and theoretical informatics - Predictive analytics, General and theoretical informatics - Statistical data analysis
Abstract: Outpatient centers comprised of many concurrent clinics increasingly see higher patient volumes. In these centers, decisions to improve clinic flow must account for the high degree of interdependence when critical personnel or equipment is shared between clinics. Discrete event simulation models have provided clinical decision support, but rarely address high-volume clinics with shared resources. While highly complex models are now capable of representing clinics in detail, validation techniques often do not evaluate model predictive performance when presented with new data. Cross-validation provides a means to evaluate the robustness of model treatment time predictions when ongoing data collection in clinics is impractical. Ensuring robust predictions assures validity in the use of models to optimize clinic performance. We apply cross-validation in evaluating a model of glaucoma clinic service at Duke Eye Center. In-person observation is used to verify the accuracy of operations data collected through electronic health records (EHR). From the EHR data, we formulate a stochastic reward net model, employing phase-type distributions to represent treatment durations, and solved through discrete event simulation. The model is formulated in two configurations to represent (1) concurrent demand on clinic staff, or (2) independently functioning clinics. Evaluating these two alternatives in cross-validation studies, we find model prediction accuracy improves when interdependence is explicitly modeled in the examined setting.

Keywords: General and theoretical informatics - Predictive analytics, Bioinformatics - Cancer genomics, Neuro genomics, Cardio genomics
Abstract: Pan-cancer analysis is a significant research topic in the past few years. Due to many advancing sequencing technologies, researchers possess more resources and knowledge to identify the key factors that could trigger cancer. Furthermore, since The Cancer Genome Atlas (TCGA) project launched, using machine learning (ML) techniques to analyze TCGA data has been recognized as a useful solution in the line of research. Therefore, this study uses RNA-sequencing data from TCGA and focuses on classifying thirty-three types of cancer patients. Five ML algorithms include decision tree (DT), k nearest neighbor (kNN), linear support vector machine (linear SVM), polynomial support vector machine (poly SVM), and artificial neural network (ANN) are conducted to compare the performances of their accuracies, training time, precisions, recalls, and F1-scores. The results show that linear SVM with a 95.8% accuracy rate is the best classifier in this study. Several critical and sophisticated data pre-processing experiments are also presented to clarify and to improve the performance of the built model.

Keywords: Integrated sensor systems, Wearable body sensor networks and telemetric systems, Physiological monitoring - Instrumentation
Abstract: A wearable multi-sensor system allowing synchronized unobtrusive measurements of 4 vital signs at a dedicated location of interest is presented. The 4xU sensor is capable of synchronously measuring magnetic impedance, reflective photoplethysmography, capacitive electrocardiogram and seismocardiography (ballistocardiography). The hardware of all modalities is described and some preliminary results are reported.

Keywords: Sensor systems and Instrumentation, Wearable wireless sensors, motes and systems, Physiological monitoring - Instrumentation
Abstract: A wide variety of sensors have been developed in the biomedical engineering community for telemedicine and personalized healthcare applications. However, they usually focus on sensor connectivity and embedded signal processing, at the expense of the sensing part. This observation lead to the development and exhaustive evaluation of a new ECG-based cardiorespiratory IoT sensor. In order to improve the robustness of our IoT-based sensor, we discuss in detail the influence of electrodes placement and nature. Performance assessment of our sensor resulted in a best-case sensitivity of 99.95 % and a precision of 99.89 % for an abdominal positioning of wet electrodes, while a sensitivity of 99.47 % and a precision of 99.31 % were observed using a commercial-grade dry electrodes belt. Consequently, we prove that our sensor is fit for the comfortable medical-grade monitoring of the cardiorespiratory activity in order to provide insights of patients health in a telemedicine context.