Keywords: Motor learning, neural control, and neuromuscular systems
Abstract: Understanding how actions are learned through trial and feedback, requires mechanistic insight into how self-paced actions are initiated, how they can be selected/initiated again, and how feedback can shape their execution and organization. We used behavioral, genetic, electrophysiological, and optical approaches to uncover that dopaminergic neurons are transiently active before self-paced movement initiation. This activity is not action-specific and modulates both the probability of initiation and the vigor of future movements, but does not affect ongoing movement. Dopamine is supposed to have opposite effects on downstream striatal direct and indirect pathways. Contrary to what is classically postulated, we found that both striatal direct and indirect pathways are active during movement initiation. The activity in both pathways is action-specific, organized in specific spatiotemporal patterns, and has complementary but different roles in movement initiation. Furthermore, when animals organize their individual movements in sequences or chunks, activity related to the initiation or termination of these chunks emerges in dopaminergic and striatal circuits. The behavioral and neuronal re-organization that accompanies sequence learning requires plasticity between the cortex and striatum. Finally, using closed-loop brain machine paradigms, we revealed that this plasticity is necessary to select, reinforce and shape specific neural that lead to desirable outcomes.

Keywords: Motor learning, neural control, and neuromuscular systems, Neuromuscular Systems - Learning and adaption, Brain Physiology and Modeling - Neural circuits
Abstract: Stable and efficient locomotion requires precise coordination of whole-body movements. Learned changes in interlimb coordination can be induced by exposure to a split-belt treadmill that imposes different speeds under each side of the body. This form of motor learning, known as locomotor adaptation, is used as a rehabilitation therapy in human patients. Here we investigated neural circuit mechanisms underlying locomotor adaptation in mice. We developed a transparent split-belt treadmill that provides high resolution readouts of locomotor behavior using our previously described, noninvasive, markerless LocoMouse tracking system [1]. Quantitative behavioral analysis revealed that mouse locomotor adaptation is specific to measures of interlimb coordination and has spatial and temporal components that adapt at different rates. The remarkable similarities between human and mouse locomotor adaptation suggest that this form of locomotor learning is highly conserved across vertebrates. Further, using a variety of genetic and lesion approaches, we demonstrate that split-belt adaptation in mice specifically depends on intermediate cerebellum, but is insensitive to large lesions of cerebral cortex. Finally, cell-type specific chemogenetics combined with quantitative behavioral analysis reveal distinct neural circuit mechanisms underlying spatial vs. temporal components of locomotor adaptation.

Keywords: Neurorehabilitation, Neurological disorders - Stroke, Motor learning, neural control, and neuromuscular systems
Abstract: Stroke is one of the leading causes of disability in world. Here we provide a rationale for a new immersive gaming approach to arm rehabilitation that focuses on movement quality rather than task completion. A pilot study shows that the new approach yielded better motor recovery than standard of care.

Keywords: Neurological disorders - Epilepsy, Clinical neurophysiology, Neural signal processing
Abstract: The aim of this study was to evaluate whether transfer-learning with pre-trained deep convolutional neural networks (deep CNNs) can be used for assessing patient outcomes in epilepsy. Transfer-learning with the GoogLeNet InceptionV3 CNN model pre-trained on the large ImageNet dataset (~1.2 million images) was able to differentiate upper (n=12) and lower (n=9) response quartile mesiotemporal lobe epilepsy patients in the NeuroPace® RNS® System clinical trials with ~76% classification accuracy based on chronic ambulatory baseline electrocorticographic (ECoG) data. These promising findings justify further research using deep CNNs for assessing patient outcomes in epilepsy.

Keywords: Brain-computer/machine Interface, Brain-Computer/Machine Interface - Robotics applications, Human performance
Abstract: Electroencephalography (EEG) based brain-computer interfaces (BCIs) have attracted increasing attention in the last decade. Previous studies have shown promising potential of using BCIs to provide alternative communication channels between human brains and external devices for people with disabilities. However, for able-bodied people with both hands occupied, the feasibility of using BCIs to provide additional communication channels to augment multitasking capability still needs further exploration. This study attempted to use the steady state visual evoked potential (SSVEP)-based BCI to provide an additional “hand” to cope with more tasks for operators under multitasking when their hands were fully occupied. The performance metrics of multitasking and SSVEP-BCI were analyzed. The results showed that it had little to no impact on the performance of multitasking to use SSVEP-BCI concurrently. On the other hand, the effect of multitasking on the performance of SSVEP-BCI tended to depend on the operators. For some operators, multitasking had very little impact on SSVEP-BCI, while for the others the impact of multitasking tended to be stronger. These findings show the potential of using SSVEP-BCI to perform additional tasks for the operators whose hands are fully occupied by multitasking.

Keywords: Human Performance - Cognition, Human Performance - Attention
Abstract: We present an Interactive System for Language and Communication Stimulation (ISLACS) directed to young children. The system is a technological educational application intended to stimulate learning abilities, developed as a native mobile application for devices with Android operating system. It is specially designed for children between 2 and 4 years of age, to stimulate language, contingency and attention, through cognitive tasks in the form of interactive mini-games. The activities teach the association between images with the corresponding infrequently used spoken words, and images of letters with the corresponding sounds (i.e. phonemes). The mini-games are based on multimedia resources, mainly videos of an educator, real objects and animals. The software evaluates progress by measuring interaction parameters such as correct answers and response time and has an algorithm to randomize the trials across children. The application is tested in 314 2-4 year-old children. Each child played a minimum of 8 sessions every other day, each one presenting 6 words and 4 letters, and testing for 3 words and 2 letters. Results showed that ISLACS is an attractive game for young children, which significantly succeeded to teach children the sounds of the letters and a repertory of nearly 48 infrequent words in Spanish, in a brief intervention.

Keywords: Motor Neuroprostheses - Prostheses, Neuromuscular Systems - Peripheral mechanisms, Neural Interfaces - Implantable systems
Abstract: Osseointegration, which describes the direct biological connection between bone and metal, has emerged as a promising method of increasing function and mobility in amputees. This technology holds significant potential, including the ability to harness an integrated neuromuscular interface to allow for volitional motor and sensory control of an external neural prosthesis. While orthopaedic osseointegration has been underway internationally for many years with a variety of implant systems, the practice is still in its early phases in the United States. Here we present the early outcomes of the first American trial for the Osseoanchored Prosthesis for the Rehabilitation of Amputees (OPRA). Our study demonstrates that one year following osseointegration, patients experience a significant improvement in function and health-related outcomes.

Keywords: Neural signal processing, Neural Signal Processing - Nonlinear analysis, Neural Signal Processing - Time frequency analysis
Abstract: In this paper, an unsupervised machine learning technique is proposed to localize Subthalamic Nucleus (STN) during deep brain stimulation (DBS) Surgery. DBS is one of most common treatments for advanced Parkinson's disease (PD). The purpose of this surgery is to permanently implant stimulation electrodes inside the STN to deliver electrical currents. It is clinically shown that DBS surgery can significantly reduce motor symptoms of PD (such as tremor). However, the outcome of this surgery is highly dependent on the location of the stimulating electrode. Since STN is a very small region inside the basal ganglia, accurate placement of the electrode is a challenging task for the surgical team. During DBS surgery, the team uses Micro-Electrode Recording (MER) of electrophysiological neural activities to intraoperatively track the location of electrodes and estimate the borders of the STN. In this work, we propose a composite unsupervised machine learning clustering approach that is capable of detecting the dorsal borders of the STN during DBS operation. For this, MER signals from 50 PD patients were recorded and used to validate the performance of the proposed method. Results show that the approach is capable of detecting the dorsal border of STN in an online manner with an accuracy of 80% without using any supervised training.

Keywords: Neuromuscular Systems - Neurorehabilitation
Abstract: Targeted muscle reinnervation and end-to-side coaptation are surgical techniques to reconstruct neural-function of residual peripheral nerves. These techniques have made some clinical achievements. The present study investigate the effect of sensory function reconstruction of median nerve and medial cutaneous nerve of the forelimb by combining targeted muscle reinnervation and the end-to-side coaptation techniques. The results showed that the signals presented different level fluctuation correlated with the tactile stimulation. The hematoxylin-eosin staining showed that the tactile corpuscles were remaining following the combined surgery. These results demonstrate the feasibility of combining the techniques and the rat model.

Keywords: Neurological disorders - Stroke, Neuromuscular Systems - EMG models, processing and applications, Neurorehabilitation
Abstract: Post-stroke physical training resulting in fatigue may affect motor rehabilitation. In this study, we compared the effects of fatigue-controlled and forced treadmill running on motor recovery based on a rat intracerebral hemorrhage (ICH) model. Twelve Sprague-Dawley rats with ICH received electromyography (EMG) electrodes implantation in the gastrocnemius muscle in the affected hindlimb. They were randomly distributed into three groups: control (n=4), forced exercise (n=4) and fatigue-controlled (n=4) groups. The training intensity in the fatigue-controlled exercise was monitored by calculating the real-time mean power frequency (MPF) of EMG. The training intervention started from forty-eight hours after ICH surgery. Modified neurological severity score was applied daily during the following 13-day intervention to evaluate motor recovery. The results showed that fatigue-controlled group achieved the best motor recovery compared with the other two (P < 0.05).

Keywords: Brain-computer/machine Interface, Neurorehabilitation - Wearable systems, Neural signal processing
Abstract: Brain-computer interface (BCI) has been emerging as an assessment tool for patients with disorder of consciousness (DOC). With the advantages of high time resolution, low cost and portable design, EEG based BCI systems are especially suitable for bedside measurement. Recent studies showed the successful application of an EEG based BCI on DOC assessment and communication. However, the effect of repeated BCI measurement on this patient group is not clear. In this study, a tactile BCI paradigm was repeated 12 runs for 10 consecutive days on 5 DOC patients. Although the BCI performance varied among runs and days, every patient reached at least once the accuracy above 60%. Moreover, the Coma Recovery Scale Revised improved on two out of the five patients. This study addressed the significance of repeating a tactile BCI on DOC patients, and indicates a promising recovery effect of a tactile BCI on DOC patients.

Keywords: Neuromuscular Systems - EMG models, processing and applications, Neuromuscular Systems - Computational modeling and simulation
Abstract: Difficulties accessing amputee populations has resulted in the widespread adoption of able-bodied subjects in virtual environments for the development of myoelectric prostheses. Factors such as scar tissue, different physiologies or surgical outcomes, and reduced visual and proprioceptive feedback, however, may contribute to differences in electromyogram (EMG) patterns between these groups. As such, studies have consistently found worse results when comparing the performance of amputee subjects to that of their able-bodied counterparts under the same conditions. To identify the source of this performance degradation, a topology-based data analysis method, called Mapper, was employed to visualize the ``shape'' of EMG feature spaces derived from amputee and able-bodied subjects. The information content of amputee EMG features was found to differ from those of non-amputee subject in three ways: 1) the loss of nonlinear complexity and frequency information, 2) the loss of time-series modeling information, and 3) the segmentation of unique information. The empirical effects of these differences were visualized by classifying motion classes using consistent and migratory features from functional feature groups. In summary, this work characterized inconsistencies in EMG features between amputee and able-bodied populations by theoretical means, highlighted the empirical effects when these are ignored, and proposed a solution for future studies with able-bodied subjects.

Keywords: Neurological disorders - Stroke, Neurorehabilitation, Neural Signal Processing - Time frequency analysis
Abstract: Alpha oscillatory activity and its dynamics have a key role in motor and sensory functions. Stroke affects different brain structures, which can result in pathological changes in alpha oscillations. We studied the relationship between the amplitude of alpha oscillations in resting state and their modulation during the attempt of movement in 37 patients with severe paralysis after stroke. As previously observed in healthy subjects, resting-state alpha activity significantly correlated with the alpha event-related desynchronization (ERD) during the attempt of movement of the paralyzed hand. Further, alpha ERD correlated with the presence or absence of damage in cortical structures, but resting-state alpha power did not. This result provides new insights on the understanding of the brain changes after stroke, which may help in future therapies to help the patients to recover their lost motor function.

Keywords: Brain Functional Imaging - Connectivity and Network, Neural signal processing, Neurological disorders
Abstract: This study explores neural connectivity in resting state through coherence and spectral graph based methods across large populations with electroencephalography (EEG). Using the Neural Engineering Data Consortium (NEDC) EEG Corpus we extract EEG data in a 10-20 montage and accompanying patient characteristics. Non-medicated subjects with clinically normal EEG are used as the normative population (n=1,167) while a group with a similar age distribution of medicated subjects with clinically abnormal EEG are used as the abnormal population (n=2,940). Parameters and properties of spectral coherence connectivity graphs are computed across frequency bands. We establish default mode networks (DMN) for the different populations on several frequency bands. We find that frequency bands differ across the populations more than specific graph properties. However, we find that there is an increased level of connectivity in the abnormal population. These results may lead to neural connectivity based diagnostic aides.

Keywords: Brain Functional Imaging - EEG and Evoked Potentials, Human Performance - Attention, Brain Functional Imaging - Source localization
Abstract: This study developed a method for investigating mind wandering (MW) in a fatigue-inducing motor task. To develop research protocols to assess fatigue in clinical cohorts, it is important that participants perform the task at hand to the best of their ability and with their complete attention. Therefore, it is important to know if the participant fails in the task due to fatigue or lack of sustained attention as a result of MW. Two cohorts of 12 healthy controls and Chronic Fatigue Syndrome (CFS) subjects performed a hand-grip fatigue-inducing motor task while EMG and EEG were simultaneously recorded. Frontal midline theta (FMT) and parietal alpha power were calculated throughout the task. While no significant differences were obtained in the FMT power for both cohorts, significant differences in parietal alpha power for the healthy subjects across the task shows they may have experienced MW unlike CFS subjects, who had to put in consistent effort to sustain attention during the task, which could imply fewer MW events. Assessing MW using EEG can serve as an objective marker for evaluating performance in a task and, for assessing the impact of fatigue on the ability to sustain attention.

Keywords: Neural signal processing, Neurological disorders - Epilepsy, Brain Functional Imaging - Classification, spatiotemporal dynamics
Abstract: Brain functional connectivity has been used to investigate the interaction between brain regions. It provides important information related to brain diseases, injuries, and high level cognitive functions. Statistical methods have been widely used to model brain functional connectivity based upon which insights of brain function are expected to be revealed. Most statistical approaches were developed based upon an assumption that connectivity patterns are static during the recording. This is not true because the connectivity changes over time. A dynamical modeling of connectivity patterns allows to characterize these variations. In this work, a simplified dynamic Bayesian modeling approach, parallel Hidden Markov Model (PaHMM), was investigated by characterizing temporal variations of cortical functional connectivity patterns computed using epileptic electroencephalogram (EEG) data. The performance of the PaHMM was evaluated based on an experimental study of epilepsy detection and classification, where multisubject epileptic EEG data from Temple University Hospital EEG Data Corpus were used. Experimental results show that an accuracy of 93.5% was obtained for the epilepsy detection, and an overall accuracy above 81% was achieved for the seizure type classification. This indicates that the method can efficiently capture temporal variations of functional connectivity patterns, and is potentially applicable in clinical settings to detect epilepsy and differentiate seizure types.

Keywords: Neurorehabilitation - Wearable systems, Neurological disorders, Human Performance - Sensory-motor
Abstract: Childhood dystonia has been associated with injury to the basal ganglia, however there is evidence suggesting the involvement of sensory cortex, cerebellum and brainstem. Even though dystonia is considered a movement disorder, recent studies have shown dysfunctional sensorimotor integration that further contributes to the dystonic symptoms. Such aberrant circuitry may prevent children with dystonia from acquiring new motor tasks. The use of EMG-based biofeedback has been proposed as a promising technique to augment sensory information and consequently improve motor function. The aim of this study is to test the effects of a newly designed vibrotactile EMG-based biofeedback device to induce changes of muscle patterns in children with dystonia during a continuous figure-eight task. We show a change in muscle activation task components when participants receive the biofeedback while performing the task. Those changes suggest new neuromotor solutions in the framework of “motor exploration” as a strategy in the early phases of motor learning.

Keywords: Neurorehabilitation - Neurofeedback, Brain Functional Imaging - EEG and Evoked Potentials
Abstract: Meditation is a mental practice to achieve focus of mind and emotional clarity. Meditation has been used for cognitive enhancement, rehabilitation and reducing stress and anxiety. In the present study, we are doing a comparative analysis between various levels of meditators based on EEG as psychophysiological indicator; and possibility of EEG as a neurofeedback for meditators. An analytical experiment on three categories of subjects (A: an expert meditator, B: five moderate meditators and C: five non-meditators) was done. Each subject was guided to perform two visual tasks; first to sit relaxed with eyes closed (REC) and second to gaze on a dot on screen (RDOT); supplied, EEG being recorded in parallel. The first subject was recorded with absorbed state of meditation (Samādhi). For psychophysiological analysis, wavelet transform based features from each recording of EEG was evaluated. Topographical mapping of brain functioning based on features were plotted and analyzed. It was observed that theta, alpha and beta were comparatively higher for expert meditator in frontal and central region during REC and RDOT. Also, during absorbed meditative state, the alpha and beta are higher at midline central region (Cz) and theta is higher at C3 and C4.

Keywords: Human Performance - Fatigue, Human Performance - Attention, Human Performance - Cognition
Abstract: Mental workload (MW) assessment is important for numerous mentally-demanding applications, including first responders, air traffic control, amongst others, as it quantifies the cognitive capabilities of the operator. Recently, there has been a push for wearables based MW monitoring for real-time feedback and human performance augmentation. Most previous studies have focused on immobile subjects. Realistic applications, however, rely on ambulant users under varying types and levels of physical activity. Movement artifacts are known to hamper the quality of signals measured by wearable devices, thus the impact on MW assessment in situ is still unknown. In this study, we compare the performance of several automated artifact removal algorithms for electroencephalograms (EEG), as well as the robustness of two classical feature sets, for MW assessment under varying physical activity levels.

Keywords: Neurological disorders, Clinical neurophysiology, Human Performance - Sensory-motor
Abstract: Non-motor symptoms of Parkinson’s disease (PD) are present in all stages of the disease, significantly affecting patient quality of life. Some previous work on temporal perception has seen abnormalities occurring in PD, highlighting the Basal Ganglia’s (BG) role on this perception. However, these studies have not considered patient perceptual ability based on the tested time scale, even though the BG's postulated influence on temporal processing is limited to certain time scales. Furthermore, it is not clear what effect Levodopa medication has on temporal perception for PD patients. This study examines the perception of vision-based temporal perception in different time scales for PD patients, and the effect of Levodopa medication via a two-forced alternative choice task using a computer-generated graphical tool. For this, perceptual ability (quantified using the subject's difference threshold obtained through cumulative Gaussian functions) of 21 patients with PD was evaluated OFF and ON Levodopa medication, compared to 17 age-matched healthy participants. Individuals with PD displayed no impairments in perceiving time in the range of milliseconds, however in the range of seconds temporal perception was significantly impaired. This provides evidence for current timing models involving cerebellar control for millisecond timing, and BG influenced timing in the range of seconds to minutes. Furthermore, Levodopa showed no significant effects on visual temporal discrimination.

Keywords: Neurorehabilitation, Neurorehabilitation - Robotics, Brain-computer/machine Interface
Abstract: Currently, one of the challenges in EEG-based brain-computer interfaces (BCI) for neurorehabilitation is the recognition of the intention to perform different movements from same limb. This would allow finer control of neurorehabilitation and motor recovery devices by end-users [1]. To address this issue, we assess the feasibility of recognizing two self-paced movement intention of the right upper limb plus a rest state from EEG signals recorded during robot-assisted rehabilitation therapy. In addition, the work proposes the use of Multi-CSP features and deep learning classifiers to recognize movement intentions of the same limb. The results showed performance peaked greater at (80%) using a novel classification models implemented in a multiclass classification scenario. On the basis of these results, the decoding of the movement intention could potentially be used to develop more natural and intuitive robot assisted neurorehabilitation therapies.

Keywords: Neurological disorders - Diagnostic and evaluation techniques, Neural signal processing, Motor learning, neural control, and neuromuscular systems
Abstract: Resting tremor affects voluntary reaching movements in patients with Parkinson’s disease (PD). Previously, we found that cutaneous afferents evoked by electrically simulating the superficial radial nerve could inhibit tremor activity [3]. Yet, the mechanism that cutaneous stimulation affects tremor as well as voluntary movements remains unclear. In this paper, we used the method of muscle synergy analysis to further investigate the muscle activation patterns of tremor and reaching movements. Three PD patients with tremor dominant symptom participated in this preliminary study. The patients performed targeted reaching movements with no visual feedback of their hands, while transcutaneous electrical stimulation was delivered to inhibit tremor activity. Kinematics and EMGs of six muscles were recorded during resting and reaching tasks. Muscle synergy analysis was performed with data during resting and task stages with and without cutaneous stimulation. Results showed that during resting state the time profiles of muscle synergy in the tremor cycle displayed two main alternating components that were distributed to antagonistic muscle pairs in upper arm, forearm and wrist. Cutaneous stimulation affected the vector patterns and time profiles of muscle synergies for tremor and reaching movements differently. Our results indicate that muscle synergies are sensitive to central descending drives of muscles, as well as evoked peripheral afferents.

Keywords: Neural signal processing, Brain Stimulation-Deep brain stimulation, Neural Interfaces - Neural stimulation
Abstract: Invasive and non-invasive electrical stimulation are increasingly being used for the diagnosis and treatment of neurological disorders, and for characterizing neural circuits involved in a range of behaviors. However, there are substantial challenges in understanding the effects of stimulation on brain activity due to contamination of electrophysiological recordings by electrical stimulation artifacts. Here, we compare the performance of several artifact removal methods on electrocorticographic (ECoG) recordings with simultaneous cortical or peripheral stimulation in humans. We systematically evaluated the effects of stimulation modality, stimulation frequency, and neural recording frequency on the ability to reconstruct neural activity amplitude and phase data. We found that no single method was most effective for all situations, however it was possible to reconstruct key neural data features in every case. The development of optimized artifact removal procedures will facilitate clearer understanding of the biological effects of electrical stimulation and allow for improved therapeutic applications.

