Keywords: Motor neuroprostheses - Prostheses, Neuromuscular systems - Peripheral mechanisms, Neural interfaces - Sensors and body interfaces
Abstract: We propose a new framework for extracting information from extrinsic muscles in the forearm that will allow a continuous, natural and intuitive control of a neuroprosthetic devices and robotic hands. This is achieved through a continuous mapping between muscle activity and joint angles rather than prior discretisation of hand gestures. We instructed 6 able-bodied subjects, to perform everyday object manipulation tasks. We recorded the Electromyographic (EMG) and Mechanomyographic (MMG) activities of 5 extrinsic muscles of the hand in their forearm, while simultaneously monitoring 11 joints of hand and fingers using a sensorised glove.

We used these signals to train a Gaussian Process (GP) and a Vector AutoRegressive Moving Average model with Exogenous inputs (VARMAX) to learn the mapping from current muscle activity and current joint state to predict future hand configurations. We investigated the performances of both models across tasks, subjects and different joints for varying time-lags, finding that both models have good generalisation properties and high correlation even for time-lags in the order of hundreds of milliseconds. Our results suggest that regression is a very appealing tool for natural, intuitive and continuous control of robotic devices, with particular focus on prosthetic replacements where high dexterity is required for complex movements.

Keywords: Motor neuroprostheses - Prostheses, Neural interfaces - Sensors and body interfaces, Neuromuscular systems - Wearable systems
Abstract: Intent recognition is essential for effective control of powered assistive devices, such as powered lower limb prostheses, exoskeletons, or wheelchairs. Currently, EMG and mechanical sensors are used for intent recognition of powered lower limb prostheses. We propose the addition of vision for improved intent recognition control, with this work focused on determining the best method for recognizing of ascending stair edges from 2D images. In this work different image processing methods were tested to determine which method produces the best line extraction. The best results were obtained using Canny, Sobel, Prewitt, and Roberts Cross edge detectors for four colorspace components. Finally, a convex/concave line decision system was developed to produce preliminary results about the presence or absence of stairs in a given image.

Keywords: Motor neuroprostheses, Neural signal processing, Neural interfaces - Implantable systems
Abstract: Clinically available myoelectric prostheses are limited by the inability to control multiple degrees of freedom simultaneously. Linear regression-based control and parallel dual-site control (an extension of conventional amplitude-based methods using intramuscular EMG) are two frequently proposed approaches for simultaneous control. Both approaches assume linearity in the EMG features, but differ in whether users are required to independently modulate the EMG amplitudes from residual limb muscles. The objective of this preliminary study was to compare these two methods for the real-time control of a 3 degree-of-freedom (DOF) wrist/hand system. Both systems used intramuscular EMG amplitudes from six forearm muscles, and differed only in how the signals were used to predict intended prosthesis activity. Five able-bodied subjects were recruited to evaluate each control system (ten subjects total). Performance in a virtual Fitts’ law task demonstrated that parallel dual-site control provided improved controllability when acquiring targets that required use of only one DOF, but linear regression control provided improved performance when acquiring targets requiring use of all three DOFs. Subjects using linear regression control were more easily able to activate multiple DOFs simultaneously, but at the expense of unintended movement when trying to isolate individual DOFs.

Keywords: Motor neuroprostheses - Neuromuscular stimulation, Neuromuscular systems - EMG models, processing and applications, Neuromuscular systems - Neurorehabilitation
Abstract: Functional electrical stimulation (FES) is a useful technique for restoring motor functions for spinal cord injured (SCI) patients. Muscle contractions can be artificially driven through delivery of electrical pulses to impaired muscles, and the electrical activity of contracted muscles under stimulus recorded by electromyography (EMG) is called M-wave. The FES-induced muscle activation which is represented by evoked EMG recordings can indicate the muscle state, and accurate control of muscle activation level by FES is the preliminary step for achieving more complicated FES control tasks. This paper proposes a real-time FES system for control of muscle activation by online modulating pulse width of stimulus. The excitation muscle dynamics is modelled by Hammerstain system with stimulus pulse width and eEMG as input and output respectively, and the model predictive control strategy is adopt- ed to compute the pulse width command sent to the Vivaltis wireless stimulator. Four reference muscle activation patterns are provided to test and validate the real-time closed-loop FES control system. Real-time control results on one able-bodied subject show promising control performances.

Keywords: Motor neuroprostheses - Robotics, Neurological disorders - Stroke, Neuromuscular systems - EMG models, processing and applications
Abstract: Here we investigate whether torque or myoelectric control was more efficient during active, robotic-assisted therapy. Eleven hemiparetic stroke patients employed the RehaArm-robot, which consist of an exoskeleton that supports the entire arm and measures in real time the joint position and moment of the shoulder, in four sessions of one hour on consecutive days. At each session, the patients repeatedly performed basic movements of the shoulder in passive and active mode. During the passive mode, subjects learnt the movements. During the active mode, subjects were asked to complete 40 task repetitions in 20 min for each modality, torque and myoelectric control. The number of repetitions achieved, completion rate (CR), was tracked for each control modality as well as subjects’ opinion about the ease of use of each modality. The results showed that the severe-to-moderate group (Fugl-Meyer Motor Assessment of the Upper Extremity<=40) achieved a significantly higher CR in myoelectric control than in torque control (p<0.05). For the mild group (>40), the CR was very similar for both control modalities. Subjectively, the two groups considered both control modalities similarly easy to use, being the myoelectric control slightly easier (higher median and middle fifty values). These results support the higher efficacy of the myoelectric control for active, robotic-assisted therapy.

Keywords: Motor neuroprostheses - Prostheses, Neural signal processing
Abstract: The reconstruction of finger movement activity from surface electromyography (sEMG) has been proposed for the proportional and simultaneous myoelectric control of multiple degrees-of-freedom. In this paper, we propose a framework for assessing decoding performance on novel movements, that is movements not included in the training dataset. We then use our proposed framework to compare the performance of linear and kernel ridge regression for the reconstruction of finger movement from sEMG and accelerometry. Our findings provide evidence that, although the performance of the non-linear method is superior for movements seen by the decoder during the training phase, the performance of the two algorithms is comparable when generalizing to novel movements.

Keywords: Neurological disorders - Diagnostic and evaluation techniques, Neuromuscular systems - EMG models, processing and applications, Neuromuscular systems - Neurorehabilitation
Abstract: Neurorehabilitation of voluntary motor function after incomplete spinal cord injury (iSCI) is often performed without periodic electromyographic (EMG) parameter analysis such as amplitude, integrated muscle activity and duration of contraction. The main objective of this study was to validate the diagnostic potential of Tibialis Anterior (TA) EMG parameters recorded during a maximum dorsiflexion velocity (MDV) protocol, designed to detect functional recovery in subjects with subacute iSCI. In this study recovery of voluntary motor function was assessed at 2 week intervals during subacute iSCI by measuring TA muscle activity, in addition to muscle strength, voluntary torque generation and gait function improvement. TA motor evoked potentials and maximum velocity of dorsiflexion protocol were also analysed. The study demonstrated that muscle strength, voluntary torque generation and gait function significantly improved during the follow up, in addition to an increase in TA EMG amplitude and a reduction in TA muscle contraction duration. TA EMG amplitude correlated with motor evoked potentials, torque and muscle balance, while short muscle duration correlated with gait function. To conclude, longitudinal assessment of limited recovery of voluntary function during subacute iSCI can be detected with specific TA EMG signature analysis during controlled movement, providing relevant diagnostic information during neurorehabilitation

Keywords: Motor neuroprostheses - Prostheses, Neural signal processing
Abstract: Powered lower limb prostheses can improve amputees’ ability to traverse stairs and ramps by providing positive mechanical work at the knee and ankle joint. EMG signals have been proposed as one way of providing seamless mode transitions by using them in combination with embedded mechanical sensors as inputs to a pattern recognition system that predicts the user’s desired locomotion mode. In this study, we have expanded the amount of mechanical sensor information to include data from an additional five degrees of freedom in the load cell, as well as calculated thigh and shank angles. The purpose of this study was to determine the impact of this additional information on the performance of an EMG-based pattern recognition system designed to predict the desired locomotion mode. Our results indicate that including the additional mechanical sensor signals decreased the error rates of the system for both steady-state and transitional steps when compared to the reduced sensor set. We also found that EMG still decreased the error rate of the system, but to a lesser extent when using the additional mechanical sensors.

Keywords: Motor neuroprostheses, Brain-computer/machine interface
Abstract: Inspired by control theoretic approaches to studying motor control, we experimentally measured how a brain-machine interface (BMI) user responds to an unexpected perturbation. We randomly applied a step cursor position offset while a monkey controlled a BMI cursor using decoded motor cortical spiking activity. The subject was able to rapidly correct for these perturbations and (re)acquire the target regardless of when in the trial this cursor jump occurred. We observed a corrective neural response in motor cortex starting 115 ms after the cursor jump. At no time did the neural response to detecting this externally-induced error manifest itself (through the decoder) as a deleterious velocity change pushing the cursor away from the target. These results show that a user of a high-performance BMI can make rapid, accurate corrections to errors and that, insofar as the neural computations needed to counteract the error may involve motor cortex, these computations do not appear to interfere with BMI cursor control.

Keywords: Motor neuroprostheses, Motor neuroprostheses - Neuromuscular stimulation
Abstract: Dysfunctional motor and sensory functions in limbs due to proximal nerve injury affects numerous people around the world. Complete prosthesis is viable using neural signal acquisition, decoding and functional stimulation. Vital to achieving a completely implantable long term prosthesis solution is the wireless power delivery and data telemetry which can seamlessly power the recording implant to acquire neural data and send control signals to the simulator implant for functional stimulation. The wireless platform eliminates transdermal/percutaneous wires required to power the implants and transfer data which otherwise requires meticulous care to prevent infections and pain, limiting the prosthesis to a temporary solution. We present the wireless platform for the peripheral nerve prosthesis implants which demand large power(100mW) and bandwidth(1.3Mbps) while strictly adhering to the IEEE safety limits on specific absorption rate. The designed platform is shown to be functional in rodents.

Keywords: Motor neuroprostheses - Prostheses, Neural signal processing - Blind source separation, Neuromuscular systems - EMG models, processing and applications
Abstract: For the past six decades, signal processing methods for myoelectric control of prostheses consisted mainly of calculating time- and frequency domain features of the EMG signal. This type of feature extraction considers the surface EMG as colored noise, neglecting its generation as a sum of motor unit activities. In this study we propose the use of motor unit behavior for classifying motor tasks with the aim of myoelectric control. We recorded high-density surface EMG of three patients who underwent targeted muscle reinnervation, and decomposed these signals into motor unit spike trains using an automatic offline EMG decomposition method. From the motor unit spike trains we used the number of discharges in each analysis interval as a feature for a support vector machine classifier. The same classifier was used for discriminating classic time-domain EMG features, for comparison. Classification accuracy was greater for motor unit information than for the classic features (97.06% ± 1.74 vs 85.01% ± 13.66), especially when the number of classes was high (95.11% ± 1.74 vs 69.25% ± 4.04 for 11 classes). These results suggest that the identification of motor unit activity from surface EMG can be a powerful way for pattern recognition in targeted muscle reinnervation patients.

Keywords: Motor neuroprostheses
Abstract: Abstract—Limb movement execution or imagination induce sensorimotor rhythms that can be detected in electroencephalographic (EEG) recordings. This article presents the interest of considering not only the beta frequency band but also the alpha band to detect the elicited EEG rebound, i.e. the increasing of oscillatory power synchronization, at the end of motor imageries. From database 2a of the BCI competition IV, it is shown that this phenomenon can be stronger over the alpha than the beta band and it is experimentally demonstrated that the analysis on the alpha band improves the detection of the end of motor imageries. Moreover a variant method to compute the oscillatory power without referring to a baseline period is proposed; such capacity is useful for self-paced brain-computer interfaces control.

Keywords: Motor neuroprostheses - Prostheses, Neural interfaces - Sensors and body interfaces, Neuromuscular systems - Peripheral mechanisms
Abstract: We propose a new framework for extracting information from extrinsic muscles in the forearm that will allow a continuous, natural and intuitive control of a neuroprosthetic devices and robotic hands. This is achieved through a continuous mapping between muscle activity and joint angles rather than prior discretisation of hand gestures. We instructed 6 able-bodied subjects, to perform everyday object manipulation tasks. We recorded the Electromyographic (EMG) and Mechanomyographic (MMG) activities of 5 extrinsic muscles of the hand in their forearm, while simultaneously monitoring 11 joints of hand and fingers using a sensorised glove.

We used these signals to train a Gaussian Process (GP) and a Vector AutoRegressive Moving Average model with Exogenous inputs (VARMAX) to learn the mapping from current muscle activity and current joint state to predict future hand configurations. We investigated the performances of both models across tasks, subjects and different joints for varying time-lags, finding that both models have good generalisation properties and high correlation even for time-lags in the order of hundreds of milliseconds. Our results suggest that regression is a very appealing tool for natural, intuitive and continuous control of robotic devices, with particular focus on prosthetic replacements where high dexterity is required for complex movements.

Keywords: Motor neuroprostheses - Neuromuscular stimulation, Motor neuroprostheses - Robotics, Neuromuscular systems - Neurorehabilitation
Abstract: Clinicians constantly face the need to rehabilitate stroke patients to re-establish coordinate reach and grasp. Current rehabilitation techniques and assistive tools face the problem aiming at single parts of the problem. We designed flexible modules for a clinical platform, able to provide hybrid reach and grasp support for interacting with common objects. Specific rehabilitation tasks can be implemented by taking advantage of the possibility to quantify the support needed by patients.

Keywords: Motor neuroprostheses, Neural interfaces - Neural stimulation, Neurological disorders
Abstract: Electrical spinal cord stimulation (ESCS) can improve motor control after various neurological disorders. However, current technologies stimulate the spinal cord continuously, irrespective of the subject's intentions. Time varying ESCS synchronized with the movement intentions of the subject may improve the therapeutic effects while reducing the period ESCS is active. Here, we introduce a neuroprosthetic platform capable of time-varying spatially selective ESCS controlled by subject’s movement intentions decoded from motor cortex neuronal activity. A macaque monkey was implanted with (i) a 96-microelectrode array in the leg area of left MI, (ii) an 8-channel electromyogram (EMG) system into eight right leg muscles and (iii) a 16-electrode epidural ESCS array placed over the lumbar spinal cord. Implants were equipped with wireless data transfer modules allowing for simultaneous recording of neuronal activity and EMGs and modulation of ESCS while the monkey walked freely on a treadmill at 1.5km/h. An algorithm predicted Foot Off and Foot Strike events based on neuronal activity and, upon prediction, modulated the ESCS protocols to selectively induced flexion or extension of the right leg during the swing or stance gait phases, respectively. Thus, we modified the locomotion without disrupting the natural rhythmic alternation of movements. Our results provide a substantial step for the development of a neuroprosthesys aimed to reestablish locomotion in paralyzed individuals.

Keywords: Neuromuscular systems - Wearable systems, Motor neuroprostheses - Robotics, Neuromuscular systems - Computational modeling and simulation
Abstract: Hand exoskeletons have been used not only for assistive purpose of the activities of daily living (ADL) for the aged, injured or handicapped people, but also for rehabilitation purpose as power assist device. Ionic polymer metal composite (IPMC) have been developed for a finger exoskeleton actuator, because it is soft and light, consumes small energy. Also, bending deformation of an IPMC strip conforms to finger flexion and extension. In this study, time-delay control (TDC) was applied to an IPMC strip in order to obtain a robust and precise tracking performance. IPMC actuators 20 mm wide and 40 mm long with various thicknesses (1.6 and 2.4 mm) were fabricated. Blocking force of an IPMC strip and bending angle of it were used as control variables. The control experiments with an IPMC strip showed that both the transient and the steady-state responses were similar to those of the desired values, and tracking errors were small. The actuator maintained 100 gf of blocking force with the inverse input gain of 0.01 and the time constant of 25 s, and 10 degree of bending angle the inverse input gain of 0.1 and the time constant of 10 s.

Keywords: Motor neuroprostheses - Prostheses, Neural interfaces - Implantable systems, Neural interfaces - Neural stimulation
Abstract: We describe the development and testing of implantable devices for recording myoelectric signals from residual arm muscles for control of prosthetic hands and a neural stimulator for restoration of touch and proprioception via selective stimulation of peripheral nerves.