Keywords: Brain-Computer/Machine Interface - Biofeedback, Brain Stimulation-Deep brain stimulation, Neural Interfaces - Implantable systems
Abstract: Increased oscillatory activities in the beta frequency band (13-30 Hz) in the subthalamic nucleus (STN), and in particular prolonged episodes of increased synchrony in this frequency band, have been associated with motor symptoms such as bradykinesia and rigidity in Parkinson’s disease (PD). Numerous studies have investigated sensorimotor cortical beta oscillations either as a control signal for Brain Computer Interfaces (BCI) or as target signal for neurofeedback training (NFB). However, it still remains unknown whether patients with PD can gain control of the pathological oscillations recorded from a subcortical site – the STN – with neurofeedback training. We tried to address this question in the current study. Specifically, we designed a simple basketball game, in which the position of a basketball changes based on the occurrence of events of temporally increased beta power quantified in real-time. Participants practised in the game to control the position of the basketball, which requires modulation of the beta oscillations recorded from STN local field potentials (LFPs). Our results suggest that it is possible to use neurofeedback training for PD patients to downregulate pathological beta oscillations in STN LFPs, and that this can lead to a reduction of beta oscillations in the cortical-STN motor network.

Keywords: Neural Signal Processing - Time frequency analysis, Brain Functional Imaging - EEG and Evoked Potentials, Human performance
Abstract: Exacerbation monitoring of obstructive sleep apnea (OSA) is important for the evaluation of treatment effectiveness and tracking the disease progression. In this study, we investigated the use of spectral features from single channel electroencephalography (EEG) for early detection and monitoring of OSA exacerbation using the Sleep Health Heart Study dataset. We have explored 22 features at different sleep stages corresponding to different frequency bands to distinguish 410 subjects in the stable and exacerbation groups. An optimal set of 15 features has been selected using the recursive feature elimination technique. It has been found that these features provide significant discriminative information (p-value ≤ 0.05) for classification. On the test dataset of 82 EEG records, a classification accuracy, sensitivity, and specificity of 79.17%, 80.85%, and 76.00%, respectively have been achieved using a Quadratic Discriminant Analysis classifier. Results demonstrate that OSA exacerbation can be detected early and monitored with this simple yet effective method using a single channel EEG.

Keywords: Motor neuroprostheses, Neurorehabilitation, Neuromuscular Systems - Locomotion, posture and balance
Abstract: Paralysis of the upper extremity (UE) and lower extremity (LE) following cervical spinal cord injury (SCI) significantly impairs the individual’s ability to live independently. While improvement in hand function is considered one of the most desired functions in tetraplegics, few therapies have been successfully developed. Spinal cord stimulation have been shown to be effective in enabling UE and LE function individually. However, improvement of both UE and LE function concomitantly, using the same therapeutic modality has not been demonstrated. This study demonstrates that noninvasive neuromodulation of the spinal cord can improve hand function and stepping abilities in an individual with severe chronic paralysis. The improvements in UE and LE function were accompanied by improvement in autonomic function and sensation below the level of the lesion. Regaining these functions in the individual with severe, chronic paralysis had dramatic impacts functionally and psychologically.

Keywords: Neural Interfaces - Neural stimulation, Neural signal processing
Abstract: Transcranial alternating current stimulation (tACS) has been known as a well-established non-invasive electrical stimulation method of human cortex. A number of researches on the efficiency of tACS discovered that it could directly modulate human cortex function by applying of oscillatory currents on the human scalp. Until now, the electrophysiological evidence of tACS remains unclear. In this study, we delivered α-tACS or sham stimulation over the prefrontal cortex of 20 healthy participants and recorded the rest EEG before and after the stimulation to calculate alpha power. The results showed that α-power of both fixed and individual α-band elevated after α-tACS, while the phenomenon did not appear in sham group. We could draw a conclusion that α-tACS could enhance α-oscillation and modulate rhythmic brain activity specifically.

Keywords: Neural signal processing, Neurological disorders - Epilepsy, Brain Functional Imaging - EEG and Evoked Potentials
Abstract: High-frequency oscillations (HFOs) are considered reliable markers of epileptic tissues. Digital filter is widely used to extract the HFOs, but it also introduces fake oscillations because of Gibbs effects. In this paper we propose a pipeline for HFOs extraction using ensemble empirical mode decomposition (EEMD) method. Stereo-Electroencephalograph (SEEG) is decomposed into intrinsic mode functions (IMFs). Rhythmical index (RI) of detected HFOs is calculated, which is corresponding to normalized angular acceleration of oscillation. The ictal SEEG is processed with nine seizures recorded from three patients, who underwent surgical intervention and got seizure free. Results show that proposed pipeline can extract HFOs from ictal SEEG automatically. RI values of HFOs in resected area are higher than non-resected area. The proposed pipeline with RI measure may serve as an effective way for extracting pathological HFOs.

Keywords: Motor learning, neural control, and neuromuscular systems, Neural Interfaces - Recording, Neurorehabilitation
Abstract: Internal action selection is an important motor control, in which patients with Parkinson's disease (PD) generally show deficiencies. Basal ganglia (BG) is proved to play an important role in decision-making and act as a specialized internal selection device within the vertebrate brain architecture. Furthermore, some studies showed there was a close relationship among striatal dopamine signaling, action selection and time interval by training mice to perform an internal selection task. However, the neural mechanism of the internal action selection is still unclear. In this study, we setup a platform for psychophysical and electrophysiological study and recorded behavioral data from normal human subjects and primates when they performing an internal action selection task. The results showed that longer trial intervals led to longer action transition time, which indicates the time interval biases internal action selection, and the effect of movement direction was not significant. Furthermore, we recorded the task-related neuronal activity in primate’s primary motor cortex (M1). Preliminary data showed there were significant firing rate changes in M1 at the transition of action selection.

Keywords: Neurorehabilitation - Robotics, Neurorehabilitation - Virtual reality, Neurological disorders
Abstract: Visuomotor feedback and its impact on performing and learning movements is an extensively studied field, both through the use of experiments under different types of visuomotor feedback, as well as through neurophysiological studies. Neurorehabilitation of the upper-limb relies heavily on repetitive targeted movements and in recent decades, the introduction of instrumented and robotic devices coupled with computer screens have substituted the existing direct visual feedback of traditional practice with an indirect feedback. However, the impact of such a shift has not been studied. Putting in perspective the literature on these different aspects, this paper shows that there seems to be little indication that the feedback type may significantly affect the neurorehabilitation outcomes. Nevertheless, despite the intrinsic difficulties in directly observing the effects of the introduction of indirect visual feedback in neurorehabilitation practices, it is of interest to investigate further this specific aspect of the newly introduced technologies.

Keywords: Neurological disorders, Neurological disorders - Diagnostic and evaluation techniques, Neurological disorders - Stroke
Abstract: White matter lesions detected on magnetic resonance imaging scans of the brain have been hypothesized to have associations with the causes of several diseases. Accurate quantification of white matter lesions is important for the hypothesis validation. However, the clinical quantification is highly variable due to subjective opinions of different raters and is likely to compromise the reliability of the assessment. This paper introduces a new method of two-dimensional fuzzy weighted recurrence networks that can numerically express the quantity of white matter lesions. The results illustrate the promising application of the proposed method that offers as a useful computational tool in brain research.

Keywords: Brain Functional Imaging - EEG and Evoked Potentials, Neuromuscular Systems - Neurorehabilitation, Human Performance - Attention
Abstract: Current brain-machine-interface (BMI) rehabilitation approaches typically focus on a specific aspect of neural activity. Auxiliary signals, derived from independent measures of neural activity and recorded in parallel might be useful in quantifying and tracking a subjects mental state and performance. In this work, we demonstrate that event-related potentials can be reliably observed in stroke survivors with chronic paralysis during a BMI intervention. The averaged evoked response remains stable over sessions and varies between subjects. A prominent negativity, positivity complex emerges whose features can be tracked across subjects and sessions.

Keywords: Neurorehabilitation - Wearable systems
Abstract: This paper describes a wearable intraoral system for speech therapy using artificial sensory feedback timed with the undesired tongue movement. The system has been implemented as a custom-made palatal retainer, which includes two optical distance sensors as proximity sensors and two stimulators to provide sensory error feedback to the tongue. Participants wore the palatal retainer to test the system performance. By testing the system with phonetic targets /t/ and /d/, we showed that the system was able to detect the tongue movement during pronunciation. The system was also tested to see if it can help a non-native English speaker with Hindi mother tongue to correct the pronunciation of alveolar consonant /t/. The subject was asked to read multiple words containing /t/ consonants repeatedly, with distracting words in between to minimize the involvement of cognition or intentional correction. Test results indicated that the error feedback via stimulation helped the subject to move his tongue forward towards the alveolar ridge in pronouncing /t/. The result suggests that intrinsic sensory feedback can be an effective way to train non-native speakers to correct their pronunciation.

Keywords: Neurological disorders, Neural Interfaces - Sensors and body Interfaces, Neural Interfaces - Recording
Abstract: Background: Since 1988, one of the most relevant criteria that support diagnosis of Parkinson Disease (PD) is the response to levodopa. Levodopa effect is evaluated using clinical scales such (MDS-UPDRS). Although this approach is informative its assessment is highly subjective. Objectives: quantify the motor response to levodopa in patients with Parkinson's disease using a portable and inexpensive RGB-D camera (Kinect) Methods: 11 PD patients were included. Descriptive analysis were used for clinical variables. Spearman´s rank correlation was used to correlate the age and the gait variables. Wilcoxon matched-pairs signed-rank test was used for the gait variables comparison between ON and OFF states. Results: In ON state PD patients showed increments in global speed [p=0.007], ankle speed [p<0.001], ankle stride length [p=0.001], arm swing speed [p=0.008], and arm swing magnitude [p=0.014], these results slightly changed when the sample was analyzed taking into account the individual behavior of each limb. In comparison with the upper limbs, levodopa response was more uniform in the lower limbs. Conclusion: As expected, in ON state PD patients improve their speed and their limbs range of movement during gait. Levodopa response can be measured within the medical consultation using a RGB-D camera.

Keywords: Neuromuscular Systems - EMG models, processing and applications, Neurorehabilitation - Robotics, Brain-computer/machine Interface
Abstract: Modern approaches to motor rehabilitation of severe upper limb paralysis in chronic stroke decode movements from electromyography for controlling rehabilitation orthoses. Muscle fatigue is a phenomenon that influences these neurophysiological signals and may diminish the decoding quality. Characterization of these potential signal changes during movement patterns of rehabilitation training could therefore help improve the decoding accuracy. In the present work we investigated how electromyographic indices of muscle fatigue in the Deltoid Anterior muscle evolve during typical forward reaching movements of a rehabilitation training in healthy subjects and a stroke patient. We found that muscle fatigue in healthy subjects changed the neurophysiological signal. In the patient, however, no consistent change was observed over several sessions.

Keywords: Neurological disorders - Epilepsy, Brain Functional Imaging - Connectivity and Network, Brain Functional Imaging - EEG and Evoked Potentials
Abstract: In order to investigate the alterations of brain network in children with epilepsy during the interictal and ictal periods, partial directed coherence (PDC) was employed as a measure of causality to analyze 22 electroencephalography (EEG) datasets recorded from 10 focal seizure children in this study. Functional brain network during interictal and ictal periods were constructed based on the computed PDC values, from which two graph-based measures, including the degree and clustering coefficient were extracted to assess the functional connectivity in seizure-linked network. Results showed that, compared to interictal period, the regional degree at the center lobe in delta band during the ictal period was significantly reduced. On the contrary, the clustering coefficients in delta band during the ictal period were significantly increased in the frontal, parietal, and temporal lobes. Our findings therefore suggest that ictal state may affect the visual, physical, mental, auditory, and other functions in epileptic children, providing a new perspective to explore the brain network alterations in children with epilepsy.

Keywords: Brain-computer/machine Interface, Brain-Computer/Machine Interface - Robotics applications
Abstract: Detection of human cognitive states using biosignals such as the electroencephalogram (EEG) is gaining relevance in different application areas, e.g., teleoperation, human-robot collaboration, and rehabilitation. Especially, the P300, which is evoked as an event-related potential (ERP), when humans perceive task-relevant infrequent events among task-irrelevant frequent events, is widely used in brain-computer interfaces (BCIs). P300 detection has been used as an indicator that a human perceives task-relevant events or detects the occurrence of task-relevant or important events. In this paper, we focus on not only perceived task-relevant events but also not-perceived task-relevant events (i.e., missed events). In fact, a human can miss task-relevant events for different reasons, e.g., reduced attention level or increased workload level during parallel task-processing situations among others. Moreover, a human can also intentionally ignore task-relevant events to manage several simultaneous tasks. However, such missed events do not often occur in real-world applications. In this paper, we propose a transfer approach to handle insufficient number of events for training a classifier. For example, task-irrelevant infrequent events are used for training of classifier to detect missed task-relevant events. We evaluated our approach in different settings of training and testing a classifier with and without classifier transfer.

Keywords: Brain-computer/machine Interface, Neural signal processing
Abstract: The problem of detecting changes in firing patterns in neural data is studied. The problem is formulated as a quickest change detection (QCD) problem. Important algorithms from the literature are reviewed. A new algorithmic technique is discussed to detect deviations from learned baseline behavior. The algorithms studied can be applied to both spike and local field potential data. The algorithms are applied to mice spike data to verify the presence of behavioral learning.

Keywords: Neural Interfaces - Sensors and body Interfaces, Motor neuroprostheses
Abstract: Improved neural interfacing strategies are needed for the full articulation of advanced prostheses. To address limitations of existing control interface designs, the work of our laboratory has presented an optical approach to reading activity from individual nerve fibers using activity-dependent calcium transients. Here, we demonstrate the feasibility of such signals to control prosthesis actuation by using the axonal fluorescence signal in an in vitro mouse nerve to drive a prosthetic digit in real-time. Additionally, signals of varying action potential frequency are streamed post hoc to the prosthesis, showing graded motor output and the potential for proportional neural control. This proof-of-concept work is a novel demonstration of the functional use of activity-dependent optical read-out in the nerve.

Keywords: Neural Interfaces - Neural stimulation, Neurorehabilitation - Neurofeedback, Neurological disorders
Abstract: Poor bladder management is a common and potentially life-threatening dysfunction among spinal cord injury (SCI) patients. In this condition, sensation from the bladder and voluntary control of micturition are lost, which might result in high post-void residual urine volume in the bladder, leading to renal impairment. Micturition can be driven using the sacral anterior root stimulator (SARS). However, commercially available SARS devices are not equipped with a closed-loop regulator for adaptive and automated control of bladder contractions. In our previous study, we developed a closed-loop control strategy for bladder emptying. In this paper we demonstrate the closed-loop neuromodulation feasibility in a SCI rat. The closed-loop strategy in this model achieved 71% voiding efficiency, higher than 40% efficiency obtained using open-loop stimulation. Our results provide a basis for developing an implantable closed-loop neural bladder prosthesis for SCI patients in the future.

Keywords: Brain-computer/machine Interface, Neural Interfaces - Recording, Neural signal processing
Abstract: Brain Computer Interfaces using motor imagery have been explored during many years. Auditory imagery,nonetheless, has been a rarely explored approach, but a one that might open new possibilities in musical interface, communication and speech synthesis.Using a non-invasive BCI, open source software, and au-ditory imagery of white noise, we tested several classification algorithms to determine the accuracy and usefulness of using non-motor imagery EEG signals.We tested 15 healthy adults with a 6 electrode EEG setup and the open source platform OpenVibe. Our results show that using a Support Vector Machine classifier and our experimental setup, we could achieve up to93% accuracy in white noise imagery in our subjects. Lin-ear Discriminant Analysis and MultiLayer Perceptron setups yielded accuracy of 73-76%.

Keywords: Brain-computer/machine Interface, Neural Signal Processing - Time frequency analysis
Abstract: Steady State Visual Evoked Potentials (SSVEP) based Brain Computer Interface (BCI) provides high throughput in communication. In SSVEP-BCI, typically, higher accuracy can be achieved with a relatively longer response time. It is therefore a research topic to reduce the response time while keeping high accuracy. We propose a new method, temporal alignments enhanced Canonical Correlation Analysis (TACCA), followed by a decision fusion to improve classification accuracy with short response time. TACCA exploits linear correlation with non-linear similarity between steady-state responses and stimulus frequencies. We compare TACCA and three state-of-the-art methods using data from 54-subjects with response time ranging from 0.5 to 4 seconds. The evaluation results show that TACCA yields mean significant accuracy increase of 10-30% in all segment lengths, especially for the shorter time segment. One-way ANOVA tests show high significant differences between single and multiple phases in TACCA performance.

Keywords: Brain-computer/machine Interface, Brain Functional Imaging - EEG and Evoked Potentials, Human Performance - Cognition
Abstract: We present a collaborative Brain-Computer Interface (cBCI) to aid group decision-making based on realistic video feeds. The cBCI combines neural features extracted from EEG and response times to estimate the decision confidence of users. Confidence estimates are used to weigh individual responses and obtain group decisions. Results obtained with 10 participants indicate that cBCI groups are significantly more accurate than equally-sized groups using standard majority. Also, selecting dyads on the basis of the average performance of their members and then assisting them with our cBCI halves the error rates with respect to majority-based performance while allowing most participants to be included in at least one of the selected dyads, hence being particularly inclusive. Results indicate that this selection strategy makes cBCIs even more effective as methods for human augmentation in realistic scenarios.

Keywords: Brain Physiology and Modeling - Neural circuits, Brain Functional Imaging - Connectivity and Network, Neurological disorders - Diagnostic and evaluation techniques
Abstract: Abstract— Spatiotemporal dynamics of brain are defined by the complex connectivity among different brain regions in three-dimensional environment. Here, we report a co-culturing platform where interconnection between 3D cultures from cortical and striatal dissociated cells can be conveniently tailored and manipulated. These 3D co-cultures were created in Polydimethylsiloxane (PDMS) mini-wells and soft-lithography was used to create micro-channels between them. We observed synchronous long burst activities in these cultures. The interconnection between cortical and striatal components of the co-culture were then severed physically to assess the effect of inter-connectivity. This platform can assist our understanding of brain plasticity, development and neural pathways.

Keywords: Brain-computer/machine Interface
Abstract: The previous visual P300-speller mostly used two-dimensional presentation on computer screens, while virtual reality (VR) technology can present stimulus in a virtual three-dimension (3D) space. This study intended to explore the effect of 3D presentation of visual stimuli on P300-speller in VR. Four different presentation paradigms, including three 3D paradigms and one two-dimension (2D) paradigm, were adopted. These 3D paradigms formed 3D visual stimuli by dynamically presenting depth information in character intensifications. The event-related potentials (ERP) features and spelling performance were analyzed, and the results showed that the electroencephalograph (EEG) class discriminant and accuracy of the 3D paradigm were superior to the 2D paradigm, and the two paradigms, Different depth information Array Character Flash (DACF) paradigm and Same depth information Array Character Jump (SACJ) paradigm were recommended.

Keywords: Brain-computer/machine Interface
Abstract: Informative patterns of neural data obtained from electroencephalography (EEG) can be classified by machine learning techniques to improve performance of human -computer interaction. A rapid serial visual presentation (RSVP)-based brain-computer interface (BCI) system relies on single-trial classification of event-related potentials (ERP) to categorize target and non-target images. The system works well in well-controlled laboratory settings; however, transitioning this approach into more dynamic, unconstrained environments poses several significant challenges. One major challenge is how to address the problem of session-to-session variability in EEG decoding. For a new session, a time-consuming and laborious calibration procedure is usually required to collect sufficient individual data for training a new classifier. This paper employed a subspace decomposition algorithm, Signal-to-noise ratio Maximizer for event-related potentials (SIM), to improve the session-to-session transfer performance of the RSVP-based BCI system. EEG data were collected from 17 subjects, each of whom performed two task sessions on two different days. Compared with the standard hierarchical discriminant component analysis (HDCA) algorithm, the classification performance was significantly improved by combining the SIM algorithm with the HDCA algorithm. The mean area under the receiver operating characteristic curve (AUC) across all subjects was improved from 0.7242 to 0.8546.

Keywords: Brain-computer/machine Interface
Abstract: Most existing steady-state visual evoked potential (SSVEP)-based brain-computer interfaces (BCIs) adopted low- or medium-frequency stimuli towards high information transfer rates (ITRs). However, the intense flashes will deteriorate the user experience. With the increase of stimulation frequency, the discomfort caused by flickering stimuli can be significantly reduced, but the amplitude of SSVEP is also decreased, which greatly increases the difficulty of SSVEP detection. To cope with the tradeoff between SSVEP intensity and user experience, this study proposed a new checkerboard stimulus scheme with spatially alternated flickering and background squares. This study compared black-white, black, and white-background checkerboard stimuli, together with a traditional single flickering stimulus in eliciting SSVEPs at different frequencies. The data analysis results showed that the signal-to-noise ratio (SNR) of SSVEP for the black-background checkerboard stimulus was comparable to the single flickering stimulus, both of which were significantly higher than the other two checkerboard stimuli in the high-frequency range (40 Hz and 60 Hz). Additionally, with a consensus of reduction in subjective visual irritation, the black-background checkerboard stimulus can lead to improved user experience of the SSVEP-based BCIs.

Keywords: Neural Interfaces - Neural microsystems and Interface engineering, Neural Interfaces - Recording, Neural Interfaces - Computational modeling and simulation
Abstract: Peripheral nerve decoding algorithms form an important component of closed-loop bioelectronic medicines devices. For any decoding method, meaningful properties need to be extracted from the peripheral nerve signal as the first step. Simple measures such as signal amplitude and features of the Fourier power spectrum are most typically used, leaving open whether important information is encoded in more subtle properties of the dynamics. We here propose a feature-based analysis method that identifies changes in firing characteristics across recording sections by unsupervised dimensionality reduction in a high-dimensional feature-space and selects single efficiently implementable estimators for each characteristic to be used as the basis for a better decoding in future bioelectronic medicines devices.