Keywords: Sensory neuroprostheses - Auditory, Neuromuscular systems - Computational modeling and simulation, Neural interfaces - Implantable systems
Abstract: Cochlear implants have been proved to be an effective treatment for patients with sensorineural hearing loss.

Among all the approaches that have been developed to design better cochlear implants, 3D model-based simulation stands out due to its detailed description of the electric field which helps reveal the electrophysiological phenomena inside the cochlea.

With the advances in the cochlear implant manufacturing technology, the requirement on simulation accuracy increases. Improving the simulation accuracy relies on two aspects: 1) a better geometrical description of the cochlea that is able to distinguish the subtle differences across patients; 2) a comprehensive and reliable validation of the created 3D model.

In this paper, targeting at high precision simulation, we propose a parametric cochlea model which uses micro-CT images to adapt to different cochlea geometries, then demonstrate its validation process with multi-channel stimulation data measured from a implanted cochlea.

Comparisons between the simulation and validation data show a good match under a variety of stimulation configurations.

The results suggest that the electric field distribution is affected by the geometric characteristics of each individual cochlea. These differences can be correctly reflected by simulations based on a 3D model tuned with personalized data.

Keywords: Sensory neuroprostheses - Auditory, Brain functional imaging - EEG, Sensory neuroprostheses
Abstract: A cochlear implant (CI) can partially restore hearing in patients with severe to profound sensorineural hearing loss. Proper programming and evaluation of the CIs are key aspects determining success in the restoration of hearing for patients. Recent evidence suggests that cortical auditory evoked potentials, elicited via an unattended oddball paradigm, can provide objective information on CI spectral discrimination abilities, which in turn may be useful for assessing speech perception performance. This study investigates the applicability of an acoustic change paradigm for objective evaluation of CI users’ ability to resolve spectral content via single channel electroencephalography. Acoustic change complex (ACC) responses were obtained from 13 CI users and correlated with psychoacoustic spectral discrimination abilities. The applicability of the acoustic change paradigm was compared to that of the unattended oddball paradigm. The neural spectral discrimination threshold, estimated via the ACC responses, showed a non-significant correlation with the behavioral spectral discrimination threshold. In contrast, the neural spectral threshold estimated via an unattended oddball paradigm showed a significant correlation with the behavioral threshold. Results suggest that the ACC is an alternative method to assess CI performance, however, the unattended oddball paradigm is a more robust paradigm than the ACC to objectively evaluate CI users’ ability to resolve spectral ripples.

Keywords: Sensory neuroprostheses, Neural interfaces - Neural stimulation, Sensory neuroprostheses - Signal and vision processing
Abstract: Despite considerable advances in the field of retinal prosthetics during recent years, significant variability remains in the quality of vision restoration for patients. One target for refinement of prosthetic vision is to selectively activate one or more of the ~20 parallel channels of visual information that are established in the retina and subsequently travel to different visual networks in the brain. These different channels result in different spike train response patterns in the retinal ganglion cells (RGCs) which constitute the sole output neuron population of the retina. Here we demonstrate, however, that the genuine visual response patterns of retinal ganglion cells can be altered by electrical stimulation, suggesting that the encoding of visual stimuli by retinal prosthesis devices may require consideration of stimulation-induced changes in the retina. Specifically, we demonstrate that ON and OFF response amplitudes increase significantly after stimulation. This leads to changes in the relative weighting of ON and OFF response types on a cell by cell basis – fundamentally altering the visual stimulus encoded by some RGCs.

Keywords: Sensory neuroprostheses - Visual, Sensory neuroprostheses
Abstract: Retinal ganglion cells (RGCs) in the vertebrate retina display a wide range of dendritic morphologies. In order to isolate the contribution of this physical property, we employed a neural morphology generator to study the differences in firing patterns among RGCs with different dendritic geometrical factors but with identical voltage-gated channel kinetics and distributions. Our results suggest that cell morphology alone is sufficient to generate quantitatively distinct electrophysiological responses, serving as a basis for understanding how structural factors influence RGC behavior.

Keywords: Neural interfaces - Implantable systems, Brain functional imaging - EEG, Neural interfaces - Neural stimulation
Abstract: The human auditory periphery is a complex mechano-electrical system that transduces sound waves into nerve action potentials. In this sensory modality sound conduction to and frequency analysis in the cochlea produce frequency-dependent signal propagation delays. A cochlear implant (CI) is a neural prosthesis that replaces the peripheral auditory system partially by stimulating the auditory nerve electrically. This modality is in turn accompanied by artificial signal transmission delays. This study deals with the question how well the timing of neural excitation in these two modalities fit one another. For this purpose, we investigated wave V latencies of auditory brainstem responses evoked either acoustically (ABR) or electrically via the CI (EABR). The sum of delays consisting of CI signal processing and EABR wave V latencies allowed an estimation of the entire CI-channel-specific delay. We compared these values with ABR wave V latencies measured in normal hearing listeners in different frequency bands. As EABR wave V latencies were shorter than those evoked acoustically, appropriate values for delay elements (FIR group delays) in the CI system were determined and compared with the already implemented group delays. Optimized interaural stimulation timing in unilateral deaf subjects provided with a CI reduces the need for central auditory compensation and can improve speech recognition.

Keywords: Sensory neuroprostheses - Auditory, Human performance - Modelling and prediction, Clinical neurophysiology
Abstract: Electrically-evoked compound action potentials (ECAPs) is an objective measure of peripheral neural encoding of electrical stimulation delivered by cochlear implants (CIs) at the auditory nerve level. ECAPs play a key role in automated CI fitting and outcome diagnosis, as long as presence of genuine ECAP is accurately detected automatically. Combination of ECAP amplitudes and signal-to-noise ratio are shown to efficiently detect true responses, by comparing them to subjective visual expert judgments. Corresponding optimal thresholds were calculated from Receiver-Operating-Characteristic curves. This was conducted separately on three artifact rejection methods: alternate polarity, masker-probe and modified-masker-probe. This model resulted in sensitivity and specificity error of 3.3% in learning, 3.5% in testing and 5.0% in verification. It was found that the following combination of ECAP amplitude and signal-to-noise ratio would be accurate predictors: 22 uV and 1.3 dB SNR thresholds for alternate polarity, 35 uV and -0.2 dB for masker-probe and 44 uV and -0.2 dB for modified-maskerprobe.

Keywords: Sensory neuroprostheses - Auditory, Neural interfaces - Computational modeling and simulation, Neural interfaces - Implantable systems
Abstract: Cochlear implants provide functional hearing to people who are profoundly deaf or hearing impaired by replacing the function of missing inner hair cells with an array of stimulating electrodes. Previous studies developed a modeling framework for predicting the optimal number of electrodes, as well as the optimal locations and usage probabilities of electrodes, from an information theoretic perspective. However, the information theoretic method does not quantify the performance of electrode place discrimination. In this paper, we apply a so-called ‘extreme-learning machine’ to the cochlear implant model to calculate the electrode classification error rates. We also investigate the locations along the electrode array where errors are most likely to occur. We conclude based on our model that i) the classification error rate increases with increasing number of electrodes and the classification errors occur predominantly between adjacent electrodes, ii) by inserting the electrode array deeper into the cochlea, more electrode locations can be distinguished and the electrodes for which most errors occur are determined by the distance and spiral twirling angle between adjacent electrodes.

Keywords: Sensory neuroprostheses - Somatosensory and vestibular, Neural interfaces - Neural stimulation
Abstract: High-performance neuroprostheses designed to reanimate a paralyzed limb following spinal cord injury must restore both movement and sensation. For the latter goal, we are developing a novel strategy focused on encode sensations using microstimulation of the cuneate nucleus (CN) of the brainstem. Here, we characterized the temporal dynamics of downstream cortical excitation and inhibition in response to CN microstimulation in a macaque. A single CN stimulus pulse evoked a fast (7 ms) excitatory response in primary somatosensory cortex (S1) followed by an inhibitory period lasting until 50 ms. The S1 response to a second CN pulse within this inhibitory period was drastically attenuated. Following the inhibition, S1 unit activity rebounded with a prolonged excitatory phase lasting until 800 ms. Within this second excitatory phase were rhythmic peaks of increased unit activity with an alpha-band frequency (8-14 Hz). The rhythmic excitation was specific for perigranular laminae and was stimulus-amplitude dependent. The results show a complex cortical response to CN stimuli and can guide future design of CN stimulus patterns to evoke salient percepts.

Keywords: Sensory neuroprostheses - Auditory, Brain functional imaging - EEG, Neural signal processing
Abstract: Recent evidence suggests that cortical auditory evoked potentials recorded by EEG may be used to obtain an objective measure of spectral sound processing abilities in cochlear implant (CI) users. As speech perception depends on both spectral and temporal processing abilities, developing an objective measure of sound processing in the temporal domain is necessary for a complete evaluation of CI speech performance. This study explored the feasibility to objectively assess sound processing in the temporal domain employing a method based on EEG and complex temporal stimuli such as the Schroeder-phase harmonic complexes. Psychoacoustic discrimination abilities were measured employing a four-interval two-alternative forced choice paradigm. Neural discrimination abilities were measured by recording single-channel EEG during an unattended oddball paradigm. Psychoacoustic and neural discrimination abilities were analyzed for correlation. A strong, but non-significant, correlation was found in three out of six subjects. Schroeder-phase harmonic complexes may have utility as stimuli in the development of an objective measure of temporal processing in CI users. Furthermore, they provide new insights on temporal processing in CI users that may benefit the development of the CI.

Keywords: Sensory neuroprostheses - Auditory, Neural signal processing - Blind source separation, Neural interfaces - Computational modeling and simulation
Abstract: Cochlear implants (CIs) continue to improve the quality of life of deaf patients worldwide. However, empirical parameter fitting sessions are necessary for setting the stimulation parameters of each intra-cochlear electrode. As this procedure requires the conscious participation of the CI recipient, CI parameter fitting for toddlers, disabled and elderly recipients remains extremely challenging. State-of-the-art studies describe alternative parameter fitting strategies using objective EEG measures of auditory perception. For this, the identification and suppression of the cochlear stimulation artifact is the key to a reliable estimation of objective measures in CI patients. Still, the high correlation between the stimulation artifact and the evoked response complicates this task. This paper presents a computational simulation framework for simulating realistic EEG datasets of objective measures in CI patients. This is particularly useful for testing and comparing the performance of signal processing methodologies in a controlled environment. The usefulness of this framework is illustrated in the case of auditory-steady state responses (ASSRs). The performance of two ICA algorithms, infomax and extended infomax, in suppressing the CI stimulation artifacts are compared. The simulations showed that the limited performance of extended infomax as compared to infomax is hard to prove uniquely by clinical EEG datasets while it highly influences the detection of the evoked ASSRs.

Keywords: Sensory neuroprostheses - Somatosensory and vestibular, Brain-computer/machine interface - Robotics applications
Abstract: This study was designed to investigate the feasibility to employ adaptation of the sensory nerve afferents in order to selectively elicit tactile sensations (pressure, low-frequency vibration, and high-frequency vibration) by transcutaneous electrical stimulation. That is, we asked ourselves how one tactile sensation can be elicited with the others suppressed by means of adaptation. Conducting a series of experiments, we observed that a high stimulation intensity, expressed as the pulse amplitude, generally resulted in high effectiveness of adaptation, but the stimulation frequency did not affect the adaptation of the pressure sensation. Also, the monophasic and biphasic pulse trains did not make a significant difference in adaptation. The perception quality of low/high-frequency vibrations was evaluated qualitatively and quantitatively before and after a two-minute high/low-frequency stimulation, indicating that pressure was perceived remarkably less and most of the subjects reported a clearer low/high-frequency vibration. Our experimental observation suggested that adaptation can be employed to selectively elicit each tactile sensation, or, if not, increase the perception quality of the desired sensation by lessening the perception intensity of the other two.

Keywords: Sensory neuroprostheses - Somatosensory and vestibular
Abstract: Vestibular prosthetics are designed to restore sensory information of head rotations; one configuration that has been used both in animal models [1, 2] and humans [3, 4] achieves this via mono-polar stimulation from electrodes in close proximity to the ampullae or vestibular nerve branches. Here we model vestibular nuclei adaptation to the onset of steady-state prosthetic stimulation after a period of aberrant activity due to loss of vestibular function. This model allows us to predict efficacy across stimulation modes and parameters.

Keywords: Human performance - Sensory-motor, Motor neuroprostheses - Prostheses
Abstract: Powered upper limb prostheses typically use EMG to control movement. EMG control is often variable and inefficient, and it is unclear if persons benefit from use of internal models, which have been shown to improve performance with traditional human-machine interfaces. We investigated how internal model use affected errors and effort across a group of 20 subjects using EMG control to perform a tracking task. To vary the ability of subjects to form an internal model, we altered the amount of available information using three visual displays: compensatory, pursuit, and preview. Subjects were more accurate and exerted less effort with visual displays that provided more information and enabled stronger internal model formation (p<0.01).

Keywords: Human performance - Modelling and prediction, Neural signal processing - Blind source separation, Neural signal processing - Time frequency analysis
Abstract: Various studies have shown that the traditional electrooculograms (EOGs) are effective for driving fatigue detection. However, the electrode placement of the traditional EOG recording method is around eyes, which may disturb the subjects' activities, and is not convenient for practical applications. To deal with this problem, we propose a novel electrode placement on forehead and present an effective method to extract horizon electrooculogram (HEO) and vertical electrooculogram (VEO) from forehead EOG. The correlation coefficients between the extracted HEO and VEO and the corresponding traditional HEO and VEO are 0.86 and 0.78, respectively. To alleviate the inconvenience of manually labelling fatigue states, we use the videos recorded by eye tracking glasses to calculate the percentage of eye closure over time, which is a conventional indicator of driving fatigue. We use support vector machine (SVM) for regression analysis and get a rather high prediction correlation coefficient of 0.88 on average.

Keywords: Human performance - Modelling and prediction, Neuromuscular systems - Locomotion, posture and balance, Motor neuroprostheses - Robotics
Abstract: A personalized balance measurement that is easily implemented in the home environment can complement physical rehabilitation protocols. For this purpose, we use the subject-specific center of mass (CoM) estimate offered by the statically equivalent serial chain (SESC) method and the zero rate of change of angular momentum (ZRAM) concept to evaluate balance for a series of dynamic motions. Two healthy subjects were asked to stand on a Wii balance board and their SESC parameters were identified. A set of dynamic motions was to evaluate the rate of change of centroidal angular momentum and the distance of the ZRAM point to the center line of the support polygon. We found a good match between the both balance metrics. As an application example, the subjects performed a dynamic motion (such as walking and abruptly stopping) and the stability was determined in real-time using the ZRAM point from the personalized CoM trajectory. This was implemented with a real-time balance visualization tool based on Kinect measurements for home-based rehabilitation.

Keywords: Brain-computer/machine interface - Biofeedback, Brain functional imaging - EEG, Human performance - Cognition
Abstract: With the growing life expectancy, the number of elderly people is increasing tremendously worldwide. The progressive decrease of synaptic plasticity and neuronal inter-connectivity in the ageing brain, concomitant with alterations in key cognitive abilities such as working memory and attention, may be delayed, stopped or reversed by neurorehabilitation. Hence, current approaches used to modify cognitive capabilities are of utmost importance to contemporary society and often divided into behavioral training procedures and techniques for direct modulation of neural mechanisms. Neurofeedback (NF), which is based on electroencephalogram signals, is used to train individuals on learning how to influence brain function by modulating their own brain rhythms. However, the potential effects of rehabilitation through behavioral training, neuromodulation and even a combined methodology are poorly understood. Differently from the frequently reported longer protocols, an alpha and theta intensive neurofeedback protocol was applied on 14 subjects ageing more than 55. Although the herein presented results suggest that the proposed protocol succeeded to modulate alpha and theta rhythms and led to moderate cognitive improvements, no modulation was apparent on post-training resting state EEG rhythms.

Keywords: Human performance - Modelling and prediction, Neuromuscular systems - Computational modeling and simulation, Motor neuroprostheses - Neuromuscular stimulation
Abstract: Even without visual feedback, humans can accurately determine the shape of objects on the basis of haptic feedback. This feat is achievable despite large variability in sensory and motor uncertainty in estimation of hand pose and object location. In contrast, most neuroprosthetic hands still operate unaware of the shape of the object they are manipulating and can thus only provide limited intelligence for natural control of the hand. We present a computational model for haptic exploration and shape reconstruction derived from mobile robotics: simultaneous localisation and mapping (SLAM). This approach solely relies on the knowledge of object contacts on the end-points, noisy sensory readings and motor control signals. We present a proof-of-principle accurate reconstruction of object shape (e.g. Rubik's cube) from single-finger exploration and propose a straightforward extension to a full hand model with realistic mechanical properties. The proposed framework allows for principled study of natural human haptic exploration and context-aware prosthetics. In conjunction with tactile-enabled prostheses, this could allow for online object recognition and pose adaptation for more natural prosthetic control.