Keywords: Brain-Computer/Machine Interface - Robotics applications, Motor Neuroprostheses - Robotics, Sensory Neuroprostheses - Somatosensory and vestibular
Abstract: Closed-loop brain-machine interfaces may help restore the autonomy of amputees and tetraplegic patients. However, additional efforts are needed towards their real-world use with prostheses. Here we have interfaced a highly versatile closed-loop mouse BMI with an online model of a real-world prosthetic arm. We describe this setup and illustrate how it allows to explore the efficiency of different input and output coding strategies given a realistic modelling of the interactions between a commercial bidirectional prosthesis and its environment.

Keywords: Neural Interfaces - Neural stimulation, Neurorehabilitation, Neural signal processing
Abstract: Overactive bladder (OAB) is a chronic medical disorder affecting nearly 18% of the world’s adult population. Recently, we discovered a novel stimulation target; the saphenous nerve (SAFN) that can induce significant bladder-inhibitory effects in both pre-clinical and clinical experimental models. The aim of this paper is to further characterize this bladder inhibitory input via the SAFN afferents by modulating the stimulation parameters. To this end, acute experiments were conducted in 15 urethane anesthetized rats to examine the effects of increasing the duration of stimulation trials on multiple urodynamic parameters. The results of this study suggest that larger doses of SAFN stimulation elicit more potent bladder-inhibitory responses. Further work is needed to translate these findings in OAB patients.

Keywords: Neural Interfaces - Biomaterials, Neural Interfaces - Neural microsystems and Interface engineering, Neural Interfaces - Regeneration and tissue-electrode Interface
Abstract: Magnetic nanoparticles are a versatile tool to modulate calcium signaling, alter intracellular vesicle dynamics, or interfere with gene expression in cerebral neurons through imposing magnetic field gradients on the nanoparticles. However, a lack of understanding of the underlying mechanism, costly experimental magnetic setups and the complexity of magnetic gradient design currently hinder advancements and further integrations into drug studies and clinical translation. Here, we present a robust, low-cost magnetic platform, which is compatible with standard cell culture assays and Petri dishes, in combination with a fully integrated computation of the superimposing magnetic field and force maps. Utilizing the magnetic Petri dish platform, we designed and studied the impact of different magnetic field patterns on force-mediated neurite growth in dissociated primary rodent cortical neurons. We found that neurite growth re-orients strongest within a symmetric bidirectional magnetic gradient pattern without impairing neurite growth. Our magnetic Petri dish platform provides convenient means to extend magnetic force studies into tissue engineering, pharmaceutical, and translational studies, bringing a variety of benefits to medical neuroengineering.

Keywords: Neural Interfaces - Recording, Neural Interfaces - Computational modeling and simulation
Abstract: Acute neurophysiology in the behaving primate typically relies on traditional manufacturing approaches for the instrumentation necessary for recording. For example, our previous approach consisted of distributing single microelectrodes in a fixed plane situated over a circular patch of frontal cortex using conventionally-milled recording grids. With the advent of robust, multisite linear probes, and the introduction of commercially-available, high-resolution rapid prototyping systems, we have been able to improve upon traditional approaches. Here, we report our methodology for producing flexible, MR-informed recording platforms that allow us to precisely target brain structures of interest, including those that would be unreachable using previous methods. We have increased our single-session recording yields by an order of magnitude and recorded neural activity from widely-distributed regions using only a single recording chamber. This approach both speeds data collection, reduces the damage done to neural tissue over the course of a single experiment, and reduces the number of surgical procedures experienced by the animal.

Keywords: Brain-computer/machine Interface, Neural Interfaces - Computational modeling and simulation, Motor neuroprostheses
Abstract: The dynamics of brain activity are inherently non-stationary and can change depending on context, for example based on the tasks performed, the stimuli received, or the attention focus maintained. Thus to decode brain states in naturalistic scenarios, it is necessary to track such changes. Further, it is becoming increasingly common to measure the brain at multiple spatiotemporal scales by recording both spike and field activities simultaneously. Thus tracking non-stationarity necessitates efficient decoders that can detect changes in spike-field neural dynamics and accurately estimate the underlying neural and behavioral states. Here, we develop a new decoding framework to address this challenge. We build a multiscale switching dynamical model that assumes the underlying hidden neural state can evolve with dynamics chosen from an arbitrary finite set. The choice of dynamics could for example be dictated by a higher-level cognitive state such as attention level, which we call a switch state. This switch state would also dictate how the neural state is represented in the recorded binary spike events and continuous field signals. We derive a new multiscale decoder that simultaneously estimates the underlying neural and switch states from these spike-field observations. We show with closed-loop simulations that our new decoder is able to accurately estimate the switch state while accurately decoding the neural and behavioral states.

Keywords: Neural signal processing, Sensory Neuroprostheses - Auditory, Clinical neurophysiology
Abstract: Cortical auditory evoked potentials (CAEPs) are a popular neurophysiological measure in the assessment of auditory processing capabilities. Research evidence suggests that CAEPs may be employed to assess auditory discrimination abilities, which is of particular interest in monitoring clinical rehabilitation outcomes in cochlear implant (CI) users. However, the electrical artefact from CI stimulation poses a challenge for the signal analysis. Numerous artefact reduction procedures have been proposed, many of which are computationally costly or subjectively biased. This study investigated a simplified approach to CI artefact reduction based on subtraction, which was enabled by introducing an enhanced auditory oddball paradigm. Outcome CAEPs were compared to the gold-standard artefact reduction based on independent component analysis (ICA). Grand average difference potentials showed successful artefact reduction for both artefact rejection algorithms in the enhanced oddball paradigm. With the enhanced oddball paradigm, measured peak-to-peak amplitudes were significantly larger with subtraction-based processing than for ICA-based processing, suggesting that the removed independent components not only contained artefact, but also neural activity of interest.

Keywords: Brain-computer/machine Interface, Neural signal processing, Brain Functional Imaging - EEG and Evoked Potentials
Abstract: We introduce adversarial neural networks for representation learning as a novel approach to transfer learning in brain-computer interfaces (BCIs). The proposed approach aims to learn subject-invariant representations by simultaneously training a conditional variational autoencoder (cVAE) and an adversarial network. We use shallow convolutional architectures to realize the cVAE, and the learned encoder is transferred to extract subject-invariant features from unseen BCI users’ data for decoding. We demonstrate a proof-of-concept of our approach based on analyses of electroencephalographic (EEG) data recorded during a motor imagery BCI experiment.

Keywords: Motor Neuroprostheses - Neuromuscular stimulation
Abstract: Functional electrical stimulation is a promising technology for restoring functional reaching motions to individuals with upper limb paralysis. We present a control architecture that combines static models of a paralyzed arm and its response to stimulation with a PID controller. The controller is used to drive the wrist of an individual with tetraplegia to a desired wrist position. We compare the performance of our controller with a feedforward component and with no feedforward component. The combined feedforward-feedback controller produced an average accuracy (defined as the distance away from the target wrist position) of 4.9 cm, and the feedback controller produced an accuracy of 4.3 cm. The combined feedforward-feedback controller produced initially larger errors than the feedback controller, but the end performance was similar. The control architecture presented has the potential to be used for arbitrary reaching motions.

Keywords: Brain-computer/machine Interface, Brain-Computer/Machine Interface - Robotics applications, Brain-Computer/Machine Interface - Biofeedback
Abstract: Brain-machine interfaces (BMIs) may generate more errors than those encountered during normal motor control. Thus, they provide an opportunity to investigate neural correlates of error processing. Characterizing neural correlates of error processing may, in turn, provide a tool for on-line correction of the errors that are made by the interface. We investigated neural correlates of error processing during BMI experiments in which monkeys controlled an animated hand on the screen to touch a ball by moving their own fingers. Short movement segments that were consistently toward or away from the target were labeled accordingly and used to train a classifier to differentiate between correct and erroneous movements based on the neural activity. The results indicate that despite the limited number of labeled segments and active neurons in the studied data, the classifier achieved a classification rate of 68% on testing. The full receiver operating curve (ROC) has been estimated and indicates that even when the false alarm is restricted to 5%, the classifier can detect 36% of the erroneous movements. Better results are expected when using more data, especially as more challenging grasping tasks are performed. Such a classifier could be used to improve the performance of BMIs by detecting and correcting erroneous movements.

Keywords: Neural Interfaces - Neural stimulation, Neural Interfaces - Recording, Sensory Neuroprostheses - Visual
Abstract: To study electroencephalogram (EEG) power alterations in the brain of retinal degenerate (RD) models, wild type (WT) and RD mice were used in the experiments. The spontaneous neural activity in visual (Vcx) and pre-frontal (Pfx) cortices was recorded from the cranium before and after prolonged transcorneal electrical stimulation (pTES). Three weeks after initial post-stimulation recording (stage one), spontaneous neural activity was recorded (post-stimulation stage two recording). Spontaneous EEG band power was determined during the pre and post stimulation stages in awake and anesthetized states of both control and experimental groups. The results showed that pTES possibly activates Vcx and Pfx thus caused an increased trend in spontaneous EEG band power of awake RD mice. In this regard, delta band was the most dominant band in Vcx and Pfx and significantly (p<0.05) had the highest spontaneous EEG power compared to theta, alpha, beta and gamma bands respectively. Three weeks after initial post pTES recording, Vcx and Pfx maintained increased spontaneous activity in only delta band of awake RD mice. With the crucial roles slow wave delta oscillations have been reported to play in neuroplasticity, our results could indicate a neuroplastic effect of pTES.

Keywords: Brain-Computer/Machine Interface - Biofeedback, Brain-computer/machine Interface, Motor learning, neural control, and neuromuscular systems
Abstract: The ability to learn various kinds of skills, whether it is motor, sensory or cognitive, is pivotal to animal’s survival through evolution. Multiple coding schemes may be involved to encode information needed in learning. Among them, temporal coding is able to provide expanded bandwidth which renders more information to be encoded. Thus, we were motivated to inspect whether animal have the capability of encoding information in temporal precision to acquire skills. In this study, we adopted the brain-machine interfaces (BMIs) to investigate whether the precise spiking patterns can be directly reinforced and volitionally modulated. Rats were trained to learn to control the pitch of an auditory cursor to achieve rewards by volitionally modulating the number of concurrent spiking pairs between two selected units. We found that, performance in the task improved over sessions with successful modulation of relative timing of firings between two selected units to a short interval within 15ms, which indicated that temporal coding can play a part in abstract skill learning. Moreover, the fine temporal structure of neural activity and the yielding behavioral outcomes involved in learning was able to be unified under one experimental framework and robustly controlled for further studies. Successful conditioning of temporal spike patterns also revealed the great potential for high-resolution BMIs and clinical applications.

Keywords: Sensory Neuroprostheses - Visual, Neural Interfaces - Neural stimulation
Abstract: For individuals blinded by retinitis pigmentosa or age-related macular degeneration, retinal prosthetic devices can restore some form of useful vision. Unfortunately however, their performance varies considerably across subjects and the best performance has not yet reached the level of legal blindness. It has been thought that the limitation of retinal prostheses comes from the lack of comprehensive understanding about how various types of retinal ganglion cells (RGCs) respond to an identical electric stimulus as well as about how changes in stimulus parameters alter their response patterns. In particular, it is crucial to better understand spiking patterns of ON and OFF types of RGCs in response to electric stimulation because these two types are known to be critical in forming visual percepts. Our previous studies reported that the two types have different sensitivities to changes in various stimulus parameters such as stimulation rate and stimulus duration, suggesting a possibility of similar difference in response to other changes. In this paper, we explored response changes of ON and OFF types of alpha RGCs to varying stimulus charges. We found that, at a fixed current amplitude, the response magnitudes of OFF types of alpha RGCs are minimally altered by doubled stimulus charge while those of ON RGCs are significantly altered by the same change. This finding may be used for optimizing overall retinal responses to electric stimulation for better artificial vision.

Keywords: Brain-computer/machine Interface, Neuromuscular Systems - Computational modeling and simulation, Neural signal processing
Abstract: Motor brain-machine interfaces (BMIs) usually collect neuronal activity from the dorsal premotor cortex (PMd) and primary motor cortex (M1). As the pattern of interconnections between PMd and M1 is complex, current BMI decoders directly decode PMd and M1 neural signals without considering the interconnections between these cortical areas. In this paper, a new architecture of recurrent neural network (Double Recurrent Neural Network, DRNN) was proposed, which took the interconnection information between PMd and M1 into account. To evaluate the performance of DRNN, we recorded the spike data and the position of a robotic arm when a rhesus monkey performed one-dimensional robotic arm reach task. When DRNN decoder offline decoded the velocity of the robotic arm, it showed high decoding accuracy of 0.92 ± 0.04 (correlation coefficient) and strong robustness to noise and recording conditions change. The DRNN outperformed basic RNN and state-of-the-art velocity Kalman filter (VKF) in both decoding accuracy and robustness. It suggested that our proposed DRNN could be a promising BMI decoding algorithm.

Keywords: Brain Functional Imaging - Connectivity and Network, Neural signal processing, Human Performance - Cognition
Abstract: Previous studies on EEG-based emotion recognition mainly focus on single-channel analysis, which neglect the functional connectivity between different EEG channels. This paper aims to explore the emotion associated functional brain connectivity patterns among different subjects. We proposed a critical subnetwork selection approach and extracted three topological features (strength, clustering coefficient, and eigenvector centrality) based on the constructed brain connectivity networks. The experimental results of 5-fold cross validation on a public emotion EEG dataset called SEED indicate that the common connectivity patterns associated with different emotions do exist, where the coherence connectivity is significantly higher at frontal site in the alpha, beta and gamma bands for the happy emotion, at parietal and occipital sites in the delta band for the sad emotion, and at frontal site in the delta band for the neutral emotion. In addition, the results demonstrate that the topological features considerably outperform the conventional power spectral density feature, and the decision-level fusion strategy achieves the best classification accuracy of 87.04% and the corresponding improvement of 3.78% in comparison with the state-of-the-art using the differential entropy feature on the same dataset.

Keywords: Neuromuscular Systems - Wearable systems, Neurological disorders - Stroke
Abstract: Automated systems for assessing upper extremity motor function post-stroke have been proposed as a higher resolution, more objective, or faster alternative to routine clinical assessment. These studies have been performed using either remote or wearable sensors attached to the subject. A common difficulty of these systems has been the quantification of hand function, a major component of post-stroke motor dysfunction. Mechanomyographic sensors have been untested in this field but have much potential as a practical way to quantify hand function, and as a complimentary modality to kinematic data for quantifying other arm movements. For this study a new automated system has been proposed which incorporates both kinematic and myographic information in the classification of arm motor function. Twenty-eight subjects with acute stroke were recruited and instructed to perform the motor function tasks of the FMA-UE. Motor features were pruned using the ReliefF feature selection algorithm and classification performed using a linear Support Vector Machine. Non hand function tasks classified using myographic data achieved a mean classification accuracy of 50.5% compared to 62.0% achieved using IMU data alone. For hand function tasks, a higher classification accuracy of 62.4% was achieved using myographic data alone.

Keywords: Neural signal processing, Brain-computer/machine Interface, Neuromuscular Systems - Computational modeling and simulation
Abstract: Reinforcement learning(RL) scheme interprets the movement intentions in Brain-machine interfaces(BMIs) with a reward signal. This reward can be an external reward (food or water) and an internal representation of the reward which mimics the learning procedure of the subjects to link the correct movement with the external reward. Medial prefrontal cortex (mPFC) has been demonstrated to be closely related to the reward-guided learning. In this paper, we propose to model mPFC activities associated with different actions as an internal representation of the reward in the RL framework. Support vector machine (SVM) is adopted to distinguish the rewarded and unrewarded trials based on mPFC signals considering different action spaces. Then this discrimination result will be utilized to train a decoder. Here we introduce the attention-gated reinforcement learning (AGREL) as the decoder to generate a mapping between motor cortex(M1) and action states. To evaluate our approach, we test on the real neural physiological data collected from rats when performing a two-lever discrimination task. Compared with AGREL using the external reward, AGREL using internal reward evaluation can achieve a prediction accuracy of 94.8%, which is very close to using the explicit reward. This indicates the feasibility of modeling mPFC activities as an internal representation of the reward in the RL framework.

Keywords: Neural signal processing
Abstract: An innovative filter design method is proposed for threshold-based spike sorting of high-density neural recordings. Threshold-based spike sorting is the process of assigning each detected spike in an extracellular recording to its putative neuron, using only linear filters and simple thresholding operations. The low computational complexity of threshold-based spike sorting makes it interesting for real-time (hardware) implementations with potential applications in the field of brain-machine interfaces. The proposed method extends our earlier work on discriminative template matching and avoids the need for a prior heuristic definition of an interference covariance matrix. A new optimal filter design objective function is proposed, which automatically selects interference-dominated signal segments based on the output signal of the filter under design. This new method leads to filters that are signal-to-peak-interference ratio (SPIR) optimal. The method is validated on ground truth data recorded in-vivo.

Keywords: Neural signal processing, Neural Signal Processing - Time frequency analysis
Abstract: Event-related potentials (ERP) provide reliable electrophysiological correlates of subsequent neural processing following sensory stimulation, offering insight into the activation patterns of participating neural structures which is of considerable value in both neuroscience research and clinical applications. 2D single-trial representations as ERP images have seen increased application in recent studies, accompanied by a rising number of approaches to improve their signal-to-noise ratio, which for the most part have been motivated from an image processing point of view (e.g., nonlocal operators, anisotropic diffusion filtering). In this paper, a brief overview of ERP image denoising prior art is given and a novel, fast denoising algorithm based on split amplitude and phase processing (i.e., phase-informed amplitude shrinkage and regularization of the phase structure) in analytic time-frequency representations of ERP single trials obtained using a perfect reconstruction wavelet filterbank is proposed. Furthermore, the performance of the proposed algorithm is subjected to a comparative evaluation using real-world chirp-evoked auditory ERP acquired from 20 normal hearing adults. Results suggest the suitability of the proposed method for a broad range of a posteriori ERP image denoising tasks, including those lacking a priori knowledge about the shape of potentially nonstationary traces in the ERP image due to, e.g., endogeneous states gradually changing during the experiment.

Keywords: Brain-computer/machine Interface, Human Performance - Cognition, Brain-Computer/Machine Interface - Robotics applications
Abstract: Cognitive information has been exploited in non-invasive Brain Computer Interface (BCI) scenarios to provide autonomous external agents with additional information. In this context, Error-related potentials (ErrPs), temporal deflections in electroencephalogram (EEG) signals when humans perceive erroneous actions, have been exploited to teach correct policies to agents. However, previous works have shared same objective criteria to evaluate actions of agents across humans. Therefore, it is yet an open question whether ErrPs are elicited when humans assess actions based on individual subjective criteria, and if such neuronal activities can be exploited in BCIs to enhance personalized human computer interactions. In this work, we evaluate whether ErrPs are elicited while humans assess actions based on individual subjective criteria. For this purpose, we analyze EEG signals while humans evaluate trajectories performed by the robot to avoid an obstacle. We show that the ErrP can be generated even while the human evaluates the action based on his/her subjective criteria, and that such neuronal activity reveals subjective aspects when assessing the appropriateness of the action. These findings open the door to the exploitation of neural correlates for the personalized human computer interactions, and thus to building adaptable, individualized neuroprosthetic devices.

Keywords: Neural Signal Processing - Nonlinear analysis, Human Performance - Sensory-motor, Neurological disorders
Abstract: Carriers of the fragile X premutation are at risk of developing the late-onset neurodegenerative movement disorder Fragile X associated tremor/ataxia Syndrome (FXTAS). Although only recently discovered, emerging research has revealed that premutation carriers exhibit various subtle deficits in certain aspects of cognition and stability, well before the typical age of onset of FXTAS. This study aimed to characterize the premutation phenotype in terms of cognition and postural stability as well as investigating the relationship between these two faculties with the aim of identifying pre-clinical symptoms of FXTAS. A group of 12 female premutation carriers and 15 healthy controls had postural stability recorded under while standing with their eyes open, closed, as well as during working memory and attention based dual-tasks. Postural sway was characterized through traditional parameters (area, path length, and velocity) as well as through entropy-based measures such as the Complexity Index. There were no differences observed in terms of traditional parameters between groups. The control group exhibited higher complexity indices than carrier in both dual-task conditions. Similarly, the complexity index of the control group was significantly higher during both dual-tasks, compared to the baseline eyes open condition, while premutation carriers remained consistent across task. This may reflect a reduction in adaptive capacity in premutation carriers.

Keywords: Brain-computer/machine Interface, Human Performance - Sensory-motor, Brain Functional Imaging - EEG and Evoked Potentials
Abstract: Current non-invasive Brain Machine interfaces commonly rely on the decoding of sustained motor imagery activity. This approach enables a user to control brain-actuated devices by triggering predetermined motor actions. However, despite of its broad range of applications, this paradigm has failed so far to allow a natural and reliable control. As an alternative approach, we investigated the decoding of states transitions of an imagined movement, i.e. rest-to-movement (onset) and movement-to-rest (offset). We show that both transitions can be reliably decoded with accuracies of 71.47% for the onset and 73.31% for the offset (N = 9 subjects). Importantly, these transitions exhibit different neural patterns and need to be decoded independently. Our results indicate that both decoders are able to capture the brain dynamics during imagined movements and that their combined use could provide benefits in terms of accuracy and time precision.

Keywords: Neural Signal Processing - Time frequency analysis, Neural signal processing, Brain Physiology and Modeling - Neural circuits
Abstract: Multi-layered brain structures contain canonical microcircuits that specialize in region-specific functions. Information flow across the layers is typically inferred using multivariate techniques that operate on local field potentials (LFPs). These methods (e.g., Granger Causality (GC)) are sensitive to the presence of a common reference that corrupts LFPs recorded with a multi-contact electrode and introduces spurious covariations. Using models of reference-noise corrupted signals with laminar interactions, we evaluated the efficacy of three dereferencing methods - bipolar subtraction, current source density (CSD), and common average referencing. We examined which method best recovered the underlying functional interactions between layers. Each dereferencing method introduced different types of error, often to alarming levels of false prediction. While CSD and bipolar subtraction methods performed best, they often predicted spurious connections, exhibited GC power in incorrect frequency bands, and/or missed salient relationships between layers. Though the confounds in this model may not be present in evaluations of functional connectivity between brain areas, these findings call for a reassessment of dereferencing methods used in the context of evaluating information flow within laminar neural tissue.