Keywords: Human performance - Sensory-motor, Human performance - Cognition, Sensory neuroprostheses - Somatosensory and vestibular
Abstract: Previous research (Ni, Chen and Andersen, 2010) has found that there are spatial limitations in propagating quantitative depth perception of an occluding surface, from its two vertical edges to the center region. In the present study we examine whether variations in the horizontal extent of subjective surface will also alter perceived depth in vertical surface edges. In two psychophysical experiments, observers were shown temporal interocular unmatched (IOUM) features based on a vertical moving line that was partially occluded by a subjective surface. Metrical depth was perceived and we found that depth perception in the vertical surface edges can be affected by its width when the horizontal surface edges are revealed by two ends of the occluded moving line.

Keywords: Human performance - Cognition, Human performance - Ergonomics and human factors, Neural interfaces - Computational modeling and simulation
Abstract: We explore the possibility of assessing the acquisition of a new skill through electroencephalographic (EEG) measurements. In particular, we propose an experiment to monitor the process of learning to type using the Colemak keyboard layout during a twelve-lessons training. As a first step, we are interested in identifying statistically significant changes in the power spectral density (PSD) of various EEG rhythms at various stages of the learning process. Those changes are taken into account only when a probabilistic measure of the cognitive state ensures the high engagement of the volunteer to the training. Our preliminary experimental results showed changes in the beta and gamma rhythms for seven volunteers during the training process. Such changes are in agreement with previous reports of changes in PSD associated to feature binding and fluid intelligence.

Keywords: Neurological disorders - Diagnostic and evaluation techniques, Human performance - Gait
Abstract: Freezing of Gait (FOG) is a disabling symptom and movement disorder, typically associated with the latter stages of Parkinson’s disease. In this paper, we propose a novel approach for real-time FOG, tremor monitoring and fall detection, consisting of a 3D camera sensor based on the Microsoft Kinect architecture. The system is capable of recognizing freezing episodes (FOG) in a standstill state, tremors and fall incidents, commonly seen in Parkinson’s disease patients. In case of an incident, it automatically alerts relatives and healthcare providers. The system was tested on seven simulated subjects in 12 events indicating that the design was able to detect 99% of the falling incidents, 91% of tremor and 92% of the freezing of gait episodes with an average latency of 300 milliseconds. The performance of the system can be further improved with the deployment of the recently released version of Kinect, capable of providing even higher levels of accuracy.

Keywords: Human performance - Cognition, Neural signal processing
Abstract: Our previous study demonstrated that incongruent audiovisual change could result in the illusory perception of the change in sound intensity. In this study, we investigated the causal cortical networks for audiovisual interactions during this visual-affected auditory discrimination using partial directed coherence. In the early stage (49-198 ms), the audiovisual trials activated more left-to-right connections than unimodal ones, especially those inter-hemispheric frontal-to-occipital and left-centroparietal-to-right-occipital connections. In the middle-late stage (271-420 ms), the audiovisual network presented a right hemispheric dominance, involving more connections to right frontal and occipital areas. More inter-hemispheric connections, including bilateral connections from parietal to frontal cortex, were activated for audiovisual mismatched information processing and decision making.

Keywords: Human performance - Sensory-motor, Neurological disorders - Diagnostic and evaluation techniques, Neuromuscular systems - Neurorehabilitation
Abstract: Clinical telegaming integrates telecare and videogaming to enable a more convenient and enjoyable experience for patients when providers diagnose, monitor, and treat a variety of health problems via web-enabled telecommunications. In recent years, clinical telegaming systems have been applied to physical therapy and rehabilitation, evaluation of mental health, and prevention and management of obesity and diabetes. Parkinson's disease (PD) is suitable for development of new clinical telegaming applications because PD patients are known to experience motor symptoms that can be improved by physical therapy. Recent research suggests that sensory processing deficits may also play an important role in these motor impairments because successful motor function requires multisensory integration. In this paper, we describe a new web-enabled software system that uses clinical telegaming to evaluate and improve multisensory integration ability in users. This software has the potential to be used in diagnostic and therapeutic telegaming for PD patients.

Keywords: Human performance - Modelling and prediction
Abstract: Human motion assessment is key for motor-control rehabilitation. Using standardized definifions and spatiotemporal features - usually presented as a movement cycle diagram- specialists can associate kinematic measures to progress in rehabilitation therapy or motor impairment due to trauma or disease. Although devices for capturing human motion today are cheap and widespread, the automatic interpretation of kinematic data for rehabilitation is still poor in terms of quantitative performance evaluation. In this paper we present an automatic approach to extract spatiotemporal features from kinematic data and present it as a cycle diagram. This is done by translating standard definitions from human movement analisys into mathematical elements of a Switching Linear Dynamic System model. The result is a straight-forward procedure to learn a tracking model from a sample execution. This model is robust when used to automatically extract the movement cycle diagram of the same motion executed in different subject-specific manner such as his own motion speed.

Keywords: Human performance - Sensory-motor, Neural signal processing, Brain-computer/machine interface
Abstract: Streams of information reach the brain in different modalities while driving (both from the vehicle user interface as well as the environment), each requiring a certain degree of attention from the driver. The driver should ideally be able to focus on the road ahead and not on secondary streams such as vehicle alerts, mobile phone, passengers or radio. However, the individual ability to selectively ignore distractions can have a direct influence on driving performance. Here, we use paired-chirp auditory late responses (ALRs) in order to assess long interval cortical inhibition (LICI) in healthy subjects, and compare it to the score in an auditory stream segregation task within a driving simulator. Results show significant correlation between LICI and task scores, suggesting that people with a higher/more effective cortical inhibition as measured by ALRs can ignore distracting streams easily, while people with less effective cortical inhibition find harder to concentrate on a single, more relevant stream. The fundamental results obtained suggest that cortical inhibition may be employed as a predictor of driving performance, useful for the design of auditory human-vehicle interfaces.

Keywords: Human performance - Ergonomics and human factors
Abstract: We propose a novel way of robotic device control with communicative eye movements that could possibly help to solve the problem of false activations during the gaze control, known as the Midas touch problem. The proposed approach can be considered as explicitly based on “communication” between a human operator and a robot. Specifically, we employed gaze patterns that are characteristic for “joint attention” type of communication between two persons. Joint attention gaze patterns are automatized and able to convey information about object location even under a high cognitive load. Therefore, we assumed that they may make robot control with gaze more stable. In a study with 28 healthy participants who were naive to this approach most of them easily acquired robot control with joint attention gaze patterns. The study did not reveal higher preference for communicative type of control. We discuss potential benefits of the new approach that can be tested in future studies.

Keywords: Human performance - Modelling and prediction, Neuromuscular systems - EMG models, processing and applications, Motor learning, neural control, and neuromuscular systems
Abstract: Continuous prediction of dynamic joint angle from surface electromyography (sEMG) signal is one of the most important applications in rehabilitation area for stroke survivors as these can directly reflect the user motor intention. In this study, new shoulder joint angle prediction method in real-time based on the biosignal: sEMG is proposed. Firstly, sEMG to muscle activation model is built up to extract the user intention from contracted muscles and then feed into the extreme learning machine (ELM) to estimate the angle in real-time continuously. The estimated joint angle is then compare with the webcam captured joint angle to analyze the effectiveness of the proposed method. The result reveals that correlation coefficient between actual angle and estimated angle is as high as 0.96 in offline and 0.93 in online mode. In addition, the processing time for the estimation is less than 32ms in both cases which is within the semblance of human natural movements. Therefore, the proposed method is able to predict the user intended movement very well and naturally and hence, it is suitable for real-time applications.

Keywords: Human performance - Ergonomics and human factors, Neuromuscular systems - EMG models, processing and applications, Neural interfaces - Sensors and body interfaces
Abstract: In subjects with complete Spinal Cord Injury (SCI) above C7, the four limbs are paralyzed (quadriplegia). Recovery of grasping movements is then reported as a priority. Indeed, most activities of daily living are achieved through upper limbs. Thus, restoration of hand and forearm active mobility could significantly increase independence and quality of life of these people and decrease their need of human aid. Although most of the subjects plebiscite pharmacological or biological solutions, only orthotics and Functional Electrical Stimulation (FES) allow, so far, to restore hand movements but they are rarely used. Limited ergonomics and comfort of piloting modes could partly explain the low usage of these systems. In this context, our aim is to explore possible solutions for subjects to interact with such devices. In this article, we propose to evaluate the capacity of active upper limb muscles contraction to be used to intuitively control FES in tetraplegic subjects. In this study, we assessed the ability to gradually contract different muscles: trapezius, deltoid, platysma and biceps. Three subjects with C6 to C7 neurological levels of lesion were included. We show that over the active upper limb muscles tested, contraction of the trapezius muscle was considered by the subjects as the most comfortable and could be employed as an intuitive mode of control of functional assistive devices.

Keywords: Human performance - Gait, Neurological disorders - Diagnostic and evaluation techniques, Neuromuscular systems - Wearable systems
Abstract: A general need for Wearable Robots (WRs) is to perform optimal selection of information for control and monitoring during daily assistance with an autonomous systems. A good feature selection algorithm is key to perform automated estimation of the mode of locomotion under environmental variations. Ambulatory body weight support (BWS) systems can be combined with WRs to provide safe ambulation and support in overground walking in cases of lower limb paralysis. This study aimed to develop a support vector machine (SVM) model for binary and multiclass classification that performs gait pattern recognition for different values of partial BWS during overground robot-aided walking. The principal component analysis (PCA) and kernel-based PCA (kPCA) were applied to improve the classification performance. As a result, the combination of temporal and kinematic features showed to improve the accuracy in the discrimination of gait patterns in healthy patients (88%). In SVM multiclass classification the “one-against-one” approach showed to have a more stable performance (true positive and true negative rate are consistent) than “one-against-all” approach and also lower computational cost both for training and SVM’s decision making.

Keywords: Human performance, Human performance - Modelling and prediction, Human performance - Ergonomics and human factors
Abstract: Fatigue is a status of human brain activities, and driving fatigue detection is a topic of great interest all over the world. In this paper, we propose a measure of fatigue produced by eye tracking glasses, and use it as the ground truth to evaluate driving fatigue detection algorithms. Particularly, PERCLOS, which is the percentage of eye closure over the pupil over time, was calculated from eyelid movement data provided by eye tracking glasses. Experiments of a vigilance task were carried out in which both EOG signals and eyelid movement were recorded. The evaluation results of an effective EOG-based fatigue detection algorithm convinced us that our proposed measure is an appropriate candidate for evaluating driving fatigue detection algorithms.

Keywords: Human performance, Brain functional imaging, Neurological disorders
Abstract: Essential Tremor (ET) is one of the most common neurological disorders. It is characterized by an action tremor, usually most prominent during goal directed movements. Even though the underlying pathological mechanisms are still under debate, several studies indicate a dysfunction of the cerebellum. The cerebellum is involved in the integration of intrinsic and extrinsic feedback for error correction during the execution of a motor task. Our goal is to study differences in error correction between ET patients and healthy controls (HC) while performing a constant torque task with the right hand. Furthermore, more insight into the underlying physiological mechanisms is gained using functional MRI (fMRI), simultaneously. Subjects had to perform a constant torque task (target torque 0.75 Nm) with: 1) visual feedback, 2) no feedback. An MR compatible robot arm was developed and tested prior to this experiment to be able to measure a wrist flexion torque. The torque data showed a significant difference between the two feedback conditions in the deviation of the flexion torque from the target force in the ET group, with a significantly larger deviation during the no feedback task. Intergroup analysis showed no significant difference between ET and healthy controls during the visual feedback condition. Imaging data showed differences in the cerebellum, SMA, motor cortex and thalamus between the ET and HC group. However, reliable between-group analysis requires larger groups.

Keywords: Neurological disorders - Stroke, Human performance - Gait, Neural interfaces - Neural stimulation
Abstract: Drop foot syndrome corresponds to the paresis of the muscles involved in lifting the foot during dorsiflexion in the swing phase. This disability can be corrected by applying electrical stimulation to the peroneal nerve or to the tibialis anterior muscle. Developing a stimulator, operating effectively in the patient’s daily life, is the focus of this work. A 2D model of the ankle-tibialis anterior muscle will be used to describe the foot movement. Using offline optimization techniques, first the model inputs will be fit to the measurement data of a patient. Different objective functions e.g. minimizing the fatigue and metabolic cost will be also investigated to find the most suitable optimality criteria which will be used later in online situation on the patients.

Keywords: Deep brain stimulation, Neural interfaces - Sensors and body interfaces, Neuromuscular systems - Wearable systems
Abstract: During deep brain stimulation (DBS) surgery for patients of Parkinson's disease, changes in rigidity are detected by an evaluator by observing the changes in resistance of patient's arm to a passive movement. The aim of this paper was to test the hypothesis that, at the moment of reduction in rigidity, the speed with which the evaluator moves the patient's arms increases, and that this change and its amplitude can be detected with an acceleration sensor. The hypothesis was tested during 9 DBS surgeries using a 3-axis acceleration sensor attached to the evaluator's wrist. Statistical tests revealed that the method was able to identify a reduction in patient's rigidity at lower stimulation amplitudes than those identified by the evaluator.

Keywords: Human performance - Gait, Neuromuscular systems - Locomotion, posture and balance, Neurological disorders - Diagnostic and evaluation techniques
Abstract: Electrical stimulation (ES) assisted correction of foot drop in hemiparetic subjects would benefit from an efficient adaptation of the stimulation pattern regarding foot and/or shank orientations. In order to satisfy practical rehabilitation constraints and to design a feasible solution we propose a body area network (BAN) of embedded inertial measurement units (IMU). This preliminary study presents gait assessment results from experiments realized on seven hemiparetic patients (5 male, 2 female, age: 26 to 65 years, height: 1.63m to 1.85m). The subjects were instructed to walk along a 5 meters GAITRite© electronic walkway (CIR Systems, Inc., Havertown, PA, USA) with two FOX HikoB© (HikoB Villeurbanne, France) inertial measurement units respectively strapped to the foot and the shank of each leg, as illustrated in Figure 1. A motion capture system (VICON© Bonita, USA) combined with an industrial object tracking software (VICON© Tracker) has been used to provide the reference measurements for each IMU and to compare reconstruction results. This work essentially aims at competing 3 existing literature attitude and heading reference system (AHRS) algorithms results to the gold standard ones from the motion capture device. We compare the performances and reliability of Martin-Salaün Mahony [2] and Madgwick [3] approaches for the estimation of the tilt angle of the foot as well as ankle angle in the sagittal plan (Figure 2) in order to select an AHRS adapted algorithm.

Keywords: Human performance - Gait, Neurological disorders, Neuromuscular systems - Wearable systems
Abstract: Despite advances in pharmacological treatments, Parkinson’s disease (PD) patients often suffer from gait disturbances. Parkinsonian gait is characterized by small shuffling steps, bradykinesia, and, in the extreme, akinesia, freezing of gait (FOG), and falling. External cueing (a temporal or spatial stimulus) has been shown to have an immediate and palpable effect on gait in PD patients. However, the positive impact of external cueing therapy in a laboratory setting does not easily carry over or generalize to daily living activities. Indeed, PD patients have difficulties applying the skills learnt in the clinic at home. Thus, permanent cueing devices are essential to follow-up and maintain the progress they achieved in the laboratory. We have developed a mobile app for Google Glass that uses audiovisual cueing to guide gait and prevent freezing episodes in PD patients. In a pilot study with 10 PD patients the effectiveness of the app was tested. The specific goals were to investigate 1) the feasibility of Google Glass as a cueing device, and 2) which cueing modalities are optimal for improving gait with the Google Glass. Patients were asked to navigate through obstacle courses that simulate real life situations, including those known to exacerbate motor deficits. Various motion parameters were measured with a 3D motion capture system. Preliminary results show that the temporal variability in gait as well as the frequency and duration of FOG was reduced.

Keywords: Motor learning, neural control, and neuromuscular systems, Human performance, Neuromuscular systems - Locomotion, posture and balance
Abstract: Rupture of the anterior cruciate ligament is a common sports injury with a high re-injury rate. Understanding motor control post reconstruction may help reduce incidence rates. Currently, the modulation of force production during cutting tasks, when most injuries occur, is not well understood. Recently, Lyapunov exponents (LyE) have been used to understand dynamic stability after ACL injury and reconstruction during gait. The purpose of this study was to apply LyE analysis to understand the stability of multidirectional force production in ACL-reconstructed and healthy, active individuals. Our study included 2 ACL reconstructed and 7 healthy individuals. Subjects stood on 2 force plates and produced GRFs continuously and alternatively to the beat of a metronome set at 60 beats per minute with the aid of real time visual feedback. Subjects completed this test in both the anterior/posterior (AP) and medial/lateral (ML) directions. For LyE analysis, AP and ML GRFs were the variables of interest. Our results indicate the ACL-reconstructed subjects were less stable compared to healthy subjects via larger LyE values. In healthy subjects, there was no difference in LyE values comparing AP and ML tasks and no differences between limbs. ACL-reconstructed subjects showed less dynamic stability than healthy subjects in both impaired and unimpaired limbs.

Keywords: Neuromuscular systems - Computational modeling and simulation, Neuromuscular systems - Locomotion, posture and balance, Motor learning, neural control, and neuromuscular systems
Abstract: A computational model of locomotion is being developed to understand possible mechanisms for the influence of a neuromuscular electrical stimulation (NMES) therapy on step behavior. The therapy in question is one that has been developed for a rodent model of spinal cord injury called NMES+RTT, in which NMES is timed to robotic treadmill training (RTT).

Keywords: Brain functional imaging - EEG, Human performance - Gait, Neural signal processing
Abstract: This study's purpose was to determine if connectivity analyses of EEG data from a cognitive task performed during walking produces reliable results unaffected by movement artifact. We recorded movement artifact using a high density EEG electrode system while 10 subjects walked on a treadmill and synchronized this data to a non gait related event. We used the same system to record EEG data from 19 subjects while they performed a Brooks spatial memory task while both standing and walking. We synchronized this data to the working memory event, which occurred at random times relative to the gait cycle. We calculated connectivity values for both data sets using direct directed transfer function. Consistent with previous seated results, we found elevated connectivity values at alpha and theta frequencies around the working memory task. The values for the cognitive task during both standing and walking were larger than the maximum values found in the movement artifact data. Our results suggest that we can track connectivity changes due to cognitive events that occur at random times relative to the gait cycle during walking using appropriate connectivity measures. These findings suggest future studies on mobile brain imaging with EEG should be able to quantitatively assess electrocortical connectivity in subjects walking in the real world.

Keywords: Neuromuscular systems - Learning and adaption, Motor neuroprostheses - Prostheses, Motor learning, neural control, and neuromuscular systems
Abstract: EMG control of powered upper limb prostheses is difficult and imprecise. One approach for improving control is to help amputees develop more accurate internal models of their prosthetic device. This may be facilitated by an intuitive mapping of neural signals to device movement, a way of providing sensory feedback, or training methods. A first step, arguably, is to understand how an amputation affects adaptation. Here we studied trial-by-trial adaptation in a simple target-directed task with transhumeral amputees and healthy controls. We found that adaptation behavior was indistinguishable between amputees using the residual limb, amputees using the intact limb, and able-bodied subjects. Transhumeral amputees completed the task with larger errors than able-bodied subjects, but there was, perhaps surprisingly, no difference between the residual and intact limb.

Keywords: Motor learning, neural control, and neuromuscular systems, Motor neuroprostheses - Neuromuscular stimulation
Abstract: In this paper, we propose a robust control strategy for control of ankle joint angle using intramuscular functional electrical stimulation (FES). Although, several robust control strategies were proposed for FES utilizing surface electrodes. However, developing a robust control strategy for FES utilizing the intramuscular electrodes is an open problem. The method is based on sliding mode control (SMC) with exponential reaching law and fuzzy neural network (FNN). A learning algorithm based on terminal sliding mode is proposed for estimation of the FNN parameter. The experiments were conducted on three rats. Experimental results show that the proposed strategy provides excellent tracking control with fast convergence. The average of tracking error over all trials of experiments and all rats is 3.6°±0.16.

Keywords: Neuromuscular systems - Neurorehabilitation, Human performance - Sensory-motor, Sensory neuroprostheses
Abstract: Human motions are a result of complex-neural interactions between the central nervous system (CNS), sensory and musculoskeletal systems. In this paper, we are focusing in investigating these interactions relying mainly on the concept of sensory- and muscle-synergies. We hypothesize that the CNS is processing and transferring data from sensors and muscles in a unique low-dimensional signaling to simplify the complexity of environmental inputs and to facilitate recruiting muscles patterns. A pilot study involving computing sensory and muscle synergies of seven healthy participants while performing posture balance tasks was conducted to validate our hypotheses. Changes on the participant’s muscles lengths during performing the task were used to represent proprioceptors and compute sensory synergies. The resultant muscles activities, on the other hand, were recorded and used to estimate muscle synergies. Experimental results suggest that the environmental inputs were translated into lower dimensional signals and used to move the target limb to the desired position immediately after the balance disturbance. Participants who showed better posture response were found to be likely to have a stronger correlation between the utilized sensory and muscle synergies. This preliminary study is considered fundamental to understand the neural strategies among the CNS, sensory and musculoskeletal systems

Keywords: Neuromuscular systems - Locomotion, posture and balance, Motor neuroprostheses, Motor learning, neural control, and neuromuscular systems
Abstract: Falls initiated by slipping are a major cause for concern for lower-limb amputees, due to their lacking the distal musculature that aids in avoiding the initiation of a slip. It has been previously demonstrated that able-bodied individuals can interact safely with slippery surfaces by adapting limb kinematics and altering muscle activity to minimize slipping. Newly developed prosthetic devices have the potential to restore specific gait modes to the user, such as walking on a slippery surface, if only more was known about how the mechanical properties should be regulated in each mode. As a first step towards understanding the mechanics relevant to slip prevention, this study sought to quantify lower-limb muscle activity during steady state walking on a range of slippery surfaces. A specific goal was to quantify how people walk on moderately slippery surfaces that pose a hazard, but are more likely to be found on an everyday basis than some of the surfaces previously studied. Our results showed a significant trend (p<0.001) towards decreasing the level of activity used at the ankle as the floor becomes more slippery. In contrast, there is a significant trend (p<0.001) towards increasing the level of activity used at the knee. These findings suggest a strategy in which the ankle becomes increasingly compliant to maximize the surface area in contact with the floor, while increased activity in proximal muscles is used to help stabilize the legs and trunk for increased safety.

Keywords: Neuromuscular systems - Neurorehabilitation, Neural interfaces - Sensors and body interfaces, Motor learning, neural control, and neuromuscular systems
Abstract: Amputees usually perceive vivid awareness of their lost body parts after the amputation (phantom limbs). Phantom limb pain (PLP) is intense pain that is felt in the phantom limb. The mechanism of PLP is still unclear, but the major hypothesis is that it is derived from dysfunction of the brain. There are a few neurorehabilitation techniques using a mirror or virtual reality (VR) that present the visual image of a phantom limb to the patients, and this produce the movement perception of their phantom limb. Here, we developed a multimodal (visual, auditory, and tactile) VR system to obtain the perception of voluntary phantom limb movements. We applied this system to five PLP patients for three tactile feedback conditions as a pilot study. In conclusion, four of the five patients reported pain amelioration, up to 86% decrease in the tactile feedback condition. In addition, our results demonstrated that the best suited condition of feedback-sense modalities depends on the patient. These results suggest that this system can be applied to a rehabilitation platform to offer flexible neurorehabilitation regimens for each patient.

Keywords: Neuromuscular systems - Neurorehabilitation, Neural interfaces - Neural stimulation, Neuromuscular systems - Learning and adaption
Abstract: People with cervical spinal cord injury have partial or complete loss of ventilatory control and require ventilator assist. Open-loop diaphragmatic pacing can be utilized to provide this assist. A closed-loop diaphragmatic pacing system could overcome the drawbacks for manual titration of the stimulation and respond to changing ventilatory requirements. We have developed a versatile custom hardware platform dubbed “Multimed” for biosignal acquisition and parallel real-time computation, data display and storage. We have also developed a new rodent model for diaphragmatic pacing. Using these we illustrate, to our knowledge for the first-time, the successful ability to perform respiratory flow-phase triggered closed-loop diaphragmatic stimulation with resultant changes in respiratory flow and tidal volume.

Keywords: Neuromuscular systems - Neurorehabilitation, Neurological disorders - Diagnostic and evaluation techniques, Neurological disorders - Stroke
Abstract: Robotic exoskeletons are increasingly being used for neurorehabilitation, due to a number of perceived advantages. Once such advantage is the potential to use the large amounts of previously unavailable measurements to provide continuous assessment of the patient. This study investigates the validity of such measurements through an experimental protocol. Reaching movements within and outside an upper-arm rehabilitation exoskeleton (ArmeoPower) of 10 healthy subjects are compared using five commonly-used kinematic metrics (Peak Speed, Time to Peak Speed, Curvature, Smoothness, Accuracy). The study finds that (1) the robotic exoskeleton significantly affects the reaching movements of healthy subjects, (2) the measurements of the exoskeleton accurately represent the movements of the wrist, and (3) evolution of the in-exoskeleton movements over multiple sessions is indicative of changes in movements outside the robot, even though differences remain – suggesting that evolution of this data may be used to monitor patient progress.

Keywords: Neuromuscular systems - Computational modeling and simulation
Abstract: Physiological tremor is an involuntary and rhythmic movement of the body specially the hands. The vibrations in hand-held surgical instruments caused by physiological tremor can cause unacceptable imprecision in microsurgery. To rectify this problem, many adaptive filtering-based methods have been developed to model the tremor to remove it from the tip of microsurgery devices. The existing tremor modeling algorithms such as the weighted Fourier Linear Combiner (wFLC) algorithm and its extensions operate on the x, y, and z dimensions of the tremor signals independently. These algorithms are blind to the dynamic couplings between the three dimensions. We hypothesized that a system that takes these coupling information into account can model the tremor with more accuracy compared to the existing methods. Tremor data was recorded from five novice subjects and modeled with a novel quaternion weighted Fourier Linear Combiner (QwFLC). We compared the modeling performance of the proposed QwFLC with that of the conventional wFLC algorithm. Results showed that QwFLC improves the modeling performance by about 20% at the cost of higher computational complexity.

Keywords: Neuromuscular systems - Neurorehabilitation, Neuromuscular systems - Wearable systems, Neuromuscular systems - Learning and adaption
Abstract: Neurorehabilitation efficiency increases with therapy intensity and subject’s involvement during physical exercises. Robotic exoskeletons could bring both features, if they could adapt the level of assistance to patient’s motor capacities.

To this aim, we developed an exoskeleton controller, based on adaptive techniques, that can actively modulate the stiffness of the robotic device in function of the subject’s activity.

We tested this control law on one healthy subject with an upper-limb exoskeleton.

The experiment consisted in learning a trajectory imposed by the robot. The early results show the different features allowed by our controller with respect to controllers commonly used for neurorehabilitation with exoskeletons.

Keywords: Neuromuscular systems - Computational modeling and simulation, Brain functional imaging - EEG, Human performance - Sensory-motor
Abstract: Proprioception is a critical component of closed-loop motor control with sensory information being used to dynamically adjust and correct movement. We mapped brain responses in healthy adults to proprioceptive stimulation at different steady state frequencies. A haptic robotic device generated small, short, recurrent force pulses producing displacements of approximately 1 cm of the subject's index finger. Four stimulation frequencies were used: 2 Hz, 3 Hz, 5 Hz and 7 Hz. Finger displacement and high resolution 63 channel EEG were simultaneously recorded.The resulting proprioceptive steady state evoked potentials (PSSEPs) and the compliant kinematics of the subject's index finger were analyzed in the frequency domain. The harmonics of the input force pulse were used to interpolate and extrapolate the estimated kinematic and proprioceptive linear transfer functions beyond the fundamental frequencies of stimulation experiments. We used the relative SNR of the EEG channels to rank and identify the most important components of the spatial and frequency response. A physiologic, task-free, quantification of proprioceptive ability could have applications to diagnose and rehabilitate people with neurodegenerative and motor disorders.

Keywords: Neuromuscular systems - Neurorehabilitation, Motor learning, neural control, and neuromuscular systems, Human performance - Sensory-motor
Abstract: Poor control over the velopharyngeal (VP) port (connection between the oral and nasal cavities) leads to unintelligible speech; this VP dysfunction (VPD) can be due to structural abnormalities, poor motor control, or lack of appropriate feedback (hearing impairment). VP control is not aided by visual feedback since the relevant anatomy is not visible to the speaker or the listener. Here we present initial data from a novel, game-based speech rehabilitation platform designed for children with VPD, in which online feedback of speech nasalization is provided based on measurements of nasal skin vibration and speech acoustics. Twelve pediatric participants (three with VPD) completed one session with the video game and were all able to easily use the game. Over 90% of the participants reported that the game was at least “kind of fun” and that the equipment at least “kind of comfortable”. Over 90% of participants and 100% of their parents said they could use the game at home. Results are promising for further development and long-term testing in individuals with VPD.

Keywords: Neuromuscular systems - EMG models, processing and applications, Neuromuscular systems - Learning and adaption, Neuromuscular systems - Computational modeling and simulation
Abstract: Recent experiments have suggested that the central nervous system (CNS) makes use of muscle synergies as a neural strategy to simplify the control of a variety of movements by using a single pattern of neural command signal. This nature of muscle coordination could have great significance in the treatment and rehabilitation of upper limb impairments for hemiparestic patients post stroke. The use of neuromuscular electrical stimulation (NMES) for neural prosthetics or therapeutic applications has been demonstrated as a promising clinical intervention for stroke patients to recover motor function of the upper extremity. However, the existing NMES systems do not provide control methods for the patient to achieve an individualized and functional rehabilitation training. In this research work, muscle synergies from the flexion-extension elbow antagonistic muscles were studied. Using motion information and EMG signals, muscle synergies were extracted using non-negative matrix factorization (NMF) method. Reconstructed signals obtained from the muscle synergies were then applied to the virtual arm (VA) model to test a synergy based NMES strategy. Results show close resemblance to the original elbow trajectory of normal movements and thus the feasibility to control movements in stroke patients for rehabilitation.

Keywords: Neuromuscular systems - EMG models, processing and applications
Abstract: In this study, we analyzed the existence of ratedependent hysteresis in the electromyography (EMG)-force relationship. Eight able-bodied subjects participated in the experiment. Surface EMG signals were acquired from flexor pollicis longus muscle from 0% to 100% maximum voluntary contraction (MVC). The subject was asked to gradually increase grasping force from 0% to 100% MVC and decrease grasping force from 100% to 0% MVC at five different frequencies (1.5, 1, 0.5, 0.25 and 0.125 Hz). Mean absolute value (MAV) was chosen to represent the EMG signals and force signals. In order to compare differences in force between contraction and relaxation periods to EMG activity among different frequency conditions, a hysteresis index (HI), defined as an area inside the hysteresis cycle, was adopted. The results showed that all mean values of HI in different frequency conditions were larger than 0, which proved that hysteresis cycles existed in all frequency conditions. The results also showed that the HI values in different frequency conditions were significantly different from each other (p < 0.005), which proved hysteresis effects in EMG-force relationship were rate-dependent. The rate-dependent hysteresis in EMG-force relationship has a huge potential to improve the estimation performance of grasping force from EMG.

Keywords: Neuromuscular systems - EMG models, processing and applications, Neural signal processing
Abstract: Projecting a high dimensional feature to a low dimensional feature without compromising the feature characteristic is a challenging task. This paper proposes a novel dimensionality reduction constituted from the integration of extreme learning machine (ELM) and spectral regression (SR). The ELM is built upon the structure of the unsupervised ELM. The hidden layer weights are determined randomly while the output weight is calculated using the spectral regression. The flexibility of the SR that can incorporate label into consideration leads a new supervised dimensionality reduction called SRELM. Generally speaking, SRELM is an unsupervised system in term of ELM yet it is a supervised system in term of dimensionality reduction. In this paper, SRELM is implemented in the finger movement classification based on electromyography signals from two channels. The experimental results show that SRELM is better than spectral regression linear discriminant analysis (SRDA) and principal component analysis (PCA). It is also comparable to uncorrelated linear discriminant analysis (ULDA). In addition, it has better class separability than SRDA and ULDA.