Keywords: Neural signal processing, Brain Functional Imaging - EEG and Evoked Potentials, Brain Functional Imaging - Connectivity and Network
Abstract: Recent advances in the field of human brain imaging by electrophysiological recordings allow to reliably quantify the connectivity patterns of spontaneous oscillatory activity. This has provided novel tools to investigate the neural basis underlying complex human behavior and to unravel the mechanisms of brain function and reorganization in response to neurological diseases. In this context, we present a pipeline integrating high-density electroencephalography (hd-EEG) analysis and graph theory. To test our pipeline, we conducted a feasibility study on hd-EEG analysis to recordings from healthy individuals, examining the frequency-specific small-world organization of brain connectivity. Here we present the preliminary results on a small sample size. The ultimate goal of our study is to extract graph-theory related metrics, such as small-worldness index, from injured patients and use them as electrophysiological biomarkers of the recovery.

Keywords: Neural Signal Processing - Blind source separation, Brain Functional Imaging - EEG and Evoked Potentials, Neural Signal Processing - Time frequency analysis
Abstract: Recent advancement in electroencephalography (EEG) signal processing and hardware can greatly reduce motion artifact, but neck muscle electrical activity is a problem during mobile brain imaging studies examining whole body movement tasks like walking and running. To test the ability of independent component analysis (ICA) to extract neural signals contaminated by neck muscle electrical activity, we broadcast ground-truth electrical signals through a head phantom device during motion. We placed the phantom on a motion platform used to replicate human head trajectories during walking and embedded neck muscle sources within the phantom. ICA was able to extract artificial neural sources from even the most contaminated data in this simulation of human walking. Performance of ICA in high muscle activity amplitude conditions was improved by including electromyographic recordings in the ICA decomposition. These results highlight the importance of recording multiple electromyographic signals from the neck during mobile brain imaging with EEG for studying electrocortical dynamics during movement.

Keywords: Motor Neuroprostheses - Neuromuscular stimulation, Neuromuscular Systems - Peripheral mechanisms, Neural Interfaces - Neural stimulation
Abstract: During the last decades, spinal cord stimulation (SCS) has attracted much attention due to its capability to modulate the motor and sensory networks. The potential of this technique has been proved, and several investigations have focused on applying it for restoring lower limb function. The majority of SCS approaches are based on electrical stimulation, and few studies have explored magnetic fields for non-invasive SCS. This paper presents a trans-spinal magnetic stimulation (ts-MS) protocol and studies its effects on spinal circuits with seven healthy subjects, considering central and peripheral nervous systems. Motor evoked potentials (MEP) and transspinal motor evoked potentials (ts-MEP) were assessed before and after the ts-MS intervention to characterize excitatory responses. After the intervention, we found an increase of almost 30% (not statistically significant) in MEP amplitude, but no changes in ts-MEP amplitude. Further research is required to confirm, in a larger population of subjects, the potential of this technology, which could be used to improve rehabilitation therapies for patients with motor disabilities.

Keywords: Brain Functional Imaging - EEG and Evoked Potentials, Brain-computer/machine Interface
Abstract: Steady-state visual evoked potentials (SSVEP) are responses elicited when a user is presented with a repetitive visual stimulus. Change in stimuli proximity has been shown to have an influence on the performance of SSVEP-based BCI, where the inter-stimulus distance has a positive correlation with the overall performance. This limits the flexibility in stimulus design by imposing a constraint on the acceptable inter-stimulus distance, consequently limiting the range of applicability for SSVEP-based BCIs in real-world applications. Another limitation that needs to be addressed is the required number of EEG channels. In this study, we investigated these two challenges. A process of selecting optimal user-specific channel set was proposed. We demonstrated that the user-specific channel set is more robust against variable inter-stimulus distance. A significant improvement in accuracy (p=10-3) of 5% and a reduction in variation (p=10-3) of 55% was achieved on average when compared to the performance using the classic 3-channel set (O1, O2, Oz) and 6-channel set (O1, O2, Oz, PO3, PO4, POz).

Keywords: Motor learning, neural control, and neuromuscular systems, Neuromuscular Systems - Computational modeling and simulation, Neuromuscular Systems - EMG models, processing and applications
Abstract: During reach-to-grasp motions, the Electromyographic (EMG) activity of the arm varies depending on motion stage. The variability of the EMG signals results in low classification accuracy during the reaching phase, delaying the activation of the prosthesis. To increase the efficiency of the pattern-recognition system, we investigate the muscle activity of four individuals with below-elbow amputation performing reach-to-grasp motions and segment the arm-motion into three phases with respect to the extension of the arm. Furthermore, we model the dynamic muscle contractions of each class with Gaussian distributions over the different phases and the overall motion. We quantify of the overlap among the classes with the Hellinger distance and notice larger values and, thus, smaller overlaps among the classes with the segmentation to motion phases. A Linear Discriminant Analysis classifier with phase segmentation affects positively the classification accuracy by 6-10% on average.

Keywords: Brain-computer/machine Interface, Neural Interfaces - Computational modeling and simulation, Brain physiology and modeling - Neural dynamics and computation
Abstract: Neural dynamics can be non-stationary in many neural applications such as brain-machine-interfaces (BMIs) and long-term brain stimulation. Adaptive modeling is a useful approach in tracking the neural non-stationarities. Our prior work has tracked time-variant linear state-space models (LSSM) to describe human electrocorticogram (ECoG) dynamics. One key design parameter in adaptive LSSM identification is the forgetting factor, which could significantly influence the accuracy of the fitted model. However, this influence has not been systematically studied yet. Here, we use comprehensive numerical simulations to investigate the effect of the forgetting factor on identifying time-variant LSSMs. We simulate non-stationary neural activity using time-variant LSSMs and use different forgetting factors to track time-variant LSSMs and predict the simulated activity. We find that the prediction accuracy of the fitted models strongly varied with the choice of the forgetting factor. We also find that the optimal forgetting factor that led to the highest prediction accuracy varied as a function of various properties of the simulated time-variant LSSMs. Our results have implications for building more accurate adaptive models to track non-stationary neural network dynamics and can facilitate the investigation of neural non-stationarity using the adaptive models.

Keywords: Brain-Computer/Machine Interface - Robotics applications, Neurorehabilitation, Brain Functional Imaging - MEG
Abstract: This study investigates the effect of channel selection on the performance of a Magnetoencephalography (MEG)-based brain-computer interface (BCI) system in terms of classification accuracy (CA). Although many efforts are currently being undertaken to develop BCI using MEG, the major concern still is low accuracy. MEG systems involve data recording from a large number of channels which may provide a better spatio-temporal resolution for assessing brain patterns, however, a large numbers of channels result in a large number of features, which further make feature learning a challenging task. In this study, we evaluated the performance of two state-of-the-art channel selection methods, i.e. class-correlation (CC) and ReliefF (RF) across six binary classification tasks with a MEG dataset of 15 healthy participants. Both CC and RF methods provided a statistically significant increase in the CA (range: 20.91 - 24.22%) compared to baseline (i.e. using 204 channels) with bandpower features from the alpha (8-12 Hz) and beta frequency bands (13-30 Hz). Moreover, both methods reduce the optimum number of channels significantly (from 204 to the range of 1-22). Reducing the number of features can significantly reduce the computational cost and increase the chances of numerical stability which are key considerations in neurofeedback (online) applications.

Keywords: Brain Functional Imaging - Connectivity and Network, Brain Functional Imaging - Source localization, Human Performance - Drowsiness and microsleeps
Abstract: Understanding how different brain areas interact to generate complex behavior is a primary goal of neuroscience research. One approach, functional connectivity analysis, aims to characterize the connectivity patterns within brain networks. In this paper, we address the problem of discriminative connectivity, i.e. determining the differences in network structure under different experimental conditions. We introduce a novel model called Sparse Multi-task Inverse Covariance Estimation (SMICE) which is capable of estimating a common connectivity network as well as discriminative networks across different tasks. We apply the method to EEG signals after solving the inverse problem of source localization, yielding networks defined on the cortical surface. We propose an efficient algorithm based on the Alternating Direction Method of Multipliers (ADMM) to solve SMICE. We apply our newly developed framework to find common and discriminative connectivity patterns for alpha-oscillations during the Sleep Onset Process (SOP) and during Rapid Eye Movement (REM) sleep. Even though both stages exhibit a similar alpha-oscillations, we show that the underlying networks are distinct.

Keywords: Brain-computer/machine Interface, Brain Functional Imaging - EEG and Evoked Potentials, Neural Signal Processing - Blind source separation
Abstract: Imagination of limb movements causes increase or decrease in the power of brain activity in different regions of motor cortex. Electroencephalogram (EEG) collected during imagination of limb movement is used as an input for Brain computer interface (BCI) systems to translate the imagination into a command. We have studied the effect of continues hand movement imagination on the event related potentials in the brain. The goal of this paper is to investigate whether continues imagination of hand movement can cause continues effect in the event related potential. For this purpose, a set of experiments were designed and performed to collect 180 trials of hand movement imagination for 3 subjects. We utilized independent component analysis (ICA) to separate the sources related to continues hand movement imagination. The statistical methods including Ansaribradley and analysis of variance (ANOVA) are used to select the components/sources with maximum correlation to the hand movement imagination task. It is shown that although power in both alpha and beta bands decrease by movement imagination, alpha band can differentiate continues imagination vs. non-imagination better than beta band. Besides that, we have found an improvement in the event related synchronization and desynchronization over different sessions for all subjects, which shows effect of training on improving the imagination of the task.

Keywords: Neural signal processing, Brain-computer/machine Interface, Neurorehabilitation - Neurofeedback
Abstract: ElectroEncephaloGram (EEG) spectral power has been extensively used to classify Mental Imagery (MI) of movements involving different body parts. However, there is an increasing need to enable classification of MI of movements within the same limb. In this work, EEG spectral power was recorded in seven subjects while they performed MI of closing (grip) and opening (extension of fingers) the hand. The EEG data was analyzed and the feasibility of classifying MI of the two movements were investigated using two different classification algorithms, a linear regression and a Convolutional Neural Network (CNN). Results show that only the CNN is able to significantly classify MI of opening and closing of the hand with an average classification accuracy of 60.4%. This indicates the presence of higher-order non-linear discriminatory information and demonstrates the potential of using CNN in classifying MI of same-limb movements.

Keywords: Human Performance - Fatigue, Human Performance - Attention, Human Performance - Cognition
Abstract: Monitoring mental workload in a fast and accurate manner is important in scenarios where the full attention of humans involved is fundamental to the security of others. Firefighters, air traffic controllers, and first responders are constantly submitted to such tasks. In many cases, in addition to a demanding mental task, humans are also under varying levels of physical strain. Measuring mental workload under such scenarios is challenging, especially when relying on wearable sensors. In this paper, we explore the combination of an automated artifact removal algorithm with spectro-temporal features for mental workload assessment ``in-the-wild,'' where varying levels of physical strain are present. Experiments show these features outperforming classical spectral ones for mental workload classification under two activity types (biking and walking/running) and three activity levels (none, low, high). Improved performance was achieved when both feature types were combined, thus suggesting complementarity for the task at hand.

Keywords: Neural Signal Processing - Blind source separation
Abstract: In recent years, muscle synergies have been proposed for proportional myoelectric control. Synergies were extracted using matrix factorisation techniques (mainly non-negative matrix factorisation, NMF), which requires identification of synergies to tasks or movements. In addition, NMF methods were viable only with a task dimension of 2 degrees of freedoms (DoFs). Here, the potential use of a higher-order tensor model for myoelectric control is explored. We assess the ability of a constrained Tucker tensor decomposition to estimate consistent synergies when the task dimensionality is increased up to 3-DoFs. Synergies extracted from third-order tensor of 1 and 3 DoFs were compared. Results showed that muscle synergies extracted via constrained Tucker decomposition were consistent with the increase of task-dimension. Hence, these results support the consideration of proportional 3-DoF myoelectric control based on tensor decompositions.

Keywords: Human Performance - Gait, Brain Functional Imaging - EEG and Evoked Potentials, Human Performance - Cognition
Abstract: In the last years Mobile Brain/Body Imaging (MoBI) has been increasingly used to study cognition in the real world to give more ecological validity to brain imaging studies currently carried out only inside the lab. To increase portability of the setup and reduce cabling it is possible to perform a unified and real-time synchronized recording of data from multiple sources. However, delays and jitter may impair the quality of the subsequent event-related potential (ERP) analyses. Here we used an online auditory oddball P300 paradigm on one subject to compare the quality of P300 ERPs obtained (i) with online synchronization and alignment and (ii) offline with conventional alignment (synchronization channel) while sitting. We showed that offline and online synchronization strategies provided comparable although slightly different P300 ERP. We also recorded the electroencephalogram (EEG) while walking indoors and outdoors. A decreasing P300 amplitude respectively from sitting to walking indoors and outdoors confirms the dual-task effect on P300. These results show that integrated real-time P300 protocols are feasible but it is also necessary to test delays and quantify the jitter among different signals when developing real-world MoBI applications.

Keywords: Neural Interfaces - Computational modeling and simulation, Neural Signal Processing - Time frequency analysis, Neurorehabilitation - Wearable systems
Abstract: Computational cost is an important consideration for memory encoding prediction models that use data from dozens of implanted electrodes. We propose a method to reduce computational expense by selecting a subset of all the electrodes to build the prediction model. The electrodes were selected based on their likelihood of measuring brain activity useful for predicting memory encoding better than chance (in terms of AUC). A logistic regression prediction model was built using spectral features of intracranial electroencephalography (iEEG) from the selected electrodes. We demonstrate our method on iEEG data from 37 human subjects performing free recall verbal short-term memory tasks. The method achieves a 36.3% reduction in the number of electrodes used for prediction, resulting in a 64.9% reduction in inference computation time with just a 0.3% loss in prediction performance compared to the case when all electrodes were used. The electrodes selected using our method provided improved prediction performance compared to those electrodes that were not selected in 31 out of 37 patients. Building upon this observation, we also developed a method to identify the subjects for whom the proposed electrode selection method would be beneficial.

Keywords: Brain-computer/machine Interface, Neural Interfaces - Computational modeling and simulation, Neural signal processing
Abstract: We study brain-computer interfaces (BCI) based on the decoding of motor imagery (MI) from electroencephalog- raphy (EEG) neuromonitoring. The robustness of MI-BCI is a major concern in practical applications, and hence various efforts in the literature have been made to enhance the MI classification accuracy from EEG signals. Recently, classifiers based on convolutional neural networks (CNN) have achieved state-of-the-art performance. In further exploration of applying CNNs to EEG data, we propose a spatial component-wise con- volutional network (SCCNet), featuring an initial convolutional layer for spatial filtering, a common processing in EEG analysis for signal enhancement and noise reduction. Through a series of optimization and validation, we show the superiority of SCCNet in MI EEG classification, outperforming other existing CNNs.

Keywords: Neural Interfaces - Recording, Neural Interfaces - Implantable systems, Neural Interfaces - Neuroimaging
Abstract: Neural electrical impedance tomography (EIT) images changes in tissue impedance, associated with neural activity, by passing a carrier signal through neural tissue and measuring changes in the boundary voltages. In this paper, the frequency limits of an EIT carrier signal are investigated across the range 1 to 32 kHz, using a nerve cuff on rat sciatic nerve in-vitro. The upper frequency limit was determined, from impedance measurements of inactive nerve tissue, to be 20 kHz, and the lower limit, from frequency analysis of an artificially evoked compound action potentials, to be 2 kHz. This operating frequency range, of 2 to 20 kHz, is broad enough to multiplex several carrier signals, which is expected to be an essential feature of future real-time EIT systems for peripheral nerve.

Keywords: Brain Functional Imaging - Source localization, Brain Functional Imaging - MEG, Brain Functional Imaging - EEG and Evoked Potentials
Abstract: One of the enduring challenges in MEG/EEG data analysis is the poor performance of source reconstruction algorithms under high noise and interference conditions, especially in case of distributed, correlated brain activity with complex spatial extent. In our previous work, we developed a source localization algorithm, Champagne, which is robust to the effects of noise, interference and highly correlated brain source activity. Champagne is ideally suited for reconstructions of sparse and highly clustered brain source activity rather than reconstruction of distributed source activity with larger spatial extents. Here, we introduce a novel Bayesian algorithm that enables reconstruction of distributed source activity. We build upon the robust performance features of the Champagne algorithm and refer to this algorithm as Smooth Champagne. Simulations demonstrate excellent performance of Smooth Champagne in determining the spatial extent of source activity. Smooth Champagne also accurately reconstructs real MEG and EEG data.

Keywords: Brain Functional Imaging - MEG, Brain Functional Imaging - Classification, spatiotemporal dynamics, Neural signal processing
Abstract: Accurate interpretation of speech activity from brain signals is critical for understanding the relationship between neural patterns and speech production. Current research on speech activity recognition from the brain activity heavily relies on the region of interest (ROI) based functional connectivity analysis or source separation strategies to map the activity as a spatial localization of a brain function. Albeit effective, these methods require prior knowledge of the brain and expensive computational effort. In this study, we investigated automatic speech activity recognition from brain signals using machine learning. Neural signals of four subjects during four stages of a speech task (i.e., rest, perception, preparation, and production) were recorded using magnetoencephalography (MEG), which has an excellent temporal and spatial resolution. First, a deep neural network (DNN) was used to classify the four whole tasks from the MEG signals. Further, we trained a sequence to sequence (Seq2Seq) long short-term memory-recurrent neural network (LSTM-RNN) for continuous (sample by sample) prediction of the speech stages/tasks by leveraging its sequential pattern learning paradigm. Experimental results indicate the effectiveness of both DNN and LSTM-RNN for automatic speech activity recognition from MEG signals.

Keywords: Brain Functional Imaging - EEG and Evoked Potentials, Brain Functional Imaging - Multimodal, Neural Signal Processing - Blind source separation
Abstract: Mobile Brain/Body Imaging (MoBI) is an emerging non-invasive approach to investigate human brain activity and motor behavior associated with cognitive processes in natural conditions. MoBI studies and analyses pipelines combine brain measurements, e.g. Electroencephalography (EEG), with motion data as participants conduct tasks with near-natural behavior. Within the field however, standard source decomposition and reconstruction pipelines largely rely on unsupervised blind source separation (BSS) approaches and do not consider movement information to guide the decomposition of oscillatory brain sources. We propose the use of a supervised spatial filtering method, Source Power co-modulation (SPoC), for extracting source components that co-modulate with body motion. Further, we introduce a method to validate the quality of oscillatory sources in MoBI studies. We illustrate the approach to investigate the link between hand and head movement kinematics and power dynamics of EEG sources while participants explore an invisible maze in virtual reality. Stable oscillatory source envelopes correlating with hand and head motion were isolated in all subjects, with median  = .13 for all sources and median  = .16 for sources passing the selection criteria. The results indicate that it is possible to improve movement related source separation to further guide our understanding of how movement and brain dynamics interact.

Keywords: Sensory Neuroprostheses - Visual, Neural Interfaces - Computational modeling and simulation, Brain Physiology and Modeling - Neural circuits
Abstract: Visual prostheses aim to restore vision to people blinded from degenerative photoreceptor diseases by electrically stimulating surviving neurons in the retina. However, a major challenge with epiretinal prostheses is that they may accidentally activate passing axon fibers, causing severe perceptual distortions. To investigate the effect of axonal stimulation on the retinal response, we developed a computational model of a small population of morphologically and biophysically detailed retinal ganglion cells, and simulated their response to epiretinal electrical stimulation. We found that activation thresholds of ganglion cell somas and axons varied systematically with both stimulus pulse duration and electrode-retina distance. These findings have important implications for the improvement of stimulus encoding methods for epiretinal prostheses.

Keywords: Brain physiology and modeling, Brain Functional Imaging - EEG and Evoked Potentials, Brain Functional Imaging - Source localization
Abstract: Despite the widespread use of Electroecephalography (EEG) as an imaging modality, neural generators of current dipoles measured by EEG at the scalp are not fully understood. Here, we use two morphologically accurate multicompartments neuron models (layer IV pyramidal cell and layer V spiny stellate cell) to characterize how spiking neurons generate current dipoles in response to synaptic input. The simulations indicate that the dipole generated by synaptic inputs required to drive a pyramidal cell to threshold is smaller than the dipole associated the action potential itself. These results suggest a greater role of spiking neural activity toward EEG signals measured at the scalp than typically assumed.

Keywords: Neural Interfaces - Microelectrode and fabrication technologies, Neural Interfaces - Neural stimulation, Motor neuroprostheses
Abstract: Direct current (DC) can be applied to a nerve to generate a complete nerve block. However, using conventional platinum electrodes, reactions occur at the nerve interface causing damage to the nerve. The electrode can be separated from the nerve using a biocompatible, ionically conducting medium, which isolates the damaging reactions in a vessel away from the nerve. This electrode has previously been referred to as the Separated Interface Nerve Electrode (SINE). Recent experiments have shown that when a complete block is applied, for a prolonged period of time, there is a delay in the recovery of the response. For many applications it would be advantageous to have instantaneous recovery. To achieve this, the SINE electrode was used to provide a complete, instantaneous nerve block at the block threshold for 10 minutes and then the current was reversed for selected time periods to determine if instantaneous recovery could be achieved. Depending on the length of time of the repolarization, the amount of time for recovery can be reduced by as much as 50%.

Keywords: Brain-computer/machine Interface, Brain Functional Imaging - EEG and Evoked Potentials
Abstract: Synchronous or asynchronous steady-state visual evoked potential (SSVEP) based brain-computer interface (BCI) has been widely studied due to its advantages of obvious spectral characteristics, less electrodes usage and no training needed for ordinary users. In this paper, a method of combining asynchronous eye-tracking switch with synchronous SSVEP-based BCI is proposed. The eye tracking is designed as an ‘upper-level’ switch of the hybrid system to provide a free-control characteristic, and the ‘lower-level’ synchronous SSVEP BCI retains the high-accuracy property for multi-command control. Results showed that the accuracy of the eye-tracking-switch based hybrid system is relatively higher than the common SSVEP BCI, which proved the feasibility and user-friendliness of the designed method.