Keywords: Neuromuscular systems - EMG models, processing and applications
Abstract: Electromyostimulation has been used for several decades by athletes or physiotherapists in order to create a muscular reinforcement. However, the efficiency of electromyostimulation is limited by muscular fatigue and by induced pain. Currently, the systems of electromyostimulation do not adapt the stimulation parameters automatically by taking into account physiological parameters such as muscular fatigue. To adapt the stimulation parameters to muscular responses and in order to optimize the rehabilitation sessions, a control of force using an indicator of muscular fatigue could be used. In this paper, we propose two ways to control the force by using a physiological model which includes the effects of muscular fatigue.

Keywords: Neuromuscular systems - Computational modeling and simulation, Neural interfaces - Computational modeling and simulation, Brain physiology and modeling - Neuron modeling and simulation
Abstract: Urinary incontinence (UI) is the involuntary loss of urine that creates a social or hygiene problem, which has detrimental effects on quality of life. Detrusor smooth muscle(DSM)instability is a major cause of UI. Different ion channels within bladder DSM play a role in generating electrical activities such as action potentials and synaptic depolarizations. The aim is to establish a computational model to quantitatively simulate DSM action potential(AP) and investigate pathophysiological mechanisms governing normal and dysfunctional bladder activities. In line with recent experimental evidence, adapting the Hodgkin- Huxley formulation in the NEURON platform, we construct mathematical models for seven ionic currents of DSM, where the magnitudes and kinetics of each ionic current are described by differential equations, in terms of maximal conductances, electro chemical gradients, and voltage-dependent activation/inactivation gating variables. These quantifications are validated by the reconstruction of individual experimental ionic currents obtained under voltage clamp. Our simulated AP has been validated by comparing with experimental recordings and shows good correspondence in terms of amplitude and shape. In summary, this mathematical model contributes an elemental tool to investigate the physiological ionic mechanisms underlying the spikes in DSM, which on turn shed light on genesis of UI.

Keywords: Neuromuscular systems - EMG models, processing and applications, Neural interfaces - Sensors and body interfaces, Neuromuscular systems - Learning and adaption
Abstract: Surface Electromyography (EMG) has been considered as a viable human-machine interface in the context of human-centered robotics. In order to interpret human muscle activities into motion intentions, various pattern classification methods was proposed for human motion/gesture classification, which provided binary command for myoelectric control. To obtain complex motions coordinated by multiple DoFs, single DoF was usually sequentially classified and activated, which is not intuitive and efficient comparing with the natural motor strategy of the human. In this work, we investigated the motion classification methods from EMG for intuitive and simultaneous activation of multiple DoFs during 3-D arm motions. In the experiments, all motions were performed naturally rather than under the condition of maximum muscle contractions or other kinematic constraints. The combination of two EMG timedomain features after principal component analysis (PCA) processing is considered as the suitable choice considering both the classification accuracy and feasibility for robot control. For the motion classification method, least-square support vector machine (LS-SVM) represents higher classification accuracy for five arm motion classification across eight subjects with respect to other four methods which were popularly used in the previous works. The proposed method is hopefully applied in a EMG-driven simultaneous and proportional kinematics estimation systems for decoding model selection.

Keywords: Deep brain stimulation, Neural interfaces - Implantable systems, Neural signal processing - Nonlinear analysis
Abstract: We have found that a set of support vector machines operating upon local field potentials sensed from an implanted DBS lead can identify the contact chosen by the physician for the patient’s STN DBS therapy with 91% accuracy. The finding is based on a small data set and thus subject to change with further data collection and cross-validation. Nevertheless, the results suggest that an algorithm for selecting an effective contact for STN DBS based on the signals sensed from the DBS lead may be feasible.

Keywords: Deep brain stimulation, Neural interfaces - Implantable systems, Neurological disorders
Abstract: In this paper we present a prototype proof-of-concept for a closed-loop deep brain stimulation system for patients with essential tremor. This system makes use of sensed movement intentions via EEG to determine when stimulation is required and automatically enables stimulation only when needed. We demonstrate this system using a healthy subject and a benchtop experimental prototype. By limiting stimulation to only when it is therapeutically required, implanted neurostimulators can be more power efficient and potentially limit the period where patients experience side-effects to only the time when therapy is needed.

Keywords: Deep brain stimulation, Brain physiology and modeling - Neuron modeling and simulation, Neuromuscular systems - Computational modeling and simulation
Abstract: Oscillatory neural activity in the beta frequency band (12-30 Hz) is elevated in Parkinson’s disease and is correlated with the associated motor symptoms. These oscillations, which can be monitored through the local field potential (LFP) recorded by a deep brain stimulation (DBS) electrode, can give insight into the mechanisms of action, as well as treatment efficacy, of DBS. A detailed physiological model of the cortico-basal ganglia network during DBS of the subthalamic nucleus (STN) is presented. The model incorporates extracellular stimulation of STN afferent fibers, with both orthodromic and antidromic activation, and the LFP detected at the electrode. Pathological beta-band oscillations within the cortico-basal ganglia network were simulated and found to be attenuated following the application of DBS. The effects of varying DBS parameters, including pulse amplitude, duration and frequency, on the LFP at the DBS electrode were then assessed. The model presented here can be further used to understand the interaction of DBS with the complex dynamics of the cortico-basal ganglia network and subsequent changes observed in the LFP.

Keywords: Deep brain stimulation
Abstract: Models of the cortico-basal ganglia network and volume conductor models of the brain can help to gain insight into the mechanisms of action of deep brain stimulation (DBS). In this study, the coupling of a network model under Parkinsonian conditions to the extracellular field distribution obtained from a 3D finite element model of a rodent's brain during DBS is presented. This coupled model is used to investigate the influence of variations in the electrical properties and thickness of the encapsulation tissue, which is formed around the electrode body after implantation, on the suppression of oscillatory neural activity during DBS. First results suggest that variation in the properties of the encapsulation tissue, within the range examined, have a limited influence on the suppression of pathological oscillatory activity during DBS in rodents.

Keywords: Deep brain stimulation, Brain physiology and modeling - Neuron modeling and simulation
Abstract: Computational models of deep brain stimulation (DBS) have played a key role in investigating the mechanisms of action of DBS therapies. By estimating a volume of tissue directly modulated by DBS, one can relate the pathways within those volumes to the therapeutic efficacy of a particular DBS setting. With the advent of higher-density DBS electrode arrays, there is a growing need for a systematic method to quantify the morphology of the modulated volumes within the brain. In this study, we applied the tools of spherical statistics to quantify such morphologies through the application of a computational model of a directionally segmented DBS array. The same statistical techniques have broad applications to characterizing distributions of in-vivo electrophysiological recordings and histological labeling of neurons.

Keywords: Deep brain stimulation, Neural interfaces - Computational modeling and simulation, Neurological disorders
Abstract: We present a novel closed-loop subthalamic nucleus (STN) deep brain stimulation (DBS) scheme for Parkinson’s disease (PD) and obsessive-compulsive disorder (OCD). The algorithm is designed to effectuate real-time, adaptive stimulation employing the outcome of the 0-1 test for chaos as a state-specific biomarker. In case of a null outcome, the system identifies optimal patterns of stimulation desynchronizing pathologic neuronal activity with minimal energy consumption, on grounds of a stochastic dynamical model and an appropriately formulated cost function. Simulations are performed utilizing microelectrode recordings (MERs) acquired during 8 and 2 DBS surgical interventions for PD and OCD, respectively.

Keywords: Brain physiology and modeling - Neural dynamics and computation, Brain functional imaging, Brain physiology and modeling - Neural circuits
Abstract:

Keywords: Brain functional imaging
Abstract: A single imaging technique can reveal only a small part of the brains’ complex machinery. Here, we present a correlative framework to combine complementary imaging technique and obtain a more comprehensive view of the encephalon. The same neuron imaged in vivo with two-photon fluorescence microscopy could be retrieved ex-vivo and characterized in terms of its ultrastructural features by means of electron microscopy. On the other hand, we describe a method to integrate data from in vivo two-photon fluorescence imaging and ex vivo light sheet microscopy, based on the use of major blood vessels as reference chart. This correlative approach allows placing in a three-dimensional anatomic context the neurons whose dynamics have been observed with high detail in vivo.

Keywords: Brain functional imaging
Abstract: The combination of optical methods with, optogenetic actuators and reporters enables today an all - optical analysis of a well - defined neuronal population within intact neuronal circuits and systems. However,these tools imply a number of constraints on the opto-stimulation method to be used. A perfect light-delivery method should be: efficient, multi-scale, allow millisecond temporal resolution and μm spatial resolution, and be robust to scattering. Several solutions have recently been proposed to fulfill these requirements, and can be divided in two main categories: laser scanning and parallel excitation methods. Laser scanning methods use galvanometric mirrors or acousto-optic devices to quickly scan a laser beam across several positions. With parallel methods all the regions in the target area are excited simultaneously. Here we review a series phase-modulation parallel methods for single-photon (1P) and two-photon (2P) patterned photostimulation and demonstrate their use in exemplary experiments in vitro and in vivo.

Keywords: Brain physiology and modeling - Neural circuits
Abstract: Capturing the dynamics of neuronal activity across whole nervous systems at high temporal resolution has been a long-standing dream in neuroscience. While point-scanning microscopy methods provide the necessary 3D resolution, their volume acquisition rates are limited. Widefield microscopes on the other hand do not provide sufficient optical sectioning capability. We recently implemented two complementary fluorescence microscopy methods that allow for simultaneous whole-animal imaging of genetically encoded calcium indicator activity in C. elegans, and whole-brain read-out in zebrafish larvae. While Wide-field Temporal Focusing Microscopy “sculpts” the spectral components of femtosecond laser pulses to achieve sectioning, Light Field Deconvolution Microscopy, a tomography related method, uses a microlens array to simultaneously capture spatial and angular information followed by computational reconstruction to acquire volumetric information from a single sensor exposure. Here, we discuss our recent results using both techniques for acquiring whole-brain functional imaging data at speeds up to tens of Hertz and near single cell resolution for small model organisms.

Keywords: Brain functional imaging
Abstract: We present a technique for fully noninvasive acquisition of real-time volumetric multispectral optoacoustic data from whole mouse brain. The neuroimaging capacity of the new methodology is demonstrated here by simultaneous label-free assessment of multiple stimulus-evoked hemodynamic responses, including blood oxygenation, total hemoglobin, cerebral blood volume, oxygenized and deoxygenized hemoglobin.

Keywords: Brain physiology and modeling - Neural circuits
Abstract: In addition to the widely-used ability to selectively target specific cell types, optogenetics combined with other neurophotonic strategies also offer an exciting path towards spatio-temporally-controlled targeting: projected patterns of light can be used to selectively and flexibly control and image activity patterns distributed across entire populations of neurons. When natural photoreception is disrupted, as in outer-retinal degenerative diseases, stimulation of surviving nerve cells offers a potential strategy for bypassing compromised neural circuits, inspiring early development of optogenetic retinal prostheses. Selectively exciting large neural populations is essential for eliciting meaningful perceptions in the brain. Here, we present our recent work on distributed neuronal interfacing with large populations of optically accessible, optogenetically transduced neurons in two-dimensions (retinas) and three-dimensions (bioengineered brain-like 'optonets'). Our results demonstrate that patterned computer-generated Holographic Optical Neural Stimulation (HONS) can achieve millisecond temporal precision and cellular resolut ion as a path towards simultaneously controlling populations of retinal ganglion cells, and that new adaptations of multiphoton temporal-focusing holography provides a powerful tool for distributed 3D imaging & control. HONS pattern projection combined with high resolution imaging provides a path towards all-optical bidirectional interfacing, and is also being translate towards in vivo applications.

Keywords: Deep brain stimulation, Neural interfaces - Computational modeling and simulation, Neurological disorders - Epilepsy
Abstract: Low intensity Local Direct Current Stimulation (LDCS) is an electrical stimulation technique that has been poorly investigated in vivo in the field of epilepsy. This study addresses the computational as well as the experimental in vivo effects of low intensity DC stimulation currents on Hippocampal Paroxysmal Discharges (HPDs), a common form of interictal discharges in mesial temporal lobe epilepsy. The results highlight the significance of polarity-dependent effects in silico as well as in freely moving epileptic mice. In conclusion, this combined in silico- in vivo approach shows that cathodal LDCS can significantly reduce the occurrence of HPDs.

Keywords: Neurological disorders - Epilepsy, Neural signal processing - Time frequency analysis, Neurological disorders
Abstract: Genetically modified Mecp2-deficient mice provide a unique model of epilepsy associated with Rett syndrome. Examination of intracranial electroencephalogram (iEEG) recordings from mice lacking mecp2 function have revealed the presence of spontaneous epileptiform activity. To date the majority of these studies have focused on the low frequencies oscillations (LFOs), however, recent findings suggest there may be a link between high frequency oscillations (HFOs) and epileptogenesis. In this study LFO-HFO modulations were examined identifying delta-like HFO modulation in male Mecp2-deficient mice diminishes after mecp2 gene reactivation therapy. These delta-like HFO modulations were found to be strongly associated with long duration epileptiform discharges found primarily in the male non-rescue Mecp2-deficient mice. The differences in HFO interactions with delta LFOs in male non-rescue and rescue mice could potentially be used as a biomarker for identifying the presence of seizure activity.

Keywords: Neurological disorders - Stroke, Brain functional imaging, Neural interfaces - Neuroimaging
Abstract: Stroke recovery involves a battery of plastic changes in the brain. Blood oxygen level dependent (BOLD) functional magnetic resonance imaging (fMRI) technology provides brain activation information with exquisite spatial resolution as a powerful tool for investigating changes in brain plasticity. In this paper, we performed a longitudinal study examining plasticity of functional activation as exhibited by BOLD-fMRI following stroke. Data were collected from 11 patients with corticospinal tract (CST) damage at three stages of recovery, i.e., acute stage (<2wks), early stage (1mon-3mons) and chronic stage (>3mons) post stroke. The evolution of cortical activations for both affected and unaffected hand motion tasks were studied. Quantitative activation measurements including the effective size and sum of t values were calculated and the correlations of these values with patient Fugl-Meyer index were analyzed across all stages. Stroke patients showed a shift from bilateral activation in acute and early stage to the ipsilesional activation in chronic stage when performing a movement task with the affected hand, which suggests a compensation effect from the contralesional hemisphere during the recovery process. The correlation analysis showed a significantly negative correlation with cingulate cortex activity at early stage from both quantitative activation measurements, implying the special role of cingulate cortex in stroke recovery.

Keywords: Neurological disorders - Stroke, Brain functional imaging
Abstract: Stroke, a common brain injury, could result in both structural and functional impairments. Sensory remapping has been thought to play a special role in cortical plasticity, which contributes to the recovery after ischemic stroke. Previous studies using imaging and electrophysiological methods have found that sensory representations after stroke would remap not only to the surrounding areas, but also to remote areas. On a rodent photothrombotic stroke model using optical intrinsic signal (OIS) imaging, the present study attempts to investigate the remapping of sensory representation in contralesional cortex to ipsilesional hindlimb stimulation. Quantitative analysis revealed an overall expansion of hindlimb representation in contralesional cortex after stroke. Moreover, results indicated that hindlimb representation become less correlated to different stimulation intensity in contrast to the positive correlation before stroke. We speculate that diaschisis in the acute stage of stroke might account for such a change in contralesional sensory representation.

Keywords: Neurological disorders - Epilepsy, Neural signal processing - Blind source separation
Abstract: In blind source separation (BSS), sparsity is proved to be very advantageous. If data is not sparse in its current domain, it can be modelled as sparse linear combinations of elements of a chosen dictionary. The choice of dictionary that sparsifies the data is very important. In this paper the dictionary is pre-specified based on chirplet modelling of various kinds of real epileptic spikes. Dictionary look up together with source separation is used to extract the closest source to the source of interest from the scalp EEG measurements. The algorithm has been tested on synthetic and real data consisting of epileptic discharges, and the results are compared with those of traditional BSS.

Keywords: Neurological disorders - Epilepsy, Deep brain stimulation, Neural signal processing
Abstract: For most people suffering from epilepsy the unpredictability of their spontaneous seizures considerably contributes to the burden of their disease. Therefore active and passive seizure prediction methods have been investigated since the 1970s. Among them, the relative phase clustering index (rPCI) has been reported to be a promising stimulation based technique that offers the possibility to localize the seizure onset zone as well as the likelihood for a seizure occurrence within different time ranges. The adaptation of this measurement to animal models of epilepsy would not only offer the possibility to realize and test closed-loop setups based on identification of a preictal period; it would also provide the possibility to test therapeutic paradigms in a much faster and more efficient way. We investigated the transferability of the concept of the rPCI to the epilepsy model of the rat. Adaptation of stimulation parameters, especially the current amplitude, will be necessary for the application in animals, due to different electrodes as well as different anatomical sizes. Therefore, the influence of the stimulation parameters on the results of this active probing are analyzed in this study. While the amplitude and the pulse width (PWD) seem to affect the determined phase clustering index, no correlation with the frequency of test stimulus application could be proven.