Keywords: Human Performance - Modelling and prediction, Sensory Neuroprostheses - Auditory
Abstract: Speech perception in background noise remains a challenge in cochlear implant (CI) users, and noise-suppression processing (e.g., Wiener filtering) has been commonly utilized to improve speech perception for CI users. It is crucial to objectively examine the perception of the noise-suppressed speech in CI users. The purpose of this work was to investigate whether the mismatch negativity (MMN) response could objectively assess the quality of the noise-suppressed speech as perceived by CI users. A vowel /a/ stimulus was masked by a steady-state noise, creating two noisy stimuli at two signal-to-noise ratios (SNRs) of -5 and +5 dB. The two noisy stimuli were processed by Wiener filtering. Electroencephalogram (EEG) data obtained from 7 CI users who participated in an auditory oddball paradigm was analyzed to extract the MMN. The two noise-suppressed stimuli served as the deviant stimuli and the clean vowel stimulus as the standard stimulus. Experimental results showed that the noise-suppressed stimuli at -5 dB SNR evoked a larger MMN amplitude than that at +5 dB SNR, accounting for the effect of SNR level on the auditory evoked cortical response to the noise-suppressed speech. The MMN may be potentially used as an objective biomarker to evaluate the perception of the noise-suppressed speech in CI users.

Keywords: Brain Stimulation-Deep brain stimulation, Neural Interfaces - Recording, Neural Interfaces - Neural stimulation
Abstract: Parkinson’s Disease can be modelled in rats to investigate the impact of dopamine loss. Unilateral lesions, using 6-OHDA, cause a degeneration of the dopaminergic neurons of the nigrostriatal pathway resulting in motor deficits. The experiments performed in this study aimed to compare the motor and non-motor behaviour of healthy rats vs. lesioned rats (PD). We also assessed the effect that deep brain stimulation (DBS) has on the animals’ locomotor activity. This paper covers the rotarod, open field, and cylinder tests, as well as sleep-awake analysis. The results of the rotarod test show that healthy rats are able to stay on the rod for longer than PD rats. The open field test revealed that PD rats spend more time freezing and close to walls or corners, whereas healthy and stimulated rats are more motivated to explore. In the cylinder test, during DBS, there is an increase in both right and left paw touches on the cylinder wall compared to pre- and post-DBS. It has also been shown that PD and control rats show different sleep patterns, further experiments could be helpful to better understand this difference.

Keywords: Brain Functional Imaging - Classification, spatiotemporal dynamics, Brain-computer/machine Interface, Brain physiology and modeling
Abstract: In this paper, we consider the problem of early decoding of EEG signals associated with tongue-hand motor execution and imagery tasks. A two-step feature extraction strategy is proposed. In the first step, the EEG data is segmented into sequences of quasi-stationary intervals, during which functional networks sustain their connectivity. The second step localizes the functional networks during each segment, to generate the dynamic functional connectivity graphs. Next, the common spatial pattern (CSP) algorithm is employed to select features for the classification problem. To take advantage of the dynamic nature of the generated graphs, a long short term memory (LSTM) classifier is used for classification. Using the first 500 ms of EEG recordings, the proposed framework is capable of classifying different tongue-hand motor execution and imagery tasks, with an average accuracy of 79%.

Keywords: Brain Functional Imaging - Source localization, Brain Functional Imaging - EEG and Evoked Potentials, Neural signal processing
Abstract: Mismarking locations of the fiducials can have a significant influence on the digitized electrode locations and cortical source estimation using high-density EEG. Understanding and quantifying how uncertainties in the fiducial locations affect the locations of cortical sources is important for interpreting EEG analyses. We systematically shifted fiducial locations to investigate the relationship between variations of fiducial locations and the corresponding estimations of the source locations. We quantified the uncertainty of the dipole locations using the enclosing volume of the dipole locations and the maximum width of the dipole cluster. Shifting fiducial locations 1.5 cm increased the uncertainty of the dipole locations to span a volume >1 cm³ and about 2.5 cm wide. Results suggest that the fiducials need to be digitized accurately within at least 0.5 cm of the absolute actual fiducial location to limit the uncertainty of a dipole location to <1 cm. Additionally, we used random fiducial shift combinations to estimate the effects of combinations of the fiducial shifts on dipole location estimation. This analysis showed that dipole locations were within the bounds of our dipole estimation uncertainty volumes. Based on the outcomes, we suggest marking fiducials carefully before placement of the cap and to use a digitization method with an accuracy of <0.5 cm.

Keywords: Neural Interfaces - Computational modeling and simulation, Neural Interfaces - Neural stimulation
Abstract: Magnetic stimulation using commercial transcranial magnetic stimulators (TMS) and coils is becoming an established tool for neurostimulation. However, when applied at the lumbar region it is not clear which neural structures are stimulated and especially, if the spinal cord (SC) can be stimulated. Computational modeling with realistic human body models is a promising tool to understand better the basic mechanisms of the stimulation. In this study we have used a realistic model to calculate the current density (J) distribution and magnitude under different output power levels of a commercial stimulator to describe the electromagnetic effects on the different tissues. Our results suggest that spinal cord stimulation is possible. However, significant muscle contraction is produced due to the high stimulation needed, which might make this stimulation non-practical. The spatial resolution of this technology is very poor to stimulate specific parts of the SC only. Although the stimulation aims at SC structures, we observed that most of the current does not reach the SC, but the cerebrospinal fluid (CSF). All together, these results represent a first step towards understanding and optimizing magnetic transpinal stimulation

Keywords: Neuromuscular Systems - EMG models, processing and applications, Neurological disorders - Stroke, Neuromuscular Systems - Neurorehabilitation
Abstract: Botulinum toxin (BT) is a safe and effective neuromuscular blocking agent that is clinically utilized to reduce spasticity after stroke. It is often injected repeatedly at a minimum of 12-week intervals. BT targets the neuromuscular junction and chemically denervates muscle fibers from their corresponding spinal motoneurons (MN). We explored the effect of BT on the amplitude of the smallest tendon tap force (i.e. force threshold) required to elicit a detectable biceps brachii surface electromyogram (sEMG) reflex response. We hypothesized that after BT injection, the force threshold would increase due to a decrease in available efferent activation. Two chronic stroke survivors were recruited. Data were collected before and up to 18 weeks after BT injection. For each subject, sEMG responses were analyzed using high-density sEMG (HDsEMG) recordings, and the threshold tapping forces were identified and mapped for all channels. Unexpectedly, median threshold forces (MTF) decreased post-BT (B01: 30%, B02: 50%). However, after the initial decrease, MTF then increased progressively compared to pre-BT and peaked around 12 weeks (B01: ~4 folds, B02: 50%). This is likely because post-BT, fewer available muscle fibers would require larger tapping forces to evoke detectable sEMG responses. In the last recording session (> 12 weeks), MTF did not return to pre-BT levels, indicating that successive botulinum toxin injections may still be effective if spaced much further apart in time.

Keywords: Human performance, Brain-Computer/Machine Interface - Robotics applications
Abstract: Electroencephalogram (EEG)-based prediction of driver emergency braking intention can help develop an assistance system to improve driving safety for brain-controlled vehicles. However, existing studies are focused on how to build an individual detection model for each participant. In this paper, to build a universal model, a convolutional neural network (CNN) is used to extract the features of brain signals and build the universal model. Experimental results from 13 subjects show that the proposed CNN-based method outperforms the linear discriminant analysis (LDA)-based method and has a comparable performance with individual models. This work lays a foundation for future developments of an EEG-based universal model of driver emergency braking intention detection.

Keywords: Brain Stimulation - Transcranial direct current Stimulation (tDCS), Neural signal processing, Neural Signal Processing - Blind source separation
Abstract: Non invasive brain stimulation is a widely used technique for several applications, generally aimed at modulating brain activity and thus behavior. While the behavioral effects can be monitored during the application of the stimulation, the electrophysiological correlates, such as electroencephalography (EEG), cannot, because the stimulation artifact dramatically affects the recorded signals. Here we addressed this problem and we analyzed the artifact that transcranial alternating current stimulation (tACS) leaves on EEG traces. We found that the stimulation noise adds itself non-linearly to the EEG signal and spectral analysis revealed a peak centered at the stimulation frequency. We then created a synthetic dataset by adding to real EEG traces numerically generated signals, matching the characteristics of the tACS artifact. We used this data to test a set of artifact removal techniques based on blind source separation (BSS) methods and wavelet decomposition and we found that the best performing technique is independent component analysis (ICA).

Keywords: Brain physiology and modeling - Neural dynamics and computation, Neural Signal Processing - Nonlinear analysis
Abstract: Electrical field (EF) is a popular tool for both basic research and clinical applications, its actions on neuronal activities have been investigated from physiological mechanism and dynamics. However, few studies explore its modulatory influence on neuronal computation from the point of view of dendritic sublinear integration caused by passive dendrites which play an important role in neuronal computation. Here with a reduced biophysical model this problem is explained by observing the impact of EF on neuronal computation and dendritic sublinear operation. It is found that the positive EF results in more linear dendritic sublinear integration because of hyperpolarization in distal dendrites together resulting in higher neuronal excitability in neuronal computation but negative EF inhibits this ability due to more pronounced dendritic sublinear operation resulting from the hyperpolarization in distal dendrites. Further, we explain the modulation of positive EF on dysfunctional neuron combining with Feature Binding Problem. This work builds the gap between neuronal computation and dendritic sublinear operation, which is helpful to understand the modulation of EFs on brain functions.

Keywords: Brain Functional Imaging - Connectivity and Network, Brain Functional Imaging - Classification, spatiotemporal dynamics, Brain Functional Imaging - EEG and Evoked Potentials
Abstract: Disrupted functional connectivity patterns have been increasingly used as features in pattern recognition algorithms to discriminate neuropsychiatric patients from healthy subjects. Deep neural networks (DNNs) were employed to fMRI functional network classification only very recently and its application to EEG-based connectome is largely unexplored. We propose a DNN with deep belief network (DBN) architecture for automated classification of schizophrenia (SZ) based on EEG effective connectivity. We used vector-autoregression-based directed connectivity (DC), graph-theoretical complex network (CN) measures and combination of both as input features. On a large resting-state EEG dataset, we found a significant decrease in synchronization of neural oscillations measured by partial directed coherence, and a reduced network integration in terms of weighted degrees and transitivity in SZ compared to healthy controls. The proposed DNN-DBN significantly outperforms three other traditional classifiers, due to its inherent capability as feature extractor to learn hierarchical representations from the aberrant brain network structure. Combined DC-CN features gives further improvement over the raw DC and CN features alone, achieving remarkable classification accuracy of 95% for the theta and beta bands.

Keywords: Neural Interfaces - Neural microsystems and Interface engineering, Neurological disorders - Epilepsy, Neurological disorders - Diagnostic and evaluation techniques
Abstract: A fast and low-power embedded classifier with small footprint is essential for real-time applications such as brain-machine interfaces (BMIs) and closed-loop neuromodulation for neurological disorders. In most applications with large datasets of unstructured data, such as images, deep neural networks (DNNs) achieve a remarkable classification accuracy. However, DNN models impose a high computational cost during inference, and are not necessarily ideal for problems with limited training sets. The computationally intensive nature of deep models may also degrade the classification latency, that is critical for real-time closed-loop applications. Among other methods, ensembles of decision trees (DTs) have recently been very successful in neural data classification tasks. DTs can be designed to successively process a limited number of features during inference, and thus impose much lower computational and memory overhead. Here, we compare the hardware complexity of DNNs and gradient boosted DTs for classification of real-time electrophysiological data in epilepsy. Our analysis shows that the strict energy-area-latency trade-off can be relaxed using an ensemble of DTs, and they can be significantly more efficient than alternative DNN models, while achieving better classification accuracy in real-time neural data classification tasks.

Keywords: Neural Interfaces - Neural stimulation, Neural Interfaces - Computational modeling and simulation
Abstract: Selectively stimulating neuron types within the brain can enable new treatment possibilities for neurological disorders and feedback in brain-machine interfaces. Prior computational work has shown that by choosing a sinusoidal signal with appropriate amplitude and frequency, one can obtain one-directional stimulation, i.e., one can stimulate a mammalian inhibitory neuron without stimulating an excitatory neuron. However, bidirectional selectivity is not achievable using just sinusoidal inputs. In this work, which is also computational, we design novel current waveforms to achieve this bidirectional selectivity. To do so, we explicitly exploit the non-linearity of neuronal membrane potential in response to stimulating currents. These current waveforms are able to stimulate either of the two neuron-types without stimulating the other. Further, we are also able to design a waveform which stimulates both neurons. Moreover, we can ensure a relatively high firing rate (~100 Hz) when a neuron-type is targeted for stimulation.

Keywords: Brain-computer/machine Interface, Motor neuroprostheses, Neural signal processing
Abstract: Local field potential (LFP) has gained increasing interest as an alternative input signal for brain-machine interfaces (BMIs) due to its informative features, long-term stability, and low frequency content. However, despite these interesting properties, LFP-based BMIs have been reported to yield low decoding performances compared to spike-based BMIs. In this paper, we propose a new decoder based on long short-term memory (LSTM) network which aims to improve the decoding performance of LFP-based BMIs. We compare offline decoding performance of the proposed LSTM decoder to a commonly used Kalman filter (KF) decoder on hand kinematics prediction tasks from multichannel LFPs. We also benchmark the performance of LFP-driven LSTM decoder against KF decoder driven by two types of spike signals: single-unit activity (SUA) and multi-unit activity (MUA). Our results show that LFP-driven LSTM decoder achieves significantly better decoding performance than LFP-, SUA-, and MUA-driven KF decoders. This suggests that LFPs coupled with LSTM decoder could provide high decoding performance, robust, and low power BMIs.

Keywords: Motor Neuroprostheses - Epidural Stimulation, Neurological disorders - Stroke, Neurorehabilitation
Abstract: Epidural electrical stimulation (ECS) combined with rehabilitative training could improve behavioral function for stroke both in clinical and animal studies. Though the primary motor cortex (M1) ECS has some effects on recovery, the performances are not good enough and better sites for stimulation need further investigation. In this study, we compared the effects of ECS on M1 and premotor cortex (PM). Twelve rats were trained to the single-pellet retrieval (SPR) task before receiving photothrombotic ischemia model and electrodes implantation. They were randomly distributed to three groups: M1 stimulation, PM stimulation and no stimulation (Control) groups. The 100Hz monopolar cathodal stimulation was given concurrent with rehabilitation training over recovery days. The performance on SPR task were scored for analysis. Our results showed that M1 ECS had a better effect on behavioral improvement compared to the other two groups.

Keywords: Brain-computer/machine Interface
Abstract: Steady-state visual evoked potential (SSVEP)-based brain computer-interfaces (BCIs) have shown its robustness in facilitating high-efficiency communication. State-of- the-art training-based SSVEP decoding methods such as extended Canonical Correlation Analysis (CCA) and Task- Related Component Analysis (TRCA) are the major players that elevate the efficiency of the SSVEP-based BCIs through a calibration process. However, due to notable human variability across individuals and within individuals over time, calibration (training) data collection is non-negligible and often laborious and time-consuming, deteriorating the practicality of SSVEP BCIs in a real-world context. This study aims to develop a cross-subject transferring approach to reduce the need for collecting training data from a test user with a newly proposed least-squares transformation (LST) method. Study results show the capability of the LST in reducing the number of training templates required for a 40-class SSVEP BCI. The LST method may lead to numerous real-world applications using near-zero-training/plug-and-play high-speed SSVEP BCIs.

Keywords: Brain-computer/machine Interface, Neural Interfaces - Computational modeling and simulation
Abstract: Characterizing the neural encoding of behavior remains a challenging task in many research areas due in part to the complex and noisy spatiotemporal dynamics of evoked brain activity. When modeling these neural encodings, it is important to separate robust, behaviorally relevant signals from background activity, which often contains signals from irrelevant brain processes and decaying information from previous behavioral events. To achieve this separation, we develop a twobranch State Space Variational AutoEncoder (SSVAE) model to individually describe the instantaneous evoked foreground signals and the context-dependent background signals. We modeled the spontaneous speech-evoked brain dynamics using smoothed Gaussian mixture models. By applying the proposed SSVAE model to track ECoG dynamics in one participant over multiple hours, we find that it predicts speech-related dynamics more accurately than other latent factor inference algorithms. Our results demonstrate that separately modeling the instantaneous speech-evoked and slow context-dependent brain dynamics can enhance tracking performance, which has important implications for the development of advanced neural encoding and decoding models in various neuroscience subdisciplines.

Keywords: Brain-computer/machine Interface, Neural signal processing, Motor learning, neural control, and neuromuscular systems
Abstract: Neuronal activities of two monkeys are analyzed to explore whether moving tasks can trigger specific motor-controlled signal as response. A multi-array contains 96-channel electrodes was used to record the long-term neuronal activity. In the time domain analysis, we found the cues of tasks can stably stimulate some spindle waves which do not exist in the task-free stages. Most spindles sparkled after the participant issued real and correct actions as responses to the cues of tasks, indicating the spindle as a motor-controlled wave-marker. In a further time-frequency analysis, the motor-controlled spindles indicate their high intensity in the spectrogram of  band. A spindle detection algorithm based on SVM was constructed to annotate the motor-controlled spindles automatically. Considering the individual and the day-today differences, the subject-specific model under leave-one-day-out cross-validation was framed to guarantee the model robustness. The algorithm performed well with F1 scores at 83:73  13:08% and 79:2616:95% on the two monkeys, respectively. This finding might be used in the future as a new marker in certain research areas, e.g., mental detection, autism spectrum disorder (ASD).

Keywords: Neural Signal Processing - Time frequency analysis, Human Performance - Modelling and prediction, Neurological disorders - Stroke
Abstract: Time-frequency coherence analysis between EEG and EMG signal represents a valuable tool to gain insight into neural mechanisms underlying motor control. However, for self-paced movements, the variable inter-trial duration limits its proper use. To overcome this obstacle, we propose a time-normalizing approach and test it on both simulated and experimental data recorded during elbow extension movements performed by a post-stroke subject. Results show that the proposed time-normalization improves both the consistency and the accuracy of time-frequency coherence calculation, detection and quantification. The proposed time-normalization overcomes a major limitation to generalization of coherence analysis and can be suggested as an essential step to perform for coherence analysis in presence of variable inter-trial duration.

Keywords: Neural Interfaces - Neural stimulation, Neural signal processing, Motor Neuroprostheses - Epidural Stimulation
Abstract: This Spinal cord injury is a type of central nerve injury that permanently changes patients’ ability of self-care and quality of life. In recent years, some scholars have attempted to rebuild the lost function due to spinal cord injury through the microelectronic neural bridge. In order to solve the problem of electrode implantation in the microelectronic neural bridge scheme, this study proposed a new method for detecting the conduction pathway of neural signals from rat primary motor cortex in spinal cord with electronic detection technology. Functional electrical stimulation was performed on the primary motor cortex of the rat, and the spinal nerve signals corresponding to the T9-T12 spinal segments were recorded. Correlation and delay between recording sites were analyzed by cross-correlation function to verify whether the neural signals were from the same source, which was the cortex stimulation site. Then the recording sites can be connected to form the conduction pathway. The experiment was carried out on SD rats, and the experimental results were consistent with the expected results, indicating that the method is feasible. And the conduction of rat primary motor cortex nerve signals in spinal cord can be detected from the perspective of electronic informatics. Subsequent experiments can provide practical data for electrode implantation in microelectronic nerve bridging experiments.

Keywords: Human performance, Human Performance - Cognition
Abstract: With the development of non-contact heart rate variability (HRV) measure, fast cognitive stress estimation via HRV has many potentially powerful applications, especially combined with mobile smart devices. This study proposed two novel HRV features, L1-Norm and L2-Norm, based on the spectrum sparsity measure, and evaluated the effect of the short HRV measure (15 s) on their stress detection performances. An experiment was conducted with nine participants under rest and cognitive stress condition. The results of two metrics, equal error rate (EER) and area under the curve (AUC) from receiver operating characteristic (ROC), showed that the prediction performance of the proposed features was comparable to the heart rate and better than the ratio of low-frequency power to high-frequency power. This study indicated that it was possible to estimate cognitive stress with short HRV measures, and it was necessary to employ the feature combinations for the performance of a single feature was not consistent on all the subjects.

Keywords: Brain Stimulation-Deep brain stimulation, Neurological disorders - Diagnostic and evaluation techniques, Human Performance - Modelling and prediction
Abstract: DBS surgery is considered for Parkinson’s Disease patients when motor complications and consequent quality of life is no longer acceptable on optimal medical therapy prescribed by neurologists. Within the operating room, the electrode placement with the best clinical outcome for the patient is quantitatively assessed via the wrist rigidity assessment. A subjective scale is used, influenced by the neurologists’ perception and experience. Our research group has previously designed a novel, comfortable and wireless system aiming to tackle this subjectivity. This system comprised a gyroscope sensor in a textile band, placed in the patients’ hand, which communicated its measurement to a Smartphone via Bluetooth. During the wrist rigidity evaluation exam, a signal descriptor was computed from angular velocity (V) and a polynomial mathematical model was used to classify the signals using a quantitative scale of rigidity improvement. In this present work, we aim to develop models that consider the 3-gyroscope-axes to acquire V and the cogwheel rigidity. Our results showed that y-gyroscope-axis remains the best way to classify the rigidity reduction, showing an accuracy of 78% and a mean error of 3.5%. According to previous results, the performance was similar and the decrease of samples to extract the V features did not compromise system performance. The cogwheel rigidity did not improve the previous model and other gyroscope-axis beyond the y-axis decreased system performance.

Keywords: Sensory Neuroprostheses, Neuromuscular Systems - EMG models, processing and applications, Neural Interfaces - Neural stimulation
Abstract: Electrical stimulation of peripheral nerves can elicit sensations in the residual and missing limbs of amputees. The goal of this study is to determine how electrical stimulation activates peripheral neural pathways by examining the relationship between elicited sensations and activation of residual muscles proximal to the amputation. In this study, high-density composite flat interface nerve electrodes were implanted on the residual nerves of a participant with a trans-tibial amputation. The sensations elicited by applying stimulating pulses to selected contacts within the electrodes were classified into three categories: tactile sensation in the residual limb, tactile sensation in the missing limb, and feeling of muscle contraction in the residual limb. Surface electromyograms (EMG) were recorded from below-knee muscles in the residual limb. Initial results suggest direct peripheral afferent activation induces tactile sensation in the participant’s missing limb. No significant elevated EMG was observed in any of the below-knee muscles when the perceived tactile sensation was in the missing limb. However, EMG activity was observed in at least one of the muscles when the perceived tactile sensation was in the residual limb. Our findings suggest that elicited tactile sensations in the missing limb are largely due to direct activation of afferent fibers. However, muscle contraction sensation in the residual limb may result from afferent activation by muscle recruitment.

Keywords: Human Performance - Drowsiness and microsleeps, Human Performance - Fatigue, Brain Functional Imaging - EEG and Evoked Potentials
Abstract: Drowsiness of the driver is one of the major causes of car accidents. In the last years, the extraction of an objective drowsiness correlate to prevent drowsiness related accidents has gained more and more attention. The aim of the current study was the assessment of a time-resolved quantification of drowsiness correlates in multimodal psychophysiological data obtained during simulated driving. A further part of the study was the feasibility assessment of a drowsiness categorization by means of psychophysiological data. The heart rate, the heart rate variability, the skin conductance level and the relative band power (θ, α, β) as well as the α/β power ratio of the EEG were examined as objective correlates of drowsiness in 30 participants. The results show that, besides the skin conductance level, all investigated time-resolved psychophysiological measures showed the expected trends described in the literature (ANOVA test, p < 0.01). The highest Pearson’s correlation coefficient calculated between the subjective rating scale and objective index of drowsiness was noticeable for the α/β power ratio in the parietal lobe. Thus, the drowsiness categorization was only realized for this index. The individual results show that this time-resolved psychophysiological index of drowsiness can reflect the participants’ actual mental state. It is concluded, that the time-resolved α/β power ratio is a suitable objective index for drowsiness.

Keywords: Motor Neuroprostheses - Neuromuscular stimulation, Human Performance - Modelling and prediction, Motor Neuroprostheses - Robotics
Abstract: We introduce capability maps visualizing the abilities of the arm of a person with a cervical spinal cord injury activated by functional electrical stimulation (FES). We map the arm’s workspace at different wrist positions using a person-specific arm model based on force data gathered during interactions with a robot. We describe four maps: 1) a map of the maximum force the person can produce in one direction, 2) a map of wrist configurations that FES can hold against gravity and other passive forces, 3) a map of the maximum force the person can apply in all directions, and 4) a map of the directions the arm can move with FES. To demonstrate these maps we applied electrical stimulation to nine muscle groups of a person with high tetraplegia, measured the resulting force with a robot attached to the person’s wrist, created a Gaussian process regression model relating the forces to the wrist positions, and used this model to create the four capability maps. The results are 2D images displaying the arm’s force production and movement capabilities for a person with high tetraplegia as a function of wrist position. As these maps predict functional benefits of specific interventions, they can reduce risk in developing new interventions to restore function to people with whole-arm paralysis.

Keywords: Brain physiology and modeling - Neuron modeling and simulation, Neural Signal Processing - Nonlinear analysis, Neural signal processing
Abstract: A morphologically realistic and anisotropic model of CA3 pyramidal neuron was developed to determine the synaptic integration modes the neuron is able to perform. Linearity and nonlinearity were identified in different synaptic locations with varying active mechanisms such as the presence of ionic channels in the dendritic arbor and the types of receptors in the synapse. Quantification of synaptic integration was performed using paired-pulse stimulation protocol and subthreshold input/output (sI/O) transformation. Results show that the mode of synaptic integration is location-dependent while the linearity or nonlinearity in the integration is mainly influenced by the active channels and receptors in the neuron. A simple CA3 pyramidal neuron could therefore be considered as a two-layer network of neurons. Its computational power is enhanced by performing linear, sublinear and supralinear integration.

Keywords: Human Performance - Modelling and prediction, Motor Neuroprostheses - Prostheses, Neural Signal Processing - Blind source separation
Abstract: The aim of this study was to characterize the accuracy in the identification of motor unit discharges and to estimate wrist kinematics from motor unit behaviors. High-density electromyography (EMG) of forearm muscles and wrist torques in three degrees-of-freedom (DoFs) were recorded concurrently during wrist movements of 8 able-bodied subjects. The EMG signals were decomposed into motor unit spike trains (MUSTs) with a blind-source separation algorithm. Two methods based on principal component analysis and regression model respectively were proposed to estimate wrist torques in three DoFs. On average, 19±6 MUSTs were identified in each trial with accuracy > 85%. For all conditions, the Pearson correlation coefficient between estimations and recordings was always > 0.8. The average normalized root mean square error of two methods was 0.15±0.03 and 0.16±0.05, respectively. These results indicated the identification of motor unit behaviors with high confidence, thus having the potential to be practical approaches for prosthesis control.

Keywords: Motor Neuroprostheses - Prostheses, Motor learning, neural control, and neuromuscular systems
Abstract: The continuous estimation of kinematics or kinetics from electromyography (EMG) signals is essential for intuitive control of prostheses and other human-machine interfaces based on bioelectrical signals. In this preliminary study, we concentrate on the continuous estimation of wrist torques under isometric contraction of three separate degrees-of-freedom (DoFs) with a stack-autoencoder based deep neural network. With this kind of deep neural network, features used for regression could be extracted autonomously other than in hand-crafted manner. Five subjects participated in the experiment under a visual feedback guide interface, in which surface EMG signals and wrist torques were concurrently recorded. It is shown that a promising estimation performance is achieved in all three DoFs. The outcomes of this study demonstrate the feasibility of this method on continuous estimation of wrist torques and reveal the potential for further being extended into continuous and simultaneous myoelectric control.

Keywords: Brain Functional Imaging - Multimodal, Brain physiology and modeling - Neural dynamics and computation, Neural signal processing
Abstract: In this paper, we use convolutional neural networks (CNNs) to model/capture the relationship between simultaneously acquired EEG and fMRI. Specifically we use CNNs to implement "neural transcoding" - i.e. generating one neuroimaging modality from another - from EEG to fMRI and vice versa. The novelty of our approach lies in its ability to resolve the source space without prior hemodynamic and leadfield estimation. The two CNNs, one for EEG-to-fMRI and the other fMRI-to-EEG transcoding, are coupled in their source space representations, and given their architecture are able to capture both linear and non-linear transformations that map two imaging modalities into a common neural source space. We present results on simulated simultaneously acquired EEG-fMRI data and show the performance of mapping each modality to the other, the accuracy of recovering source space, and the effects of noise and variation in the simulated acquisition parameters, such as MRI slice timing, on the results.

Keywords: Brain Stimulation-Deep brain stimulation, Brain-Computer/Machine Interface - Biofeedback, Neurorehabilitation - Wearable systems
Abstract: Essential tremor is a prevalent neurodegenerative movement disorder. Deep brain stimulation represents a highly effective means of treatment, especially for scenarios for which traditional medical intervention is no longer feasible. One of the major post-operative challenges is the determination of an optimal set of tuning parameters. Optimizing the deep brain stimulation parameters can impart a time-intensive task to the clinician. The smartphone in the context of a wearable and wireless inertial sensor system offers the capability to objectively quantify the characteristics of the tremor. Machine learning in conjunction with a wearable and wireless inertial sensor system, such as a smartphone, can distinguish between disparate states, such as deep brain stimulation in ‘On’ and ‘Off’ status. Multiple machine learning classification techniques are available, such as the multilayer perceptron neural network, support vector machine, K-nearest neighbors, logistic regression, J-48 decision tree, and random forest. The objective of this research endeavor is to evaluate these six machine learning classification algorithms for classification of deep brain stimulation regarding ‘On’ and ‘Off’ status for Essential tremor during a reach and grasp task.

Keywords: Brain Functional Imaging - fMRI, Neural Signal Processing - Blind source separation, Brain Functional Imaging - Connectivity and Network
Abstract: Task-induced variations in neural activity and their effects on the topological architecture of intrinsic connectivity networks (ICNs) of the brain are still a matter of ongoing research. In this exploratory study, we used spatial independent component analysis (ICA) as a data-driven technique to characterize ICNs related to two different tasks in healthy subjects who underwent 3T fMRI. The fMRI tasks consisted of (a) a viscerosensory stimulation of an internal organ (interoceptive task), and (b) passive viewing of emotionally expressive faces and pictures from the International Affective Picture System (exteroceptive emotion task). Comparison of the network volumes and peak activations during each task condition demonstrated that changes in ICN volume and corresponding peak activation differed between the interoceptive and exteroceptive emotion tasks when compared to the baseline rest. Further, salience network was the most task-activated ICN for both fMRI task conditions. However, different spatial characteristics were observed between the salience networks derived from the interoceptive task and the one derived from the exteroceptive emotion task. This study is a step in the direction of better understanding the influence of task condition on ICN topology. Future research with a larger sample size and task variations should delve deeper into what aspects of network topology really matter, with further investigations regarding the observed differences due to gender and age.

Keywords: Human Performance - Fatigue, Brain Functional Imaging - Connectivity and Network, Human performance
Abstract: Mental fatigue is a serious problem in contemporary society, which affects personal productivity, safety in life and mental health, and it has become a convergent research topic in the nascent field of neuro-ergonomics. However, we still lack a clear understanding of the neural mechanism of mental fatigue, which contradicts the realistic demands. In this study, considering the dynamic accumulation of mental fatigue, temporal brain networks were estimated from 20 participants based on functional images. First, the dynamic connection network was constructed through functional imaging data in a sustained visual attention task. Quantitative analysis of the dynamic functional connectivity (FC) was then performed in the 3D spatiotemporal architecture using our recently introduced temporal efficiency method. We found that the temporal global efficiency was decreased with the increase of mental fatigue, suggesting a significantly disintegrated spatiotemporal topology of dynamic FC underpinning mental fatigue. The finding therefore expands the research of related static brain networks and provides new evidence for the resource hypothesis of mental fatigue.

Keywords: Neural Signal Processing - Time frequency analysis, Neural Interfaces - Recording
Abstract: — Overactivity of the sympathetic nervous system (SNS) is associated to several cardiovascular and metabolic dysfunctions, such as hypertension and insulin resistance. Indirect biochemical measurements and surgical manipulations have provided preliminary evidences about a crucial role of the Carotid Sinus Nerve (CSN) in generating the SNS overactivity. However, CSN and SNS neural activities and their interplay have not been yet characterized in healthy and pathological conditions. Understanding this relationship is key for the development of electroceutical approaches to deliver therapeutic neuromodulation to the autonomic nervous system and restore insulin sensitivity. Here we show that early type 2 diabetes rats present a high frequency shift in both CSN and SNS neural activities with respect to control animals. This feature could be an important neural signature characterizing type 2 diabetes. Moreover, we show that CSN resection in early type 2 diabetes rats abolishes SNS high frequency shift confirming that normal SNS activity and insulin sensitivity may be recovered by CSN activity suppression. These findings shed new light on the pathological neural changes within the autonomic nervous system in type 2 diabetes. Moreover, they pave the way for electrical monitoring of the metabolic state of diabetic patients, a key first step for the development of electroceutical therapies.

Keywords: Neural Interfaces - Biomaterials
Abstract: In this paper, we demonstrate, for the first time, how we can pattern grid networks of human hNT astrocytes on parylene-C/SiO2 substrates down to the single cell level. We demonstrate that the functionality of the astrocyte networks by calcium release on the introduction of ATP and show that this is similar to that of a control sample. Thus, we demonstrate that the parylene-C/SiO2 platform is a viable way to investigate the glial network at the single cell level.

Keywords: Neural Interfaces - Biomaterials, Neural signal processing
Abstract: To further investigate the communication in astrocytic networks, in vitro astrocytes were patterned in simple networks by using parylene-C/SiO2 platform. We designed a 2 x 2 small network on parylene-C/SiO2 platform and demonstrated, for the first time, that the intracellular calcium response of a single human hNT astrocyte stimulated by UV laser pulses can be transmitted to neighboring astrocytes in this 2 x 2 small network. The calcium responses of the astrocytes network were recorded and analyzed.

Keywords: Human Performance - Fatigue, Motor learning, neural control, and neuromuscular systems
Abstract: The studies of grip-load force and muscle fatigue characteristics are of great significance in rehabilitation and sports medicine. Functional near-infrared spectroscopy (fNIRS) is generally used to investigate cerebral oxygenation changes during a motor task. In this study, the effects of force load and muscle fatigue on fNIRS signal features during a handgrip voluntary contraction task were quantitatively investigated. Twenty-four healthy subjects performed isometric grasping contractions at the conditions of 30% and 50% of the maximal voluntary contraction (MVC) with his/her dominant or non-dominant hand. Experimental results showed that the average change rate of oxygenated hemoglobin (HbO2) concentrations at the condition of 30% MVC hand muscle strength was significantly smaller than that at the condition of 50% MVC hand muscle strength. However, the total change of HbO2 concentration in conditions of the 30% and 50% MVC hand muscle strength were in good agreement. The total change of HbO2 concentration and the average change rate of HbO2 concentration were not significantly different between the dominant and non-dominant hands.

Keywords: Brain-computer/machine Interface, Neural signal processing, Brain Functional Imaging - EEG and Evoked Potentials
Abstract: The research in non-invasive Brain-Computer Interface (BCI) has led to significant improvements in the recent years. However, the user experience and the BCI illiteracy problem remain key issues to address for obtaining robust and resilient applications. In this paper, we address the choice of the time segment for the detection of steady state visual evoked potential (SSVEP) detection. The choice of this parameter is typically fixed and has a direct influence on the accuracy of detection, and therefore the information transfer rate. We propose to shift the problem of the time segment to the choice of the threshold for determining if a response has been properly detected. We consider an open-dataset of 10 participants to validate the rationale of the approach. The results support the conclusion that an adaptive time segment can lead to a better ITR on average across participants compared to a fixed time segment equal to the average of the mean adapted time segment for each subject. The ITR increases from 68.87 to 75.39 bpm with 12 targets, and from 54.20 to 72.66 bpm with 6 targets, highlighting the need of adaptive methods.

Keywords: Neural signal processing, Clinical neurophysiology, Brain Functional Imaging - EEG and Evoked Potentials
Abstract: Brain-function monitoring is a difficult problem as the particular characteristics of the brain of an individual must be taken into account. Therefore, some particular aspects of a single brain may not be properly modeled through the use of data from other people. In addition, there are situations where brain monitoring has to happen a single time and through multiple conditions. In such a case, it is not possible to calibrate a system with previous signals from this same person as the new incoming states are unknown. Moreover, as new unseen conditions happen over time, it is not possible to create a model in relation to these incoming conditions where no information is available. For all these reasons, the use of one-class classifiers (OCC) or novelty detection techniques for brain state monitoring provides a solution for determining a shift in the condition of the brain state of an individual. In this paper, we propose to investigate the brain states in different sedation states (baseline, mild, moderate, and recovery) with electroencephalogram (EEG) signals from seven adult participants by using bandpower features in five frequency bands on 20 sensors. We propose to use the one-class nearest neighbor classifier to evaluate the detection of a change from a sedation state to posterior state. We propose to set the threshold based on the 90th percentile of the current class. The results support the conclusion that an f-score of 87.56+/-15.85% can be obtained from the b

Keywords: Neural signal processing, Brain-computer/machine Interface, Brain physiology and modeling
Abstract: Recent work has shown that it is possible to decode aspects of continuously-spoken speech from electrocorticographic (ECoG) signals recorded on the cortical surface. The ultimate objective is to develop a speech neuroprosthetic that can provide seamless, real-time synthesis of continuous speech directly from brain activity. Instead of decoding acoustic properties or classes of speech, such a neuroprosthetic might be realized by decoding articulator movements associated with speech production, as recent work highlights a representation of articulator movement in ECoG signals. The aim of this work is to investigate the neural correlates of speech-related lip movements from video recordings. We present how characteristics of lip movement can be decoded and lip-landmark positions can be predicted.

Keywords: Neural signal processing, Motor Neuroprostheses - Neuromuscular stimulation
Abstract: This work proposes the classification of motor imagery signals for brain-machine interfaces with functional electrical stimulation in the quadriceps muscle. Five volunteers participated in the test, 3 healthy participants, aged 28 ± 3 years, and 2 paraplegic volunteers, aged 43 (ASIA-B, C7 level - 16 years) and 47 (ASIA-A, T7 level - 20 years) years respectively. In total, each participant performed 90 repetitions of motor imaging of the lower limb under electrical stimulation, with frequencies of 20Hz, 35Hz, and 50Hz and current amplitude of 20mA. The patterns were analyzed off-line and submitted to the classification architectures after application of spatial filtering to extract the characteristics. The classification of the patterns was performed using the architectures: (i) Linear Discriminant Analysis (LDA), (ii) Multilayer Perceptron (MLP), and (iii) Support Vector Machine (SVM). To validate the proposal, the performance was compared between the classifiers through the accuracy of cross validation, variance, precision, and sensitivity. With the SVM classifier, the best accuracy percentage was 86.5%. These results are promising and the trained architectures are feasible for implementation in neuroprostheses with lower computational resources.

Keywords: Brain-computer/machine Interface
Abstract: In recent several decades, brain-computer interface (BCI) technology continually yield fruitful results. The electrocorticogram (ECoG) has attracted considerable interest because of its advantages of higher signal-to-noise ratio and greater long-term stability than electroencephalography (EEG) signals. We present an optimal scheme of ECoG signals for motor imagery (MI) classification. The time-frequency features are first extracted by the modified S-transform (MST) algorithm, and then a classifier is trained by using the support vector machine (SVM). In addition, channel selection is performed to reduce the computational complexity of MI-based BCI scheme. This method was tested on BCI Competition III dataset I. The MST coupled with the SVM can obtain the satisfactory classification of 95%. Channel selection can greatly reduce the computational burden of classification and enable this scheme to classify MI tasks in real time.

Keywords: Neural Interfaces - Neural stimulation, Motor Neuroprostheses - Neuromuscular stimulation, Brain Stimulation - Optogenetics
Abstract: Peripheral optogenetic stimulation of motor activity offers enticing advantages over traditional functional electrical stimulation for the purposes of reanimating paralyzed muscles. When facilitated by intramuscular injection of viral gene therapy constructs, however, the process of transducing light sensitive ion channels along motor nerves faces several challenges including uptake of the virus at the neuromuscular junction as well as evasion of both virus and expressed gene products from the immune system. These hurdles to successful peripheral motor gene therapy are often amplified when attempting to translate these techniques to non-human primates. In this study, we examined the efficacy of a systemic immunosuppression regimen and use of a designer adeno-associated virus in prolonging functional opsin expression in targeted peripheral nerves of a macaque. Using a regimen of daily cyclosporine and either an intramuscular or intraneural injection of an rAAV2-retro based vector, we observed functional nerve expression of ChR2 via EMG activity locked to optical stimulation of a targeted nerve for up to 24 weeks post-injection. Throughout this experiment, we observed a gross timeline of expression including an initial increase of ChR2 expression over 9-13 weeks followed by an eventual decline after cessation of the immunosuppression regimen. These results suggest a potential strategy for successful translation of peripheral motor gene therapy to human subjects.

Keywords: Neural Interfaces - Neural stimulation, Neural Interfaces - Recording, Neurorehabilitation - Neurofeedback
Abstract: We have developed a real-time electrocolonogram monitoring and stimulation system for promoting peristaltic movements of colon. The system monitored electrocolonogram and applied appropriate electrical stimulation to the serosal surface of colon in caudad direction. Stimulation parameters were determined by the stimulation controller in real time based on the measured electrocolonogram. As a proof of concept, animal experiment was performed using a mouse model. We successfully monitored electrocolonogram from colon and generated its peristaltic movement using electrical stimulation applied to the serosal surface of colon. The results demonstrated that the developed system can be used to monitor the colonic activity and apply stimulation to promote peristaltic movement of the colon. This system can be used to develop a closed-loop monitoring and stimulation system for a timely stimulation of colon.

Keywords: Sensory Neuroprostheses - Visual, Neural signal processing, Brain Physiology and Modeling - Neural circuits
Abstract: The visual system is one of the most important pathways of obtaining information for human being and other animals. The retina is responsible for initial processing of visual information and transmitting signals to the second processing system by using the spiking activity patterns. This paper implements a retinal spiking neural network based on field-programmable gate array (FPGA), and uses different scopes of light stimulation to stimulate the digital retinal network and induce different spiking activities. The retina neural network contains 96 neurons, which uses Hodgkin-Huxley type neuron model to build neural network using three-layer feedforward neural network structure. The neural network is implemented using Cyclone IV EP4CE115 FPGA, and uses OV7620 camera to obtain external signals. The state machine control the input information of the retina system, and the firing patterns are finally displayed on oscilloscope device. Experimental results show that the proposed digital retinal network can generate the dual-peak response of the retinal ganglion cells. This work is meaningful for the design of the retina prostheses and is helpful for the investigation of the underlying mechanisms of the retinal activities.

Keywords: Neural signal processing, Neural Signal Processing - Time frequency analysis
Abstract: It has been reported that meditation is a progressive procedure to achieve a thoughtless and transcendental deep meditation state. However, it remains difficult to know whether and when the practitioner has achieved deep meditation state for lack of a reliable and objective measure. The aim of this study is to explore an electrophysiological biomarker for the deep meditation stage by studying the autonomic nervous system (ANS) activities during meditation using heart rate variability (HRV). We recruited 70 experienced Tibetan Buddhist monks, and recorded electrocardiogram signal for ~10 min of rest followed by ~30 min of meditation. We found two different stages of meditation, i.e., light meditation (0-10min) and deep meditation (after 10min) stages, which can be distinguished by the increased very low frequency (VLF, 0.003-0.04Hz) and low frequency (LF, 0.04-0.15Hz) power. The light meditation stage was comparable with the rest in the ANS activities, while the deep meditation stage was significantly different from the rest and light meditation stage. We speculate that the deep meditation stage could be marked by significant increases of VLF and LF. Meanwhile, ratio of low and high frequency power, standard deviation of RR intervals and fractal scaling exponents of detrended fluctuation analysis also increased in the deep meditation stage. Our results indicated that meditation is a dynamic process and the deep meditation was dominated by activated sympathetic nervous system.