Keywords: Motor learning, neural control, and neuromuscular systems, Neural signal processing, Clinical neurophysiology
Abstract: Oscillations in the alpha (8-12 Hz), beta (15-35 Hz) and gamma (35-60 Hz) frequency bands are commonly observed in recordings from the primary motor cortex. Coherence analysis based on motor unit spike trains is commonly used to quantify the degree of shared cortical input and the common modulation of motor unit discharge rates between muscles. In this study, intra-muscular coherence is used to investigate the alterations in the neural drive to the First Dorsal Interosseous muscle directly after a fatiguing contraction and following a rest period. An increase in coherence was observed for all frequency bands examined, which was statistically significant within the alpha and beta frequency ranges. There was no consistent difference between the coherence estimates obtained pre-fatigue and those reported after the recovery period. The increase in beta band coherence post-fatigue may indicate increased levels of cortical drive to the motor unit pool. Although the functional significance behind the increase in beta frequency coherence is unclear, it may aid in the coordination of muscle activity to compensate for the decline in the force generating capacity after fatigue.

Keywords: Neurological disorders - Diagnostic and evaluation techniques, Clinical neurophysiology, Neural interfaces - Neural stimulation
Abstract: Evidence of Current Perception Threshold (CPT) to assess neural fiber function in healthy subjects suggests greater discrimination for stimuli at 1, 250 and 3000 Hz than at 5, 250 and 2000 Hz. Similar data are not yet described for subjects affected by diabetes mellitus. This study proposes to provide a comparative database of parameters obtained with sinusoidal electrical stimulation applied at 1, 5, 250, 2000 and 3000 Hz in subjects affected and non-affected by the diabetes. Ninety subjects were recruited to compose the control (n=45) and diabetic (n=45) groups. The CPT intensities and the reaction times obtained for left and right feet stimulation show responses characterized by weaker intensities (533 to 1671 µA) and longer delays (1.24 to 1.42 s) at low frequencies (1 and 5 Hz) than the intensities (3965 to 5685 µA) and delays (0.96 to 1.12 s) obtained at high frequencies (2000 and 3000 Hz). Moreover, the low frequency stimulation trials evoked up to 73% of the self-reports corresponding to C-fiber sensations while high frequencies evoked up to 60% of the self-reports related to Aβ-fiber sensations. Moreover, the subjects affected by diabetes needed a stronger intensity of stimulation current in order to perceive consistently the sensations evoked by Aβ-fibers (hypoesthesia). In addition, the findings reinforce the suggestion that the discrimination between sensations related to different neural fibers is increased for stimuli at 1, 250 and 3000 Hz for both gro

Keywords: Neural signal processing, Brain physiology and modeling - Neuron modeling and simulation
Abstract: A Bayesian neural decoding model requires the use of a neural encoding model. The parameters of this encoding model are generally fitted to some training data and used subsequently in the decoding of a test stimulus. One encoding model that has been widely used for the auditory system implies the use of a generalized linear model (GLM) having three parameters accounting respectively for the spontaneous rate of the neuron, its spectro-temporal receptive field and the dynamics of the neuron. Here we present a cross-validation study of the goodness of fit of a GLM encoding model in order to quantify the effects on the fitting of using model parameters estimated from a training data set on a test data set. The goodness of fit is measured using Kolmogorov-Smirnov (KS) statistics. It is observed that using trained parameters on the test data yields a much poorer than expected goodness of fit of the auditory encoding model, with only 5% of the neurons having a suitable fit. Moreover, we show that this poor goodness of fit is the result of all three parameters of the auditory GLM encoding model being inadequate for the test data. Using such parameters in a decoding framework may thus result in a biased decoded stimulus.

Keywords: Neural signal processing - Nonlinear analysis, Brain-computer/machine interface
Abstract: Automated electroencephalography (EEG) based affect recognition has gained a lot of interest recently, with clinical, human-computer interaction, neuromarketing, and even multimedia applications. Typically, conventional EEG features such as spectral power, coherence, and frontal asymmetry have been used to characterize affective states. Recently, cross-frequency coupling measures have also been explored. In this paper, we propose a new feature set that combines some of these aforementioned paradigms. First, the full-band EEG signal is decomposed into 4 subband signals, namely theta, alpha, beta, and gamma. The amplitude modulation (or envelope) of these signals is then computed via a Hilbert transform. These amplitude modulations are further decomposed into 10 cross-frequency coupling patterns (e.g., gamma-beta coupling pattern). The mutual information between each of these 10 patterns is then calculated for all inter-hemispheric EEG electrode pairs. To gauge the effectiveness of the newly-proposed feature set, the so-called DEAP database was used. Experimental results show the proposed feature set outperforming conventional ones for estimation of arousal, valence, dominance, and liking affective dimensions. Gains of up to 20% could be achieved when the proposed features were fused with spectral power and asymmetry index features, thus suggesting complementarity between spectral and spectro-temporal features for automated affective state recognition.

Keywords: Neural signal processing - Blind source separation, Brain functional imaging - EEG, Brain-computer/machine interface
Abstract: The needs for online Independent Component Analysis (ICA) algorithms arise in a range of fields such as continuous clinical assessment and brain-computer interface (BCI). Among the online ICA methods, online recursive ICA algorithm (ORICA) has attractive properties of fast convergence and low computational complexity. However, there hasn't been a systematic comparison between an online ICA method such as ORICA and other offline (batch-mode) ICA algorithms on real EEG data. This study compared ORICA with ten ICA algorithms in terms of their decomposition quality, validity of source characteristics, and computational complexity on the thirteen experimental 71-ch EEG datasets. Empirical results showed that ORICA achieved higher mutual information reduction (MIR) and extracted more near-dipolar sources than algorithms such as FastICA, JADE, and SOBI did while the performance of ORICA approached that of the best-performed Infomax-based algorithms. Furthermore, ORICA outperforms most of ICA methods in terms of the computational complexity. The properties of fast convergence and low computational complexity of ORICA enable the realization of real-time online ICA process, which has further applications such as real-time functional neuroimaging, artifact reduction, and adaptive BCI.

Keywords: Neural signal processing - Nonlinear analysis, Neural interfaces - Sensors and body interfaces
Abstract: The features of the application of multifractal analysis to assess the interrelation of the short-term signals of microwave radiation of the human brain and heart rate variability (HRV) are described. Data are presented for periods (20,40)s of microwave radiation signals and periods (6.5,25)s of HRV showing low level of systematic divergence. Similar properties for periods (50,70)s of microwave radiation signals and periods (25,300)s of HRV are found. The results indicate the interface between microwave radiation fluctuations of the human brain and HRV signals, which represent autonomic nervous system activity.

Keywords: Neural signal processing - Nonlinear analysis, Neurological disorders - Epilepsy, Neural signal processing
Abstract: This study investigated the seizure and non- seizure state transitions in the intracranial electroencephalogram (iEEG) recordings of extratemporal lobe epilepsy patients. Cross-frequency coupling between low and high frequency oscillations in conjunction with an unsupervised learning algorithm – namely, hidden Markov models – was used to objectively identify seizure and non-seizure states as well as transition states. Channels consistently capturing two and/or three distinct states in a 32-channel iEEG array were able to identify regions of interest located in resected tissue of patients who experienced improved post-surgical outcomes.

Keywords: Neural signal processing
Abstract: Cascading series of events are a salient feature of neural networks, where neuron spikes may stimulate or inhibit spike activity in other neurons. Only individual spike times associated with each neuron are observed, usually without knowledge of the underlying relationships among neurons. This paper addresses the challenge of tracking how spikes within such networks stimulate or influence future events. The proposed approach is an online learning framework well-suited to streaming data, using a multivariate Hawkes point process model to encapsulate autoregressive features of observed events within the network. Recent work on online learning in dynamic environments is leveraged not only to exploit the dynamics within the underlying network, but also to track that network structure as it evolves. Regret bounds and experimental results demonstrate that the proposed method performs nearly as well as an oracle or batch algorithm.

Keywords: Neural signal processing, Neurological disorders - Epilepsy
Abstract: High frequency oscillations (HFOs) have been considered as a promising clinical biomarker of epileptogenic regions in brain. Due to their low amplitude, short duration, and variability in patterns, the visual identification of HFOs in long-term continuous intracranial EEG (iEEG) is cumbersome. The aim of our study is to improve and automatize the detection of HFO patterns by developing analysis tools based on an unsupervised k-means clustering method exploring the time-frequency content of iEEG. The clustering approach successfully isolated HFOs from noise, artifacts, and arbitrary spikes. We tested this technique on three subjects. Using this algorithm we were able to localize the seizure onset area in all of the subjects. The channel with maximum number of HFOs was associated with the seizure onset.

Keywords: Neural signal processing - Blind source separation, Motor neuroprostheses - Prostheses, Neuromuscular systems - Peripheral mechanisms
Abstract: The correlation structure of natural hand & finger movements suggests that their motion is controlled in a lower--dimensional space than would be possible given their mechanical nature. Yet, it is unclear whether this low dimensional embedding is relevant to how the brain represents motor actions and how we can decode it for Brain-Machine Interface applications. We collected large data set of natural hand movement kinematics and analysed it using a novel sparse coding and dictionary learning approach -- Sparse Movement Decomposition (SMD), which captures the embedding of the data in terms of spatial and temporal structure. We show that our sparse codes over natural movement statistics give a more parsimonious representation than the simple correlation structure. This suggest that, like V1 neuron receptive fields can be predicted from sparse code over natural image statistics, motor control may be encoded in such a manner. We further show how our sparse coding can help understand the temporal structure of behaviour, and thus our technique may be used for behavioural fingerprinting in diagnostics and for more naturalistic neuroprosthetic control.

Keywords: Neural signal processing - Time frequency analysis, Neural interfaces - Neural stimulation, Deep brain stimulation
Abstract: Neurological implantable devices with electrophysiological sensing capabilities may enable new medical therapies and diagnostics. Of specific interest is phase-dependent delivery of therapy—such as electrical stimulation—as a potential method to enhance therapy effectiveness in improving diseased or damaged physiological processes with respect to rhythmic biomarkers, while minimizing the average energy required to deliver the therapy. To address this need for phase-detection, within the constraints of device power limits, we have developed a computationally efficient, causal, real-time Fourier transform (RTFT) for use as a phase detection method that is both general and highly configurable. The application of this method to theta-band local field potentials recorded from the brains of sheep is demonstrated.

Keywords: Neural signal processing - Time frequency analysis, Brain-computer/machine interface, Brain functional imaging - EEG
Abstract: EEG-based Brain Computer Interfaces (BCIs) are quite noisy brain signals recorded from the scalp (electroencephalography, EEG) to translate the user's intent into action. This is usually achieved by looking at the pattern of brain activity across many trials while the subject is imagining the performance of an instructed action -- the process known as motor imagery. Nevertheless, existing motor imagery classification algorithms do not always achieve good performances because of the noisy and non-stationary nature of the EEG signal and inter-subject variability. Thus, current EEG BCI takes a considerable upfront toll on patients, who have to submit to lengthy training sessions before even being able to use the BCI. In this study, we developed a data-efficient classifier for left/right hand motor imagery by combining in our pattern recognition both the oscillation frequency range and the scalp location. We achieve this by using a combination of Morlet wavelet and Common Spatial Pattern theory to deal with non-stationarity and noise. The system achieves an average accuracy of 88% across subjects and was trained by about a dozen training (10-15) examples per class reducing the size of the training pool by up to a 100-fold, making it very data-efficient way for EEG BCI.

Keywords: Neural signal processing, Neural signal processing - Time frequency analysis
Abstract: This study aims to decrease the gap between invasive high-resolution data acquisition and non-invasive measurement in the animal model under auditory stimuli. We approach this problem by analyzing the degree of shared information between the phase of local field potentials (LFPs) of auditory cortices and the phase of auditory evoked responses (AER) at different frequency domains. It has been extensively illustrated in previous studies that the phase of evoked responses align reliably in presence of a repetitive stimulus. Yet this implies that changes in the instantaneous phase over a series of stimulus presentations must also be mirrored in the laminar activity. To estimate the impact of laminar specific activity on the AER dynamics over a series of acoustic stimulation, we employ an information theoretic approach (mutual information) for quantifying the relevant information encoded in the phase of laminar LFPs and AERs.

Keywords: Neural signal processing, Neural signal processing - Blind source separation, Neural interfaces - Computational modeling and simulation
Abstract: In the CA1 region of the rat hippocampus, fast field oscillations termed sharp wave ripples have been identified as playing a crucial role in memory formation and learning. During ripple activity, particular sequences of neurons fire in a phenomena called replay. So termed because the spiking encodes patterns of past experiences, the exact role of the content of replay is an active subject of investigation in order to determines its relationship with learning and memory guided decision making. A need arises for systems that can decode replay activity during ripples in real time. This necessitates fast algorithms for both spike sorting and ripple detection with the lowest possible latency. A low latency implementation makes possible feedback experiments where decoded ripple sequences can, with minimal delay, trigger stimulating pulses that can disrupt particular kinds of decoded information before they can contribute to behavior. In this study, we optimize and implement a recently proposed online spike sorting algorithm for an increasingly popular electrophysiological software suite and measure improvements that greatly enhance its multi-tetrode decoding capabilities. Synchronizing with online ripple detection, this novel framework will allows experimenters to study the effects of disrupting replay activity with a degree of granularity hitherto unavailable.

Keywords: Neural signal processing, Neural signal processing - Time frequency analysis, Brain physiology and modeling
Abstract: Aggregate signals that reflect activities of a large number of neurons in the cerebral cortex, local field potentials (LFPs) have been observed to mediate gross functional activities of a relatively small volume of the brain tissues. There are several bands of the oscillations frequencies in LFPs that have been observed across multiple brain areas. The signature oscillation band of the LFPs in the primary motor cortex (MI) is over beta range and it has been consistently observed both in human and non-human primates around the time of visual cues and movement onsets. However, its dynamical behavior has not been well characterized. Furthermore, dynamics of beta oscillations has been documented based on the phase locking of beta oscillations, but not in terms of the inherent dynamics of the oscillations themselves. Here, we used the complexity measure derived from cluster coefficients of a recurrence network and analyzed a pair of wide-band signals, one including beta band of the LFPs and the other ranging the low gamma band in MI recorded from a non-human primate. We show rather unique temporal profiles of the evoked responses using complexity of the dynamical behavior in both bands of the oscillation, either of which is not simply resembling either the power of the oscillation or the phase locking of beta oscillations. Therefore, the current method can reveal a new type of dynamics of the underlying network complexity during the task simply simply based on

Keywords: Neural signal processing
Abstract: Linkage between behavioral states and neural activity is one of the most important challenges in neuroscience. The network activity patterns in the awake resting state and in the actively behaving state in rodents are not well understood, and a better tool for differentiating these states can provide insights on healthy brain functions and its alteration with disease. Therefore, we simultaneously recorded local field potentials (LFPs) bilaterally in motor cortex and striatum, and measured locomotion from healthy, freely behaving rats. Here we analyze spectral characteristics of the obtained signals and present an algorithm for automatic discrimination of the awake resting and the behavioral states. We used the Support Vector Machine (SVM) classifier and utilized features obtained by applying discrete wavelet transform (DWT) on LFPs, which arose as a solution with high accuracy.

Keywords: Neural signal processing, Brain physiology and modeling - Neural dynamics and computation
Abstract: Large scale CMOS-MEAs are an emerging neurotechnology enabling extracellular recordings in-vitro and in-vivo with thousand’s electrodes simultaneously. This is on the way to provide the unprecedented capability of acquiring signals from several thousands of single-units, thus opening novel perspectives for electrophysiology, but also novel challenges for analysis and management of large datasets. Here, we propose an analysis platform designed for managing unprecedentedly large datasets of electrical recordings acquired with a 4096-electrode array platform. Furthermore it provides a computational framework to facilitate the development and integration of new analysis tools exploiting high-resolution electrical recordings.