Keywords: Brain Functional Imaging - Connectivity and Network, Human Performance - Attention, Brain Functional Imaging - EEG and Evoked Potentials
Abstract: Voluntarily shifting attention to a location of the visual field improves the perception of stimulus at the attended location and suppresses information processing at unattended locations, something referred to as visual spatial attention. Previous studies have shown that alpha band phase synchronization between frontal and posterior regions was modulated by visual spatial attention. However, the causality of such long-range connectivity is still unknown, which makes it difficult to assess its relation to the underlying neural mechanisms of visual spatial attention. In this study, 64-channel scalp EEG was collected from 26 healthy adults when performing a visual spatial-cueing attention task, and a recently proposed algorithm, phase transfer entropy(PTE), was used to study the directed alpha phase interactions between frontal and posterior regions during attention orienting. We found that: (i) The PTE from frontal regions to posterior regions significantly increased after cue onset; (ii) The PTE from frontal regions (especially the right hemisphere) to posterior regions showed significant lateralization during late cue-target interval, with the greater PTE of contralateral pair than that of ipsilateral pair relative to cue direction. These results suggest that alpha phase interactions might reveal top-down control signals from frontal to posterior regions during visual spatial attention, and support the right frontal regions as the causal role in top-down attentional control.

Keywords: Brain-computer/machine Interface, Motor neuroprostheses, Motor learning, neural control, and neuromuscular systems
Abstract: Brain-machine interfaces (BMIs) use signals from the brain to control cursors or robotic arms, with potential applications for restoring the ability for users to interact with the physical world around them. BMIs that are kinematically redundant allow for many viable solutions for the same task. While natural motor control involves the coordinated movements of kinematically redundant limbs, it is unclear how the brain might control the redundant degrees of freedom (DOF) in a BMI. In this study, we analyze a previously collected dataset where a macaque controlled a 4 DOF virtual arm in 2D space. A Kalman filter was used to decode neural signals from motor cortices into the four joint angle velocities. The monkey was instructed to move the virtual arm from a center target to eight peripheral targets, distributed evenly around a circle in a self-initiated center-out task. The monkey was able to achieve high accuracy in the task in the first day, but reach times continued to decrease over learning and endpoint trajectories became more stereotyped. We found that the neural activity fired in more correlated patterns over days with increased firing rates, suggesting a consolidation of neural activity into a high-level representation of the joint angles, optimizing endpoint control.

Keywords: Brain-computer/machine Interface, Brain-Computer/Machine Interface - Biofeedback, Brain physiology and modeling - Neural dynamics and computation
Abstract: Learning often involves adapting behavior in response to the inferred causes of success and failure. At the neural level, this can be the result of repeating activity patterns of neurons that lead to favorable outcomes. However, it is not clear how the contributions of individual cells to an ongoing behavior is assessed. Using a calcium imaging based closed loop Brain-Machine Interface (CaBMI), we trained mice to perform a neuroprosthetic task using the coordinated activity of a small ensemble of neurons in layer 2/3 of sensorimotor cortex. We find that that after an initial period of exploration, neurons that do not directly drive the effector decrease in variance and event frequency over the course of learning. However, a large fraction of these ‘indirect’ cells demonstrate robust spatiotemporal dynamics both before and after an animal achieves reward. Throughout a single 30 minute session, these spatiotemporal sequences increase in frequency and become more consistent. Our findings suggest that neuroprosthetic control is the result of an emergent, spatially organized network level solution, rather than the direct modulation of a few chosen output neurons.

Keywords: Brain-computer/machine Interface, Brain physiology and modeling - Neural dynamics and computation, Brain Physiology and Modeling - Neural circuits
Abstract: Prediction of future outcome plays a key role during neuroprosthetic control. It informs about current performance, so the subject can correct and refine its actions. During performance an error signal emerges representing the difference between what was predicted and what actually happened. Several studies have focused on understanding how error signals emerge; however, little is known about the predictions needed to compute them. It has been suggested that striatum, in particular ventral striatum, may be the responsible for encoding a prediction function. So far, there is no clear evidence that these predictions are present in the ventral striatum or the different roles that dorsal and ventral striatum play during neuroprosthetic control. To that effect, we trained rats in a brain-machine-interface (BMI) task while recording the activity in striatum. We showed that neuronal activity of ventral striatum could predict future outcome and that this prediction was aligned with the state prediction of a temporal difference model of reinforcement learning. These results highlight the relevance of ventral striatum during neuroprosthetic control and suggest the use of neuronal prediction information to improve brain-machine-interfaces

Keywords: Neural signal processing, Brain physiology and modeling - Neuron modeling and simulation, Neural Interfaces - Neuroimaging
Abstract: Interpreting the neural code involves decoding the firing pattern of sensory neurons from the perspective of a downstream population. Performing such a read-out is an essential step for the understanding of sensory information processing in the brain and has implications for Brain-Machine Interfaces. While previous work has focused on classification algorithms to categorize stimuli using a predefined set of labels, less attention has been given to full-stimulus reconstruction, especially from calcium imaging recordings. Here, we attempt a pixel-by-pixel reconstruction of complex natural stimuli from two-photon calcium imaging of 103 neurons in layer 2/3 of mouse primary visual cortex. Using an optimal linear estimator, we investigated which factors drive the reconstruction performance at the pixel level. We find the density of receptive fields to be the most influential feature. Finally, we use the receptive field data and simulations from a linear-nonlinear Poisson model to extrapolate decoding accuracy as a function of network size. Based on our analysis on a public dataset, reconstruction performance using two-photon protocols might be considerably improved if the receptive fields are sampled more uniformly in the full visual field. These results provide practical experimental guidelines to boost the accuracy of full-stimulus reconstruction.

Keywords: Neural Interfaces - Sensors and body Interfaces, Neural Signal Processing - Time frequency analysis
Abstract: Nowadays, Virtual Reality (VR) is bringing great benefits to Anxiety Disorder treatments, as well as to other brain cognitive dysfunctions. The advantage of VR is that it can provoke stimuli to the same degree as real-life situations. However, measurement methods of physiological changes caused by the aforementioned stimuli, which apply to VR Anxiety Disorder treatments, have not been examined extensively. As a result, clinicians who use biosignal sensors tend to ask their patients to remain motionless during simulations in order to achieve accurate measurements from the sensors. It is clear that this practice limits the level and range of benefits yielded when using VR simulation. As a consequence, the patients’ experience is restricted and so is the potential of the sensors’ application in the treatment methods. Furthermore, the data gathered from the sensors is handled using conventional analysis affecting the conclusions drawn about the patients’ state. This study aims to emphasise the importance of interacting with the stimuli during the VR treatment through the proposal of an Electrodermal Activity (EDA) Sensor System architecture that can be combined with VR simulations while still allowing the patient to move and interact within the Virtual Environment, without compromising the sensor’s measurements. Continuous Deconvolution Analysis is used to draw conclusions from the gathered biosensor data.

Keywords: Brain Functional Imaging - EEG and Evoked Potentials, Brain Functional Imaging - Classification, spatiotemporal dynamics, Brain-computer/machine Interface
Abstract: The problem of detecting the presence of Social Anxiety Disorder (SAD) using Electroencephalography (EEG) for classification has seen limited study and is addressed with a new approach that seeks to exploit the knowledge of EEG sensor spatial configuration. Two classification models, one which ignores the configuration (model 1) and one that exploits it with different interpolation methods (model 2), are studied. Performance of these two models is examined for analyzing 34 EEG data channels each consisting of five frequency bands and further decomposed with a filter bank. The data are collected from 64 subjects consisting of healthy controls and patients with SAD. Validity of our hypothesis that model 2 will significantly outperform model 1 is borne out in the results, with accuracy 6--7% higher for model 2 for each machine learning algorithm we investigated. Convolutional Neural Networks (CNN) were found to provide much better performance than SVM and kNNs.

Keywords: Brain physiology and modeling - Neuron modeling and simulation, Brain physiology and modeling - Neural dynamics and computation
Abstract: Neurostimulation demonstrates success as a medical treatment for patients suffering from neurodegenerative diseases and psychiatric disorders. Despite promising clinical results, the cellular-level processes by which they achieve these favorable outcomes are not completely understood. Specifically, the neuronal mechanisms by which neurostimulation impacts ion channel gating and transmembrane ionic flux are unknown. To help elucidate these mechanisms, we have developed a novel mathematical model that integrates the Poisson-Nernst-Planck system of PDEs and Hodgkin-Huxley based ODEs to model the effects of this neurotherapy on transmembrane voltage, ion channel gating, and ionic mobility. Using a biologically-inspired domain, in silico simulations are used to assess the impact of TES and DBS on neuronal electrodynamics. Results show that an instantaneous polarization of the membrane's resting potential occurs in a location specific manner, where the type and degree of polarization depends on the position on the membrane. This polarization in turn leads ion channel gating and transmembrane ionic flux to change in a site specific fashion. In addition, results show differences in polarization, membrane voltage, and transmembrane ion mobility resulting from highly distinct forms of neurostimulation, namely transcranial electrical stimulation and deep brain stimulation.

Keywords: Brain-computer/machine Interface, Neural signal processing, Neural Interfaces - Computational modeling and simulation
Abstract: This paper presents a new preprocessing method to correct blinking artifacts in Electroencephalography (EEG) based Brain-Computer Interfaces (BCIs). This Algorithm for Blink Correction (ABC) directly corrects the signal in the time domain without the need for additional Electrooculogram (EOG) electrodes. The main idea is to automatically adapt to the blink’s inter- and intra-subject variability by considering the blink’s amplitude as a parameter. A simple Minimum Distance to Riemannian Mean (MDRM) is applied as the classification algorithm. Preliminary results on three subjects show a mean classification accuracy increase of 13.7% using ABC.

Keywords: Brain-computer/machine Interface
Abstract: Motion-onset visual evoked potential (mVEP) has been successfully used for spelling in both EEG and intracranial EEG based brain-computer interface (BCI). However, its speed is relatively slow compared to P300 and SSVEP paradigms. In order to improve the speed, we proposed a novel bi-directional (leftward and rightward) visual motion BCI paradigm, which decreased the stimulus presentation time of single trial by 50%. Offline experiments were conducted on 5 subjects, which revealed a unique symmetrical spatial and temporal mVEP pattern. The N200 peak of mVEP first appeared on the hemisphere contralateral to the visual motion onset position of stimuli, which reflected the hemisphere transmission delay caused by the optic chiasm. Based on this observation, we developed a BCI system capable of discriminating not only target mVEP responses from non-target ones, but also response to leftward motion from rightward ones. Our new system achieved an averaged AUC 0.93±0.044 in BCI classification, and an information transfer rate (ITR) boost of 43±16 over original uni-directional mVEP BCI in offline evaluation. The results demonstrated the capacity of bi-directional visual motion paradigm in doubling the BCI spelling speed, and laid the foundation for a high speed online visual motion BCI.

Keywords: Neurorehabilitation - Wearable systems, Neuromuscular Systems - Wearable systems, Neuromuscular Systems - Neurorehabilitation
Abstract: .In this study we demonstrate the development of a muscle performance metric to facilitate neurorehabilitation. The muscle performance metric measures the change in power of muscle activity over time as repetitive exercise is performed. It will be demonstrated that a time-constant can be constructed to measure progress of muscle activity based on the exercise. We will present herein that this metric is consistent with an exercise and can be used to predict outcomes. We propose that this muscle performance metric can be used to optimize therapeutic interventions and to design user-specific performance targets based on exercise strategies.

Keywords: Brain functional imaging, Brain Functional Imaging - Classification, spatiotemporal dynamics
Abstract: In this paper, we propose a novel regularized mixture model for clustering matrix-valued image data. The new framework introduces a sparsity structure (e.g., low rank, spatial sparsity) and separable covariance structure motivated by scientific interpretability. We formulate the problem as a finite mixture model of matrix-normal distributions with regularization terms, and then develop an Expectation-Maximization-type of algorithm for efficient computation. Simulation results and analysis on brain signals show the excellent performance of the proposed method in terms of a better prediction accuracy than the competitors and the scientific interpretability of the solution.

Keywords: Brain-computer/machine Interface
Abstract: Brain-machine interfaces (BMIs) use signals acquired from the brain to control actuators such as computer cursors or robots, with potential to restore motor function to individuals with disabilities. While the process of learning and controlling a BMI is complex, involving cortico-striatal networks, it has been well-established that the brain is able to learn to control BMI actuators using relatively few neurons as direct inputs into the decoder. In particular, neurons that are used as inputs to a BMI decoder experience changes in direction tuning and modulation depth, eventually forming a stable neuroprosthetic map. Furthermore, previous work has shown that indirect neurons (those that are not inputs to the decoder) also form a stable neuroprosthetic map that differs from manual reaching. However, it is still unclear how these changes in indirect units are formed over the course of learning. We found that over learning, indirect neurons adapted similarly to that of decoded (direct) neurons. Indirect neurons formed a stabilized tuning map, decreased neural dimensionality, and consolidated firing activity into more correlated patterns. Moreover, direct and indirect neurons adapted together, and neurons with low modulation depth (MD) decreased firing activity while those high MD increased firing activity over learning. Together, our results show that indirect neurons change alongside direct neurons, suggesting a large-scale neural search and adaptation for direct neurons.

Keywords: Human Performance - Modelling and prediction, Brain Functional Imaging - Classification, spatiotemporal dynamics
Abstract: This paper explores the discrimination ability and stability of electroencephalogram (EEG) over time and eye movement signals for classifying five emotions: happy, sad, fear, disgust and neutral. We develop a multimodal emotion dataset called SEED-V with 16 subjects. Two classifiers are trained based on the EEG and eye movement signals, and topographic maps are used to depict the neural patterns of EEG signal. The classification result based on EEG, eye movement and feature level fusion (FLF) reaches the average accuracies of 70.8%, 59.87% and 75.13%, respectively. The experiment result indicates that: a) the EEG and eye movement signals has good discrimination ability for five emotion classification problem; b) the beta and gamma bands of EEG signal have better discrimination ability than the delta, theta and alpha bands; c) the stable neural patterns of different emotions do exist and are common across sessions and d) the neural pattern of disgust emotion has high gamma response in the frontal area, while fear emotion has low activation at the top of brain in the gamma band.

Keywords: Human Performance - Modelling and prediction, Brain Functional Imaging - Classification, spatiotemporal dynamics
Abstract: Recently, various multimodal approaches to enhancing the performance of affective models have been developed. In this paper, we investigate the complementary representation properties of EEG and eye movement signals on classification for five human emotions: happy, sad, fear, disgust, and neutral. We compare the performance of single modality and two different modality fusion approaches. The results indicate that EEG is superior to eye movements in classifying happy, sad and disgust emotions, whereas eye movements outperform EEG in recognizing fear and neutral emotions. Compared with eye movements, EEG has the advantage of classifying the five emotions, with the mean accuracies of 69.50% and 59.81%, respectively. Due to the complementary representation properties, the modality fusion with bimodal deep auto-encoder significantly improves the classification accuracy to 79.71%. Furthermore, we study the neural patterns of five emotion states and the recognition performance of different eye movement features. The results reveal that five emotions have distinguishable neural patterns and pupil diameter has a relatively high discrimination ability than the other eye movement features.

Keywords: Brain Functional Imaging - Classification, spatiotemporal dynamics, Brain Functional Imaging - Connectivity and Network, Brain Functional Imaging - fMRI
Abstract: Visualization of how the external stimuli are processed dynamically in the brain would help understanding the neural mechanisms of functional segregation and integration. The present study proposed a novel temporal autoencoder to estimate the neurodynamics of functional networks involved in rhythm encoding and reproduction. A fully-connected two-layer autoencoder was proposed to estimate the temporal dynamics in functional magnetic resonance image recordings. By minimizing the reconstruction error between the predicted next time sample and the corresponding ground truth next time sample, the system was trained to extract spatial patterns of functional network dynamics without any supervision effort. The results showed that the proposed model was able to extract the spatial patterns of task-related functional dynamics as well as the interactions between them. Our findings suggest that artificial neural networks would provide a useful tool to resolve temporal dynamics of neural processing in the human brain.

Keywords: Human Performance - Fatigue, Neuromuscular Systems - Computational modeling and simulation
Abstract: Listening related fatigue is a common difficulty, especially for people with hearing loss. To reduce this negative impacts on their daily life, we have to understand the origins and the development of this fatigue. For this reason we developed a quantitative model of cognitive fatigue. We consider the development and course of fatigue, using a Simulink model including factors like internal and external demand, the discrepancy between expected and actual performance and the relation to reward. Besides, control and base motivation act as influencing factors on actual motivation. The functions changing according to the design of the fixed parameters are demanded effort, exerted effort, distress and fatigue. Fatigue acts as an additional input on motivation. Our focus lies mostly on listening-related fatigue. This type of fatigue occurs in difficult listening situations, especially in individuals with hearing loss. This model can be applied of a wide range of listening situations and provides a good basis for further research in this field.

Keywords: Brain Functional Imaging - Connectivity and Network, Brain Functional Imaging - EEG and Evoked Potentials, Brain functional imaging
Abstract: In this study, we investigate the role of cross-frequency coupling (CFC) between the low frequency frontal (F) oscillations and high frequency parieto-occipital (PO) oscillations in creative ideation. These long distance brain interactions have been associated with cognitive processes like working memory (WM). We analyzed the alpha-gamma, theta-alpha and theta-gamma coupling for three different phases (early, mid and late) of creative ideation. We used Alternate Uses (AU) task as experimental paradigm for generating creative thinking and Object Characteristics (OC) task as the control. The results indicate significantly higher F to PO connection for AU compared to OC across all the three tested CFCs, in the early phase of creative ideation. Further, our results suggest that the late phase of creative ideation is associated with significantly lower theta-alpha and theta-gamma coupling but higher alpha-gamma coupling compared to OC. The results demonstrate that these F to PO cross-frequency interactions, associated with the WM subnetworks, play a major role in creative ideation. Further, our results suggest alpha-gamma interaction in the early phase of creative ideation to be a potential biomarker for individual differences in creativity.

Keywords: Brain-computer/machine Interface, Neural signal processing, Neural Signal Processing - Time frequency analysis
Abstract: Recently, filter bank analysis has been used in several detection methods to extract selective frequency features across multiple brain computer interface (BCI) modalities due to its effectiveness and simple structure. In this work, we propose filter bank technique as a standard preprocessing method for popular training free multi-channel steady-state visual evoked potential (SSVEP) detection methods to overcome subject-specific performance differences and a general improvement in detection accuracy. Our study validates the effectiveness of filter bank extensions by comparing performance differences of multi-channel methods with their filter bank counterparts using a forty target SSVEP benchmark data set collected across thirty five subjects. The results demonstrate that the proposed two stage (a filter bank stage followed by SSVEP detection) implementation of popular multi-channel algorithms provide significant improvement in performance at short data lengths of < 2.75 s (p < 0.001) and can be viewed as a potential standard detection approach across all SSVEP identification problems.

Keywords: Brain Functional Imaging - Connectivity and Network, Human Performance - Attention, Brain Functional Imaging - EEG and Evoked Potentials
Abstract: Understanding modular organization between brain regions can provide deeper insight into the complex neural mechanisms associated with processes like vigilance decrement. Distinct but interacting modules in brain connectivity networks have been known to support integration of specific mechanisms relevant in high-order cognitive processes. To investigate the neuronal mechanisms associated with vigilance decrement, we conducted an experiment where the participants performed a driving task. EEG graph metrics within communities, like clustering coefficient (C_intra), efficiency (E_intra), density (D_intra), and between communities, like intermodule density (D_inter), were computed from the source-localized surface brain signals. Further, we also calculated the nodal out degree to investigate the difference in information flow in the brain during vigilance decrement. Increase in the intermodule density, D_inter, was observed from the left fronto-parietal cluster to the right temporo-parietal cluster. Moreover, significant reduction in the intramodule metrics, E_intra and Cintra was observed in the right temporo- parietal cluster. Thus, our findings signify a flexible topographical architecture to compensate the hub disruption effect caused due to decline in vigilance.

Keywords: Brain Functional Imaging - Connectivity and Network, Brain Functional Imaging - EEG and Evoked Potentials, Brain functional imaging
Abstract: Perception of olfactory stimuli involves complex brain processing which can be directly associated with cognition and emotion. Neural structures embedded deep within the brain and several cortical entities collaborate for the processing of this unique sensory modality. In this study, we investigate the dynamic changes in the neural responses associated with prolonged and repeated exposure to pleasant odor stimuli. Graph metrics computed from EEG functional connectivity like clustering coefficient (p = 0.0008), characteristic path length (p = 0.014) and local efficiency (p = 0.0005) were seen to undergo statistically significant changes, indicating inhibition in the global and local information processing that can be attributed to olfactory adaptation. Moreover, dominant but diminishing activity was observed in the left cerebral hemisphere, signifying recruitment of various neuronal ensembles associated with complex cognitive processes for perception of pleasant odor stimuli.

Keywords: Brain-computer/machine Interface, Neural Signal Processing - Nonlinear analysis, Brain physiology and modeling
Abstract: We propose a novel non--linear source separation technique for single--channel, multi--trial Electroencephalogram (EEG). First, a generative model is posited as the generating process behind bandpassed traces. In particular, the inputs are conceived as the state variable of a switching mechanism that samples temporal snippets from two distributions corresponding to a background component and a phasic event or wave packet counterpart. In order to non--linearly separate the sources, we propose a neurophysiologically principled, non--linear mapping to a space of ell_2--norms via the Embedding Transform. In this way, the estimated phasic event component---an ideal time series where neuromodulations are emphasized---is isolated for further processing. The algorithm is tested on the Brain--Computer Interface (BCI) Competition 4 dataset 2a. The results not only surpass classic power--based measures, but also highlight the discriminative nature of scale--specific wave packets in motor imagery tasks. The inherent switching mechanism that generates the traces suggests a transient, temporally sparse feature of the neuromodulations that can be further exploited in applications where compression is advantageous.