Keywords: Neural signal processing, Neurological disorders - Diagnostic and evaluation techniques, Brain-computer/machine interface
Abstract: Recent work has shown that a P300-based brain-computer interface (BCI) can provide effective long-term communication for individuals with amyotrophic lateral sclerosis (ALS). BCI users can experience significant variation in day-to-day BCI performance that can both frustrate and discourage users and caregivers alike. This study seeks to characterize this performance variation using measures of causality between electrode locations in scalp-EEG recorded from individuals with and without ALS during use of a P300-based BCI. Results show that there are statistically significant causal relationships between channels, particularly in the high beta frequency range, that are consistent across subject groups. Moreover, the connectivity patterns in the group with ALS appear to be more diffuse when compared to controls. These preliminary findings suggest that there may be differences in brain activity between individuals with and without ALS, as well as in the activity across successful and unsuccessful task sessions using a P300-based BCI. Ultimately, this information may lead more reliable BCI use for people with ALS.

Keywords: Neural signal processing, Neural signal processing - Time frequency analysis, Brain functional imaging - EEG
Abstract: Eye-Fixation Related Potentials technique is a joint analysis of both electrical brain activities and ocular movements. It allows to extract neural components synchronized with ocular fixations. However, the extracted brain responses, elicited by adjacent fixations, can be distorted by overlapping processes due to short inter fixations intervals. In this work, the ability of two algorithms, Adjar and xDawn are compared to correct these distortions. The Adjar algorithm is based on assumptions regarding the temporal distributions which become too restrictive for EFRP studies. On the hand, the xDawn algorithm is based on a more general and flexible model that is better adapted for EFRP studies.

Keywords: Neural signal processing - Time frequency analysis, Brain functional imaging - EEG, Human performance - Cognition
Abstract: Evidence from behavioral studies has suggested a capacity existed in working memory (WM). As functional connectivity in brain network has been introduced into research field of WM mechanism, the aim of this study is to investigate what happens in functional connectivity and causal flow in networked brain while WM load reaches the capacity. 32-channel electroencephalography (EEGs) was recorded from 8 healthy subjects while they performed a visual working memory task with load 1-6. Short-time Fourier transform was used to determine the principal frequency range (theta) during WM. Functional connectivity among theta components of EEGs was estimated by directed transform function (DTF). Information transform was described by causal flow. The results averaged in 10 trials for each subject show that the connectivity strength increased with load increasing from 1 to 4, peaked at load 4, and decreased after the load reached 4. The causal flow with source Fz showed the similar tendency as DTF. These findings could lead to improve understanding the capacity-related neural mechanism in WM from the view of functional connection and causal flow in the networked brain.

Keywords: Neural signal processing, Brain physiology and modeling - Neural circuits
Abstract: Working memory (WM) provides temporary information storage for performance of cognitive tasks. Neural signals in hippocampus-prefrontal cortex (HPC-PFC) circuit interact and construct a network. The question raised here is how the neural signals connect and transfer in the HPC-PFC network to perform a WM task? In this study, 32-channel local field potentials (LFPs) were recorded with two electrode arrays respectively implanted in HPC and PFC during a rat Y-maze working memory task. The principle frequency band of LFPs during the task was theta, determined via short-time Fourier transform. Functional connectivity strength was further calculated quantitatively and a causal network was defined by directed transfer function (DTF). The information transfer in the network was described by information flow. The results show that (1)the DTF curve peaked before the choice point. (2) The information flow in working memory was from HPC to PFC. These findings suggest that the functional connectivity strengthens at WM state and HPC is the WM information source in the HPC- PFC network.

Keywords: Neural signal processing
Abstract: Vagal nerve activity has been shown to be impacted in the case of Heart Failure. A direct recording of vagal nerve activity could be useful to investigate its characteristics, especially in the case of the vagal nerve stimulation therapy, which has been shown to be useful in heart failure patients. The objective of this paper is to validate the measurement of vagal nerve activity in a heart failure sheep model and to examine its characteristics and relationships with other cardiovascular parameters. Three sheep were implanted with an electrode on the right cervical vagus nerve and nervous activity were recorded together with left ventricular pressure and intraventricular electrocardiogram. Baseline periods showed fluctuations of vagal activity at the respiratory rate, correlated with fluctuations of ventricular pressure. Those fluctuations disappeared after a distal section of the nerve. Valsalva-like manoeuvres were induced in the sheep, and nervous activity was raised during the continuous positive pressure and lowered during the apnea. Another experiment on a pig also showed that the modifications of vagal activity during a Valsalva were suppressed by the application of the local anaesthesia xylocaine.

Keywords: Neural signal processing - Nonlinear analysis, Neural signal processing, Brain-computer/machine interface
Abstract: The performance of steady-state visual-evoked potential (SSVEP)-based Brain-Computer Interfaces (BCIs) have shown great improvement with multi-channel classification techniques. These methods fundamentally involve developing spatial filters that linearly combine the EEG channels so as to improve SSVEP strength and suppress noise. This paper proposes a nonlinear spatial filter using Maximum Contrastive Networks (MCNs). Essentially, MCNs are deep networks trained to maximize the contrast between signal and noise components in EEG. In other words, the network attempts to enhance the signal-to-noise ratio (SNR) of the SSVEPs in EEG. Networks of varying configurations and sigmoid functions are experimented on the EEG recordings. After random initialization, the network is pre-trained using a denoising autoencoder. Then the network is trained by back-propagation to maximize SNR. The results obtained from the MCNs are compared with the classifiers based on Minimum Energy Combination and Canonical Correlation Analysis. In this initial study, results show that MCNs significantly improve performance over the MEC and CCA based classifiers across all sessions for the trained subject. The cube-root sigmoid MCNs proved to be more accurate compared to the hyperbolic tangent MCNs. Since significantly higher accuracies were attained for lower EEG time segments, subject-specific trained MCNs with optimal configuration likely possess a large potential for online SSVEP detection.

Keywords: Neural signal processing
Abstract: The variability and complex dynamics of cell morphology make the automated segmentation of neurons in microscopic images a rather difficult task. To fully leverage modern computational power in large-scale analysis of such biological images, automation is necessary. In this paper, we present an automated approach to segmenting individual cells from their surroundings, and test it on time-lapse images of hipppocampal neurons during neurite initiation and extension. Noting that active contour based methods are usually accurate, but computationally expensive and slow, we propose a fast hybrid approach that combines Chan-Vese active contour segmentation with Bayesian thresholding for segmentation of neuron and measurement of neurite growth dynamics. Our approach demonstrated upto two-hundred-fold faster quantification of growth dynamics compared to the pure Chan-Vese segmentation.

Keywords: Neural signal processing, Neural interfaces - Microelectrode and fabrication technologies, Neural signal processing - Nonlinear analysis
Abstract: Neurons communicate in the brain via spikes. Understanding the balance between the fast-firing minority and slow-firing majority in a neuronal population is therefore a fundamental step to unravel the nature of communication and of operation within and across neuronal assemblies. Recent in vivo observations show that many functional and structural parameters of the brain follow a skewed nature and typically manifest lognormal distributions of patterns. Here, we show for the first time that high-density microelectrode array (HD-MEA) reveal such a lognormal-like distribution of the firing patterns also in in vitro grown hippocampal neuronal networks, and already after 10 minutes of recording. Additionally, we demonstrate that the electrode density plays a key role for obtaining such a distribution in cultures. Overall, our findings show that in vitro neural networks recorded with CMOS-MEAs might contribute in investigating the organization and function of neuronal networks by revealing, with similarities with results obtained in vivo, the relationships among different skewed distributions at multiple scales, i.e. from synapses, single cells and micro-circuits up to large-scale networks.

Keywords: Neural signal processing, Neural signal processing - Nonlinear analysis, Human performance - Sensory-motor
Abstract: Dimension reduction is critical in identifying a small set of discriminative features that are predictive of behavior or cognition from high-dimensional neuroimaging data, such as EEG and fMRI. In the present study, we proposed a novel nonlinear supervised dimension reduction technique, named PCA-SIR (Principal Component Analysis and Sliced Inverse Regression), for analyzing high-dimensional EEG time-course data. Compared with conventional dimension reduction methods used for EEG, such as PCA and partial least-squares (PLS), the PCA-SIR method can make use of nonlinear relationship between class labels (i.e., behavioral or cognitive parameters) and predictors (i.e., EEG samples) to achieve the effective dimension reduction (e.d.r.) directions. We applied the new PCA-SIR method to predict the subjective pain perception (at a level ranging from 0 to 10) from single-trial laser-evoked EEG time courses. Experimental results on 96 subjects showed that reduced features by PCA-SIR can lead to significantly higher prediction accuracy than those by PCA and PLS. Therefore, PCA-SIR could be a promising supervised dimension reduction technique for multivariate pattern analysis of high-dimensional neuroimaging data.

Keywords: Neural signal processing, Neural signal processing - Blind source separation, Neural signal processing - Time frequency analysis
Abstract: Scalp encephalograms (EEG) are often contaminated by various types of biological and non-biological noise that affect the performance of source localization, signal decoding and/or event estimation methods. The statistics and structure of EEG noise are usually unknown and time-varying. As these characteristics may vary substantially between subjects, as well as within subjects both in time and space, it may be difficult to select a unique noise model and/or to update its parameters. This study proposes a data-driven approach, based on the Empirical Model Decomposition and the auto-correlation function, for estimating and comparing the spatio-temporal statistical characteristics of EEG noise. The proposed approach was applied to a dataset of continuously recorded (over several hours) scalp EEGs from 3 patients with focal epilepsy. It is shown that the statistical characteristics of EEG noise vary substantially in time and space. Thus, adaptive signal processing methods may be most appropriate for denoising and/or estimation of data-driven and possibly subject-specific noise distributions, with the ultimate goal to improve source localization, waveform feature extraction and signal decoding.

Keywords: Neural signal processing
Abstract: Automated diagnosis of skin cancer is an active area of research with different classification methods proposed so far. However, classification models based on insufficient labeled training data can badly influence the diagnosis process if there is no self-advising and semi supervising capability in the model. This paper presents a semi supervised, self-advised learning model for automated recognition of melanoma using dermoscopic images. A Deep architecture is constructed using labeled data together with abundant unlabeled data, with the fine tuning using an exponential loss function to maximize separation of labeled data. In parallel a self-advised SVM algorithm is used to enhance classification results by counteracting the effect of misclassified data. To increase generalization capability and redundancy of the model, polynomial and radial basis function based SA-SVMs and Deep network are trained using randomly chosen training samples via a bootstrap technique. Then the results are aggregated using LSE weighting. The proposed model is tested on a collection of 100 dermoscopic images. The variation in classification error is analyzed with respect to the ratio of labeled and unlabeled data used in the training phase. The classification performance is compared with some popular classification methods and proposed model outperforms most of the popular techniques including KNN, ANN, SVM and semi supervised algorithms like Expectation maximization and transductive SVM.

Keywords: Neural signal processing
Abstract: In this paper, we investigate the workings of threshold (TH) based spike detection for neuronal extracellular field potential spikes. Thresholding is the most used spike detection method. In general, it is employed by setting the TH as per convention and without considering either the undetected or spurious spikes. In this paper, we provide insight in to the workings of thresholding, and proposed a new objective way to set the TH based on spike count histogram analysis. We illustrate the method with 2D and 3D simulations and analysis of measured data.

Keywords: Clinical neurophysiology, Neurological disorders - Epilepsy, Neurological disorders
Abstract: Scalp recordings of cortical activations, electroencephalography (EEG), are used for a variety of clinical and research purposes. EEG is commonly used clinically to detect diseases or injuries to the underlying cortical physiology. Unfortunately, the EEG signal is also artifact prone and these artifacts can exhibit a similar temporal and spectral profile as that caused by the potential disease. We have created a model of the EEG and artifacts to explore their interplay and the theoretical limits of detection when artifacts may not be separable from clinical events of interest.

Keywords: Neural signal processing
Abstract: Many scientists have focused their attention on the interplay between functional-effective connections and neuronal dynamics at different level of complexity (from single neurons to brain area) and on different experimental models (from simple in-vitro networks to the whole brain). In this work, we focus on an in-vitro reductionist model constituted by neuronal cultures coupled to Micro-Electrode Arrays (MEAs). At present, there are no specific and dedicated software to analyze the connectivity of this experimental model. Thus, we designed and developed a new toolbox in order to reliably estimate functional links of such in-vitro neuronal networks coupled to MEAs. This toolbox, named ToolConnect, implements two correlation and three information theory based methods (i.e. cross and partial correlation, transfer entropy, mutual information and joint entropy). We tested the implemented toolbox on in silico neuronal networks based on realistic computational model, using specific validation procedures based on the Receiver Operating Characteristics Curve (ROC curve) and the accuracy.

Keywords: Brain-computer/machine interface, Brain-computer/machine interface - Biofeedback, Brain-computer/machine interface - Robotics applications
Abstract: This study aimed to improve the detection performance of movement related cortical potentials (MRCP) in EEG for brain-computer interface (BCI) applications. For this purpose, we apply an outlier-resisting manifold learning method to reduce false detection. Experimental data were collected as described in [1] from 12 healthy subjects. The subjects performed brisk ankle dorsiflexions in a cue-based paradigm in 4 conditions combining two speeds (slow and fast, S and F) and two force levels (low and moderate, 20% and 60% of maximum force). The 4 classes are denoted by S20, S60, F20 and F60, respectively. The continuous EEG data were segmented into two classes: 1) signal intervals that contain MRCPs during motor imagination; and 2) noise signal intervals where no motor task was performed. We extract robust locality preserving features of these two classes via an outlier-resisting manifold learning method, called L1-norm based Locality Preserving Projection (LPP-L1) [2]. A nearest-neighbor classifier was then applied to these features. The detection performance was evaluated on the experimental data and compared with the previously proposed Locality Preserving Projection (LPP) method [3]. A 5-fold cross-validation was used when analyzing the detection performance of the two algorithms.

Keywords: Brain-computer/machine interface, Brain-computer/machine interface - Biofeedback, Brain-computer/machine interface - Robotics applications
Abstract: The purpose of this study was to improve the detection performance of movement related cortical potentials (MRCP) in EEG for brain-computer interface (BCI) applications and make it efficient both in processing speed and memory requirement, so it would be more suitable for ambulatory applications. For this purpose, we apply a method called Spectral Regression based Locality Preserving Projection (SRLPP) [1]. Experimental data were collected as described in [2] with 12 healthy subjects. The subjects performed brisk ankle dorsiflexions in a cue-based paradigm in 4 conditions combining two speeds (slow and fast, S and F) and two force levels (low and moderate, 20% and 60% of maximum force). The 4 classes are denoted by S20, S60, F20 and F60, respectively. The continuous EEG data were segmented into two classes: 1) signal intervals that contain MRCPs during motor imagination; and 2) noise signal intervals where no motor task was performed. We extract locality preserving features of these two classes via SRLPP. A nearest-neighbor classifier was then applied to these features. The detection performance was evaluated on the experimental data and compared with the previously proposed Locality Preserving Projection (LPP) method [3]. A 5-fold cross-validation was used when analyzing the detection performance of the two algorithms.

Keywords: Clinical neurophysiology, Human performance - Cognition
Abstract: Mismatch negativity (MMN) is an event related brain potential based on the automatic detection of a deviant stimulus in a sequence of standard stimuli. It may then be used to quantify the aptitude to discriminate two sounds. Are found discrimination thresholds with MMN comparable to those found in psychoacoustic? It has been showed for frequency or intensity discrimination (gold standards of MMN paradigm), but less frequently for frequency and amplitude modulation. Looking for those discrimination thresholds in MMN was the occasion to optimize the data collection protocol and develop automatic threshold detection on a population with event related brain potential.

Keywords: Neurological disorders - Diagnostic and evaluation techniques
Abstract: The cerebral oxygenation is one of the most important parameters in determining tissue impairment and guiding treatment in patients treated for severe brain injury. Two different algorithms for optical estimation of oxygen saturation have been investigated as a step toward setting up a bedside system for monitoring of local brain oxygenation. An oxygen sensitive Clark electrode was used as reference during the experiments to calibrate the probe based system. Optical phantoms mimicking brain tissue were prepared by mixing various concentrations of hemoglobin and Intralipid in saline. The experiments resulted in two calibration models where each predictor were related to the expected oxygen saturation through the hemoglobin dissociation curve.