Keywords: Neural Interfaces - Microelectrode and fabrication technologies, Neural Interfaces - Implantable systems, Neural Interfaces - Regeneration and tissue-electrode Interface
Abstract: Myelin sheath of the axons provides an electrical insulation required for fast propagation of the nerve signal. The same insulation reduces signal-to-noise ratio of electrical readings and increases potential requirements for nerve electrical stimulation. Water/lipid insoluble molecular tellurium slowly metabolizes within the animal body affecting enzyme squalene monooxygenase, which requires for myelin sheath maintenance by the Schwann cells. Hereby, we describe the coating of the inert gold electrodes by molecular tellurium via electro deposition from ionic liquid solutions. Resulting coated electrodes decreases viability of Schwann cells, but well tolerated by neuronal cells in culture and in vivo. When implanted around the peripheral nerves in rats the coated electrodes decrease the basic myelin protein content and locally disrupt myelin sheaths after implantation.

Keywords: Neural Interfaces - Microelectrode and fabrication technologies, Neural Interfaces - Implantable systems, Neural Interfaces - Biomaterials
Abstract: The applications of polyimide (PI) based thin film electrodes in neural implants has been steadily increasing in the last decades. Beside the advantages of factors such as size reduction and chemical inertness, the adhesion between PI and metals lacks the necessary robustness for chronic implantation. Adhesion promoters like silicon carbide (SiC) and titanium help increase the long-term stability. However crack formation is still probable in thin-film metallization of active contact sites, which can lead to detrimental thin-film metal delamination. This work presents our process to integrate 3D groove structures into the active contact sites, thereby reducing the probability of crack formation. The contact sites were entirely covered with a sputtered iridium oxide film (SIROF) and showed more conformable electrochemical properties compared to regular planar SIROF contact sites.

Keywords: Brain physiology and modeling - Neural dynamics and computation, Neural Interfaces - Recording, Neural Interfaces - Neural stimulation
Abstract: Active high-density electrode arrays can record the spiking activity of a large number of single-neurons in brain circuits. This offers the opportunity to develop closed-loop neural interfaces for studying complex brain circuits contingent on their cellular activity. However, this requires adapted solutions to process large-volumes of data acquired by thousands of electrodes and, in turn, generate feedbacks. Here, we present of a closed-loop system that exploits System on Chip resources of a Xilinx ZedBoard Zynq-7000 to processes the instantaneous mice retina activity recorded by 4096 closely spaced microelectrodes. This is used to infer in real-time the functional properties of retinal ganglion cells (RGCs) all over the retina. Results show the performances of two interactive algorithms designed for (i) data reduction by clustering single-unit sub-millisecond correlated spike-trains from closely spaced neighboring electrodes (closed-loop latency of 22.9±1.5 ms per second of recording, n=5 retina); (ii) the classification of major types of RGCs consisting in ON and OFF types of functional responses (closed-loop latency of 117.3±30.9 ms, n=5 retina).

Keywords: Neural Interfaces - Microelectrode and fabrication technologies, Neural Interfaces - Implantable systems, Neural Interfaces - Neural microsystems and Interface engineering
Abstract: This paper describes the fabrication of silicon neural probes based on low-cost semiconductor technologies, such as photolithography, thin-films deposition and blade dicing. Each 8 mm long fabricated neural probe has 13 platinum recording sites (microelectrodes) with different dimensions, ranging from 50 × 50 µm2 to 300 × 300 µm2. The platinum microelectrodes impedance were characterized through electrochemical impedance spectroscopy with results ranging from 5 kΩ to 600 kΩ at 1 kHz. Local field potentials were recorded when performing in-vivo electrophysiological recordings on an adult mouse. The results demonstrate the functionality of the neural probe, by acquiring local field potentials activity even with a microelectrode area of 50 × 50 µm2. Moreover, the results suggest a robust and versatile fabrication process avoiding etching processes.

Keywords: Neural Interfaces - Recording, Neural Interfaces - Computational modeling and simulation, Neural Interfaces - Neural stimulation
Abstract: Neural recordings can provide useful information regarding specific sensory or motor function that can be applied to functional electrical stimulation systems for persons with spinal cord injury or stroke patients. Among myriad neural interfaces, the nerve cuff electrode offers an attractive tool for communicating with peripheral nervous system. The tripolar cuff electrode is the most common nerve cuff design; however, the physical symmetry of this configuration does not allow for the differentiation of action potentials propagating in opposite directions (i.e., afferent vs. efferent activity). The main goal of this study was to test the feasibility of using a tetrapolar configuration to achieve high SNR, directionally-sensitive recording of nerve activity. The objectives for this design were to combine (1) low-noise recordings from two sets of tripolar contacts with (2) the directional information provided by differentially measuring the two tripolar signals.

Keywords: Neural Interfaces - Implantable systems, Neural Interfaces - Neural microsystems and Interface engineering, Neural Interfaces - Microelectrode and fabrication technologies
Abstract: Brain machine interfaces notoriously face difficulties in achieving long term implanted recording stability. It has been shown that damage and inflammation, caused during insertion by electrodes that are too large and stiff, provoke a sustained inflammatory tissue response. This is commonly referred to as the foreign body response, resulting in encapsulation and thus increased electrode impedance over time. Accordingly, neural interfaces with ever smaller and more flexible electrodes are continually in development, but unfortunately face challenges of their own, first and foremost of which is buckling and bending during insertion. This work presents the development of a prototype insertion method, comprising an insertion device and novel probe architecture, that promotes straight insertion without buckling, while simultaneously minimizing the insertion force for multi-microwire electrode probes. When compared against insertion of probes with unsupported free electrodes, the prototype method achieved significantly straighter electrode insertion, resulting in both a smaller distance between electrode recording tips and a greater average insertion depth. While achieving less straight insertion than probes with sucrose coated electrodes, a common technique for promoting reliable insertion without buckling, the prototype method was able to maintain significantly lower insertion forces.

Keywords: Neural Interfaces - Implantable systems, Brain-computer/machine Interface, Neural Interfaces - Neuromorphic engineering
Abstract: This paper presents for the first time a real-time closed loop neuromorphic decoder chip-driven intra-cortical brain machine interface (iBMI) in a non-human primate (NHP) based experimental setup. Decoded results show trial success rates and mean times to target comparable to those obtained by hand-controlled joystick. Neural control trial success rates of approximately 96% of those obtained by hand-controlled joystick have been demonstrated. Also, neural control has shown mean target reach speeds of approximately 85% of those obtained by hand-controlled joystick. These results pave the way for fast and accurate, fully implantable neuromorphic neural decoders in iBMIs.

Keywords: Neural Interfaces - Neural stimulation, Neural Interfaces - Implantable systems
Abstract: Neurostimulation is used to treat a variety of neurological diseases. Historically, these implantable neurostimulator systems have relied on providing therapeutic benefit to patients using tonic, or open-loop, stimulation. In this paper, we review the state of the art in current spinal cord stimulation therapy to identify the needs of the field for a new tool to enable new research. We then present an overview of the design and capabilities of the Nexus-I Research Development Kit for investigational human-use research activities for patients who have been implanted with the Model 97715 Intellis neurostimulator.

Keywords: Neural signal processing, Neural Interfaces - Sensors and body Interfaces
Abstract: Electroencephalogram (EEG) systems commonly used in laboratory environments are rapidly moving towards real-world applications. In this paradigm shift, new methods for recording and classifying EEG signals are necessary. However, it is challenging to validate the efficacy of new electrodes or other data acquisition components when the target signal cannot be controlled. Here, we propose and validate a method for using an attenuated tACS unit to generate a ground-truth signal in conjunction with porcine skin to determine electrode effectiveness. We highlight the utility of this approach by objectively comparing the signal quality of a chirp signal though porcine skin as measured by four different EEG electrodes, three “dry” and a standard hydrogel. We believe a similar approach could be used with transdermal signals on live humans to observe effects of motion, electrode type, or environment on EEG in order to improve design for use in noisy, non-laboratory settings.

Keywords: Neural Interfaces - Neural microsystems and Interface engineering, Brain-Computer/Machine Interface - Robotics applications, Motor learning, neural control, and neuromuscular systems
Abstract: This work explores the design of a low cost microcircuit platform that is light weight, energy efficient and highly capable in stimulating insect brain, logging sensory data and using it to model and control its neuromuscular pathway, all in real-time. The results are encouraging and opens up avenues for miniature bio-robots.

Keywords: Neural Interfaces - Sensors and body Interfaces, Neural Interfaces - Implantable systems, Neurorehabilitation - Wearable systems
Abstract: Chronic active implanted medical devices rely on a robust transcutaneous energy supply and signal transmission. A robust and precise alignment of the external communication unit to the subcutaneous counterpart is crucial in multichannel transceiver systems. In this work, we present a method for an active alignment monitoring that can be easily integrated. The system consists of three insulated electrodes in a triangular configuration with electrical connecting tracks. Transcutaneous signal transmission to the middle electrode leads to output signals at the other electrodes, which can be measured extracorporeally. Ultrapure water served as first in vitro model for human skin. This simple model was validated with means of electrochemical impedance spectroscopy with human skin samples for an ex situ alignment characterization of the proposed system. The results suggest that the proposed approach can reliably achieve alignment in the sub-mm region.

Keywords: Neuromuscular Systems - Wearable systems, Neuromuscular Systems - EMG models, processing and applications
Abstract: Surface electromyography (sEMG) allows to evaluate neurophysiological aspects of movements. The traditional muscle-specific sEMG approach requires considerable expertise to correctly position bipolar derivations over each muscle of interest. This reduces the availability of myoelectric imaging tools for non-professional users in everyday life conditions. High-density EMG (HDsEMG) consists of recording the sEMG via a dense array of electrodes that restricts the imaging surface to a narrow area, thus limiting usability. Here we demonstrate the usability and advantages of a medium-density sEMG (MDsEMG). We recorded the muscle activity of a subject performing repeated arm flexion and extension using both muscle-specific sEMG (4 derivations over the biceps and triceps) and MDsEMG (28 monopolar derivations, positioned around the arm circumference). Grand-average dynamic-time-warped amplitudes showed similar activation patterns for both modalities. However, MDsEMG allowed to extract maps reflecting the spatial activation of muscles at different movement phases (minimum and maximum acceleration). MDsEMG also proved superior in decoding the elbow angle and allowed to map the regions of maximum significance. MDsEMG is a viable alternative to muscle-specific and HD sEMG. It does not require expertise in electrode positioning and is able to cover large surfaces. It thus can pave the way to easier myoelectric imaging in everyday life and more effective rehabilitation treatments.

Keywords: Neurorehabilitation - Neurofeedback, Brain-Computer/Machine Interface - Biofeedback, Neural Interfaces - Recording
Abstract: Chronic neuropathic pain is a debilitating disorder. Many patients receive pain relief when implanted with a Spinal Cord Stimulation (SCS) system. However, there is currently no way to automatically determine the most effective parameters to use in stimulation, so some patients must return to the clinic periodically to have their device re-programmed. Further, postural changes can shift the position of the stimulating electrodes and require programming changes to optimize pain relief. Evoked compound action potentials (ECAPs) could be used as feedback signals to determine optimal patient stimulation parameters. To determine the feasibility of closed-loop feedback in SCS, we have designed a recording system that simultaneously records evoked compound action potentials (ECAPs) during stimulation. Using differential recording and high dynamic range amplifiers, this system is capable of simultaneous stimulation and recording without saturating. Here we show in vitro the capability of this system to record simulated evoked compound action potentials across all channels during stimulation without saturation and with low noise. This data acquisition system provides the ability to add recording to existing clinical stimulation trial systems with minimal changes to standard clinical practice.

Keywords: Neural Interfaces - Recording, Neural Interfaces - Neural microsystems and Interface engineering, Brain-computer/machine Interface
Abstract: We present a design and analysis of nonlinear distortions for low-area integrated neural preamplifier with pseudoresistor-based AC coupling. We evaluate the distortions as a function of frequency, signal amplitude and sizing of the pseudoresistors. We describe a preamplifier design in 0.18 µm SOI CMOS technology with Total Harmonic Distortions (THD) below 1% in the full range of frequencies and amplitudes of extracellular neural signals. The circuit noise is 5.82 µV rms in the Local Field Potential (LFP) frequency range (1-300 Hz) and 4.45 µV rms in the action potential (AP) range (300 Hz - 5 kHz). The preamplifier occupies silicon area of 0.0046 mm² and is suitable for recording systems with >10,000 channels per cm² of the chip area, providing high-fidelity amplification for large-scale neural interfaces.

Keywords: Neuromuscular Systems - Wearable systems, Neuromuscular Systems - EMG models, processing and applications
Abstract: Alternatives to conventional wet Silver/Silver Chloride (Ag/AgCl) hydrogel electrodes could greatly improve the ability to record long-term surface electromyography (sEMG). This paper presents preliminary evaluation of dry epidermal electronic system (EES) electrodes over 8 days measurements. We tested 3 subjects and recorded sEMG signals of right rectus femoris during static and dynamic tasks from epidermal electrodes and traditional Ag/AgCl electrodes. For epidermal electrodes that remained intact and maintained adhesion to the skin, signal-to-noise ratio (SNR) values were greater than 16.9 dB and signal-to-motion ratio (SMR) values were greater than 13.1 dB, which were both above the accepted thresholds for good signal quality (SNR > 15 dB and SMR > 12dB). We also developed a new metric, the Signal Quality Index, SQI to help determine whether higher SNR or higher SMR were necessarily indicative of better EMG signal quality. Our results indicated that the single use dry epidermal electrodes were able to maintain high signal quality over 8 days and suggested that dry epidermal electrodes have the potential ability for long-term sEMG recording.

Keywords: Sensory Neuroprostheses - Somatosensory and vestibular, Neural Interfaces - Neural microsystems and Interface engineering, Neural Interfaces - Microelectrode and fabrication technologies
Abstract: This paper reports direct stimulation of bladder pelvic nerve using a battery-free neural clip interface. We implanted a flexible neural clip interface combined with a triboelectric nanogenerator (TENG) on a bladder pelvic nerve in rats to directly modulate bladder functions for the first time. The stimulation parameters such as currents, pulse width, and charges generated by the TENG were characterized. In an in vivo experiment, different numbers of pulses were applied to stimulate the pelvic nerve while monitoring changes in bladder pressure and the occurrence of micturition. The bladder contraction with micturition was observed for all experiments, except when applying a single pulse for the stimulation. The results demonstrate that this technology may potentially be used for direct modulation of bladder function.

Keywords: Neural signal processing, Neural Interfaces - Microelectrode and fabrication technologies, Neural Interfaces - Neuroimaging
Abstract: The ECoG has been widely used in human brain research, while 2-photon microscopy has been broadly applied to basic neuroscience studies using animal models. Bridging the gap between the 2-photon microscopy and the ECoG is critical for transferring the vast amount of neuroscience knowledge obtained from animal models to human brain studies. Here we develop an LSTM recurrent neural network model to decode the ECoG high gamma power from the cellular calcium activities obtained by multimodal ECoG recordings and 2-photon calcium imaging enabled by transparent graphene microelectrode arrays. In both awake and anesthetized states, our model can successfully decode the stimulus-induced ECoG high gamma power increases and its spontaneous fluctuations in the absence of stimulus.

Keywords: Sensory Neuroprostheses - Visual, Neural Interfaces - Neural stimulation, Brain-computer/machine Interface
Abstract: Retinal prostheses aim to restore visual perception in patients blinded by photoreceptor degeneration, by stimulating surviving retinal ganglion cells (RGCs), causing them to send artificial visual signals to the brain. Present-day devices produce limited vision, in part due to indiscriminate and simultaneous activation of many RGCs of different types that normally signal asynchronously. To improve artificial vision, we propose a closed-loop, cellular-resolution device that automatically identifies the types and properties of nearby RGCs, calibrates its stimulation to produce a dictionary of achievable RGC activity patterns, and then uses this dictionary to optimize stimulation patterns based on the incoming visual image. To test this concept, we use a high-density multi-electrode array as a lab prototype, and deliver a rapid sequence of electrical stimuli from the dictionary, progressively assembling a visual image within the visual integration time of the brain. Greedily minimizing the error between the visual stimulus and a linear reconstruction (as a surrogate for perception) yields a real-time algorithm with an efficiency of 96% relative to optimum. This framework also provides insights for developing efficient hardware. For example, using only the most effective 50% of electrodes minimally affects performance, suggesting that an adaptive device configured using measured properties of the patient’s retina may permit efficiency with accuracy.

Keywords: Neural Interfaces - Neural microsystems and Interface engineering, Neural Interfaces - Implantable systems, Brain-computer/machine Interface
Abstract: We present a platform technology encompassing a family of innovations that together aim to tackle key challenges with existing implantable brain machine interfaces. The ENGINI (Empowering Next Generation Implantable Neural Interfaces) platform utilizes a 3-tier network (external processor, cranial transponder, subdural probes) to inductively couple power to, and communicate data from, a distributed array of freely-floating mm-scale probes. Novel features integrated into each probe include: (1) an array of niobium microwires for observing low frequency local field potentials (lf-LFPs) along the cortical column; (2) ultra-low power instrumentation for signal acquisition and data reduction; (3) autonomous, self-calibrating wireless transceiver for receiving power and transmitting data; and (4) a hermetically-sealed micropackage suitable for chronic use. We are additionally engineering a surgical tool, to facilitate manual and robot-assisted insertion, within a streamlined neurosurgical workflow. Ongoing work is focused on system integration and preclinical testing.

Keywords: Neural Interfaces - Neural stimulation, Neural Interfaces - Neural microsystems and Interface engineering, Brain-computer/machine Interface
Abstract: Fully-implantable, bi-directional brain-computer interfaces (BCIs) necessitate simultaneous cortical recording and stimulation. This is challenging since electrostimulation of cortical tissue typically causes strong artifacts that may saturate ultra-low power (ULP) analog front-ends of fully-implantable BCIs. To address this problem, we propose an efficient hardware-based method for artifact suppression that employs an auxiliary stimulator with polarity opposite to that of the primary stimulator. The feasibility of this method was explored first in simulations, and then experimentally with brain phantom tissue and electrocorticogram (ECoG) electrode grids. We find that the canceling stimulator can reduce stimulation artifacts below the saturation limit of a typical ULP front-end, while delivering only ~10% of the primary stimulator’s voltage.

Keywords: Sensory Neuroprostheses - Somatosensory and vestibular, Sensory Neuroprostheses, Brain-computer/machine Interface
Abstract: Intracortical microstimulation (ICMS) of somatosensory cortex can be used to provide sensory feedback from a brain-controlled robotic arm, leading to improved manipulation of objects. ICMS-evoked percepts have been shown to depend systematically on stimulation parameters, including pulse charge, pulse patterning, and electrode location. However, while the dependence of the sensory percept on ICMS frequency has been established, previous experiments were limited to a narrow range of frequencies and did not conclusively establish the sensory consequences of changes in frequency. Indeed, while animals could discriminate between pulse trains that varied in frequency, they may have relied on intensive cues to make their discrimination judgments, given the effect of frequency on perceived magnitude. In the present study, we investigate frequency discrimination of ICMS over a wide range of frequencies and investigate the degree to which frequency discrimination relies on differences in perceived magnitude.

Keywords: Neural Interfaces - Biomaterials, Neural Interfaces - Implantable systems, Neural Interfaces - Microelectrode and fabrication technologies
Abstract: The long term performance of implanted medical devices is critically affected by their packaging materials. Traditional hermetic enclosures are a hindrance to miniaturisation which might be avoided by polymer encapsulation. However, much uncertainty remains regarding such an approach for chronic implants. It is difficult to extract meaning from lifetime experiments performed on arbitrary test structures and before many highly reliable devices have actually failed. In this study, we describe an accelerated aging experiment on devices that employ test structures fabricated on a silicone encapsulated 0.35 µm CMOS technology. Samples are aged at 47, 67 and 87 °C under a constant 5V DC bias. After over 300 days of aging, of the 36 samples under test, there have only been four failures. Three failures are attributed to the interconnect to the measurement system, with the remaining failure due to delamination of the silicone encapsulant over wirebonds. By employing a Bayesian reliability analysis of the near-zero failure data we demonstrate a 1st failure percentile for the CMOS components of greater than 92 days when aged at 87 °C, calculated at a 95% confidence level, which approximately corresponds to over 8 years at 37 °C.

Keywords: Brain-Computer/Machine Interface - Biofeedback, Neurorehabilitation - Wearable systems
Abstract: Brain-Computer Interface (BCI) combined with assistive robots has been developed as a promising method for stroke rehabilitation. However, most of the current studies are based on complex system setup, expensive and bulky devices. In this work, we designed a wearable Electroencephalography(EEG)-based BCI system for hand function rehabilitation of the stroke. The system consists of a customized EEG cap, a small-sized commercial amplifer and a lightweight hand exoskeleton. In addition, visualized interface was designed for easy use. Six healthy subjects and two stroke patients were recruited to validate the safety and effectiveness of our proposed system. Up to 79.38% averaged online BCI classification accuracy was achieved. This study is a proof of concept, suggesting potential clinical applications in outpatient environments.

Keywords: Neural Interfaces - Neural stimulation, Neural Interfaces - Microelectrode and fabrication technologies, Brain-computer/machine Interface
Abstract: Stimulation of deep cortical structures has been demonstrated to provide symptomatic alleviation to people with depression, Parkinson’s disease and epilepsy. However, electrodes used clinically for therapies requiring deep-brain-stimulation are currently limited to those that penetrate directly through delicate neural tissue following craniotomy or burr-hole surgery. Using an endovascular neural interface, the risks associated with open-brain surgery can be mitigated. Previously, we demonstrated proof-of-concept that a stent-electrode array can access the brain via blood vessels and can record high-fidelity neural information over a chronic duration. However, it was unknown whether focal stimulation could be delivered from within a cortical vessel. This work demonstrates that a stent-mounted electrode array can stimulate various regions of the sheep motor cortex to elicit visually observable and independent movements of the lip, face, jaw, neck and limb. While inter-animal differences were observed with respect to the elicited movement and stimulation threshold required to generate a response, there was reasonable consistency regarding the response and the location of the stimulating electrodes from the branching central sulcal veins. Six animals were observed to have multiple different and independent muscle movements time-locked to electrical stimulation.