Keywords: Neural signal processing, Neural interfaces - Neural microsystems and interface engineering, Neural interfaces - Neural stimulation
Abstract: This paper investigates the simultaneous use of two separate multicontact cuff (MCC) electrodes for selective recordings in electroneurography (ENG) and peripheral nerve selective stimulation.Multicontact VNS yields the appealing idea that one could perform selective VNS by using adapted spatio-temporal patterns and thus minimize side effects. In heart-related applications, the sensitivity of VNS parameters is often studied based on physiological modifications like heart rate (bradycardia) or blood pressure. This study aims at completing these observations by investigating simultaneously electrocardiography (ECG) and ENG activity. ENG activity is quantified by the analysis of compound action potentials that are evoked by the different VNS configurations whereas ECG analysis is done by studying heart rate variability during VNS. It is shown that, for a given temporal pattern, certain VNS spatial configurations have a significant effect on both heart and neural activities. The relations between both effects are studied in this paper and it allows bringing some insights in the source location of heart fibers from measurements at the nerve periphery. We demonstrate that MCC electrodes are well indicated to recruit specific fibers in a topographical way and also to record regional fiber activity. These results entail the perspective of using ENG signals as control signals in VNS medical applications in order to limit side-effects and use ENG as additional input for VNS close-loop.

Keywords: Brain-computer/machine interface, Brain-computer/machine interface - Robotics applications, Brain functional imaging - EEG
Abstract: This paper reports our results on comparing the performance of a movement artifact removal algorithm between wireless and wired EEG data acquisition systems. We adopted the g.Nautilus, 32-channel wireless acquisition system with a sampling rate of 250 Hz, and the Biosemi Active Two system, 32-channel with a sampling rate of 2048 Hz. We analyzed the EEG spectral power of the two systems by using MATLAB after applying the artifact removal algorithm called Automatic Subspace Reconstruction (ASR). One healthy volunteer with no history of neurological or physical deficits performed three head movement tasks (swinging its head up and down, swinging left and right parallel to the ground, swinging left and right vertical to the ground). The experiment protocol for each performance was as follows: rest (20s), action (20s), and rest (5s). To evaluate the difference between wireless and wired EEG systems, we compared each tasks’ power spectral density of the filtered EEG signals in the 0-30 Hz frequency range. The results show the power spectral density of the Cz electrode according to the 10-20 system and the red lines indicate the event time when the movements were performed. The wireless system shows more contrast between the rest phase and the action phase. Our results indicates the ASR algorithm performed better with the wireless EEG system on separating active and rest phases in movement related tasks compared to the wired EEG system in spite of the lower sampling rate.

Keywords: Neural signal processing
Abstract: Sound encoding is achieved by the auditory transducers of the cochlea, the inner hair cells (IHCs), which are innervated by the auditory nerve fibers (ANFs) that convey auditory information to brainstem. Masking occurs when a first stimulation (the masker) is follow by a second one (the probe). The masker is responsible of a reduction in the probe-firing rate. The interval needed to the ANF before being able to encode properly the probe is called the post-stimulation recovery time. Despite the fact that recovery is crucial for the temporal coding of sounds, little is known about the physiological mechanisms underlying the auditory nerve recovery. In this study, we developed an electrophysiological approach to address this question. Results show that a prior stimulation (forward masking) affects preferentially the temporal precision of the first spike latency. This finding suggests that the mechanism underlying the auditory fiber recovery is located at the synapse between the IHC and afferent fiber. The forward masking could then disturb the neurotransmitter release or the neurotransmitter reuptake by supporting cells. Whatever the mechanism, this experimental approach provides an interesting tool to examine the auditory nerve recovery.

Keywords: Neural signal processing
Abstract: The auditory system operates over a vast range of sound pressure levels (100-120 dB) and is nearly constant in accurate (~1 dB) across almost the entire range. To explain this sound coding dynamic, recent works have shown that auditory nerve fibers could adapt their dynamic to the sound level statistics. The aim of this study was to find out if this adaptation mechanism (initially discovered using single-fiber recordings) can be observed with compound action potential (CAP) of the auditory nerve. This technique, which is less invasive than single-fiber recording and potentially applicable to humans, provides a promising tool to track the mechanism underlying the wide sound coding dynamic of our auditory system.

Keywords: Neural signal processing
Abstract: This study proposes an automatic thresholding technique to detect eye blink artifacts contaminating EEG signals. The proposed method for the automatic detection of eye blink artifacts does not require any labeled training data, and thus the system can be more readily applied to practical applications including wearable EEG systems. The performance of the proposed method was compared with those of conventional automatic thresholding algorithms.

Keywords: Clinical neurophysiology, Sensory neuroprostheses - Auditory, Neural signal processing
Abstract: Auditory evoked potentials (AEPs) are electrical brain potentials generated automatically in response to an auditory stimulation. They can be successfully used to measure objectively both auditory detection and discrimination of normal-hearing (NH) people and can be applied to cochlear-implanted (CI) patients to evaluate the CI efficiency. However the CI produces a large electrical artifact which contaminates the AEPs. Thus their estimation can help develop efficient processing methods which can then be applied for clinical use. Here we propose a new setup of acquisition which allows recording simultaneously both the AEPs of a NH subject and the CI artifact from a phantom. Both datasets can be mixed providing realistic EEG data of an artificial CI patient. Then several signal processing methods of AEPs-CI artifact extraction can be tested. The proposed methodology, which was tested and validated for ICA, appears to be useful to develop and evaluate new signal processing methods prior to their clinical application in CI patients.

Keywords: Neural signal processing, Neural signal processing - Blind source separation, Neural interfaces - Implantable systems
Abstract: Current source density (CSD) analysis is the method used to derive the current sources and sinks that generate local field potentials (LFP). The CSD methods require multichannel electrodes to record simultaneously the evoked LFP on multiple locations within the brain structure. In practice, these multichannel recordings are sensitive to measurement errors. In this study, we introduce a finite element method (FEM) model which can be used to study different CSD methods and asses the sensitivity for different types of measurement errors. Within a volume conductor in the FEM model a CSD distribution is formed by 2 Gaussian sinks and 1 source. The potential generated by the CSD is measured on a 5x16 measurement grid, representing 5 subsequent LFP recordings with a 16 channel electrode. From the potentials the CSD is estimated with the inverse and kernel CSD method. By calculating the error in the CSD estimation we studied the effect of 2 experimental errors. i.e. a random displacement of the 16 channel electrode and a random deviation of the signal gain per channel. The results showed the kCSD method is affected significantly less by channel gain deviations then the iCSD method. However, both CSD methods are affected similar by the random displacement of the electrode. We conclude that using a FEM model to describe different experimental errors seems to be an intuitive way to study the influence on the estimated CSD.

Keywords: Motor neuroprostheses, Brain-computer/machine interface, Neural signal processing
Abstract: Spinal cord injury (SCI) disconnects supraspinal and spinal circuits, leading to paralysis. Neuronal networks embedded in lumbosacral segments retain the capacity to generate complex motor behaviors, while the cortical network retains the capacity to control prosthetic devices. In order to restore voluntary locomotor functions and promote neural repair, the development of corticospinal neuroprostheses connecting real-time decoded cortical drive to spinal electrical intervention has been proposed. In this work we investigate the locomotor correlates found in cortical activity during bipedal treadmill stepping in a clinically relevant model of severe thoracic contusion SCI in rats. We also address the challenge of online detecting feedback information viable for a closed-loop functional electrical stimulation system. Results show that hind-limb kinematics are encoded in the sensorimotor cortex of rats even after a severe contusion SCI. Moreover, gait events can be decoded online from cortical activity. Extracted information can be exploited to drive highly-selective spinal neuromodulation therapies resulting in an improved spatiotemporal activation of lumbosacral circuitries. These results are directed towards the goal of providing to patients with SCI an artificial interface between the brain and the sublesional spinal cord, making voluntary control of locomotion possible after paralysis.

Keywords: Neural signal processing
Abstract: Speech intelligibility in quiet is critically dependent on the temporal envelope coding of sounds. The temporal envelope was extensively investigated with psychoacoustic methods in human but little is known about the underlying mechanisms of envelope coding in the cochlea. The aim of this study was to develop a new electrophysiological method able to measure the auditory nerve response to amplitude-modulation stimuli. This method is clinically important to detect auditory neuropathies characterized by timing-related perception deficits.

Keywords: Neural interfaces - Computational modeling and simulation
Abstract:

Keywords: Brain functional imaging - EEG, Neural signal processing - Time frequency analysis
Abstract: Electroencephalography (EEG) studies have largely reported the P300 Event Related Potential (ERP) elicited by target stimulus processing compared to non-target stimulus. These studies used constrain experimental paradigms during which participants did not move their eyes. However, during more ecological paradigms where participants can move their eyes, the P300 potential might be more difficult to identify due to the overlapped potentials elicited by consecutive ocular fixations. In this study, we use the xDawn algorithm for denoising the overlapped Eye-Fixation Related Potentials (EFRP), and we observe the P300 potential elicited on the first, but also on the consecutive subsequent fixations landed on the target stimuli.

Keywords: Brain functional imaging - EEG, Human performance - Cognition, Neural signal processing
Abstract: Training is a process to improve one's capacity or performance through the acquisition of knowledge or skills specific for the task. Although behavioral performance would be improved monotonically and reach a plateau as the learning progresses, neurophysiological process shows different pattern like an inverted U-shaped curve. One possible account for the phenomenon is that the brain first works hard to learn how to use specific task-relevant areas, followed by improvement of efficiency derived from disuse of irrelevant brain areas for good task performance. In this study, we employed the functional connectome approach to study the changes in global and local information transfer efficiency of the functional connectivity induced by training of a piloting task. Our results have demonstrated that global information transfer efficiency of the network, revealed by normalized characteristic path length in beta band, once decreased and then increased during the training sessions. We show that graph theoretical network metrics can be used as biomarkers for quantifying the degree of training progresses, in terms of efficiency, which can be differed based on cognitive proficiency of the brain.

Keywords: Brain functional imaging, Neurological disorders - Diagnostic and evaluation techniques, Clinical neurophysiology
Abstract: Hypometabolism has been observed in the contralesional cerebellum hemisphere after various cortical lesions, and is referred as crossed cerebellar diaschisis. It is unknown whether it can be measured after awake surgery for brain diffuse low-grade glioma. Using resting state fMRI we suggest that crossed cerebellar diaschisis can be assessed postoperatively.

Keywords: Brain functional imaging - EEG, Neural signal processing - Time frequency analysis, Human performance - Cognition
Abstract: EEG-based emotion recognition has received increasing attention in advanced human-computer interaction, where the choice of independent variables to discriminate emotions from the frequency range of EEG and electrode locations is not very self-evident, thus this work tried to find the correlation between the emotional states and both EEG frequency ranges and EEG channels. 12 healthy volunteers were emotionally elicited by movie clips to experience five basic emotional states of neutral, happy, sad, tense and disgust states. Fisher discriminant ratio (FDR) was employed to find the discriminative bands and electrodes with statistical differences. Finally, a support vector machine (SVM) with 5-fold cross validation was performed. Average recognition rates have achieved 93.31% and 85.39% for two feature sets.

Keywords: Brain functional imaging - NIR, Transcranial direct current and magnetic stimulation (TCDS/TMS), Neurological disorders - Stroke
Abstract: Cerebral vascular status can be evaluated with cerebrovascular reactivity (CVR) that reflects the capacity of blood vessels to dilate, and is an important marker for brain vascular reserve. Here, transcranial direct current stimulation (tDCS) can up- and down- regulate cortical excitability depending on current direction, and anodal tDCS can increase regional cerebral blood flow during stimulation. Impairments in CVR have been associated with increased risk of ischemic events. Here, near-infrared spectroscopy (NIRS) is a cerebral monitoring method that can be used for non-invasive and continuous measurement of cerebral vascular status under various clinical conditions. This paper describes the development of a 4-channel continuous wave NIRS combined with tDCS in an FPGA-based hardware that captured the hemodynamic changes in the frontal cortex of the brain, as a measure of CVR, before and after anodal tDCS. We recruited 14 patients with established and acute ischemic stroke (<1 month) localized to a single hemisphere (10 male and 4 females from age 42 to 73). The affected hemisphere with impaired circulation showed significantly less (0.26 +/- 0.28), p<0.01, change in cerebral hemoglobin oxygenation than the healthy side (3.43+/- 0.86) in response to anodal tDCS. Thus, combining NIRS with tDCS can lend to low-cost point of care testing of cerebral vascular status so we present a NIRS-tDCS based adaptive neuro-fuzzy inference system implemented in a FPGA-based hardware.

Keywords: Brain functional imaging
Abstract: Several studies in the last years have explored the potential of using a frequency domain approach in the analysis of functional brain connectivity. However, few studies have integrated the advantages of such approach with the brain network model given by graph theory. Furthermore, the use of such methods has rarely been explored with fixed frequency stimulation resting-state protocols. In this paper, we propose a method to model functional connectivity using both a frequency domain approach and a network model. By estimating coherence with the Welch method, it is possible to represent the brain activity using binary connectivity matrices, and to characterize such networks in terms of global and local measures. We tested our method with experimental data from an autonomic stimulation protocol and compared the results with the more common correlation analysis. We were able to see significant differences between the two sessions in the frequency range of the stimulation. Such differences involved brain areas associated with the central autonomic network and didn’t show up in the time domain approach.

Keywords: Brain functional imaging, Human performance - Cognition
Abstract: This paper describes an fMRI study with 76 middle-aged and older subjects using the Wisconsin Card Sorting Task, widely utilized to investigate executive function and frontal lobe dysfunction. Several cognitive processes involved in the task are correlated with the cognitive decline observed due to aging. With functional imaging we demonstrate that subjects with a better performance show an increased activation in the frontal lobe and cingulate cortex, which are important regions for the processes involved in the task.

Keywords: Brain functional imaging - EEG, Brain-computer/machine interface, Human performance - Cognition
Abstract: Development of brain-computer interfaces interacting with cognitive functions is a hot topic in neural engineering since it may lead to innovative and powerful diagnosis, rehabilitation, and training methods. This paper addresses the problem of measuring sustained visual attention using electroencephalography and presents an experiment inspired by continuous performance tasks used in neuropsychology along with the classification results obtained when trying to discriminate between low and high attention states. Following a leave-one-subject-out validation approach, 76% accuracy was obtained when discriminating thirty second epochs and 69% accuracy using five second epochs.

Keywords: Brain functional imaging - EEG, Brain-computer/machine interface, Neural signal processing
Abstract: Brain computer interface (BCI) is a direct communication pathway between the human central nervous system and external devices primarily aiming at restoring damaged functions such as sight, hearing and movement. Although great achievements have been made for the development of reliable BCI systems to assist people with upper-limb disabilities, researches on BCI development related to lower-limb are still rudimentary. In the current study, based on the regularized common spatial pattern analysis (R-CSP) method and statistical dependency, we have developed an improved feature selection method for lower-limb movement pattern classification. High-resolution electroencephalogram (EEG) signals were recorded from four healthy male subjects undergoing real lower-limb movements. Compared to the conventional CSP, R-CSP, and PCA methods, the proposed method achieved the best average accuracy of 83.5% for single trial classification of left and right lower-limb movement. Our findings thereby have insightful implications for developing practical BCI systems for lower-limb movement.

Keywords: Brain functional imaging - EEG, Neural signal processing
Abstract: A challenging issue in cognition is how to precisely identify brain networks at very short temporal scales. So far, very few studies have addressed this problem as it requires high temporal and spatial resolution simultaneously. The recent past years have seen a noticeable increase of interest for electroencephalography (EEG) to analyze functional connectivity through brain sources reconstructed from scalp signals. Here, we performed a novel study based on EEG source connectivity to identify large scale networks with high temporal and spatial resolution. We show clear evidence of the ability of EEG source connectivity to identify brain networks with high time/space resolution during the visual processing period of picture naming task. Our qualitative and quantitative observations show that the identified brain networks are in accordance with fMRI-based results reported in the literature regarding involved brain areas.

Keywords: Brain functional imaging - EEG, Neural signal processing
Abstract: Due to its excellent temporal resolution, the Electroencephalogram (EEG) has become a key neuroimaging technique to analyze functional brain networks at scalp level (electrodes) and at reconstructed sources (inverse problem). However, a tool that can analyze EEG recordings, from raw signals 2D/3D brain networks is still missing. Here, we propose a MATLAB-based pipeline, called EEGNET, to process EEG signals and represent corresponding functional brain networks. It includes: 1) Preprocessing of the EEG signals, 2) Solving the inverse problem / reconstructing the cortical sources, 3) Computing the functional connectivity at source level, 4) Calculating the network measures and 5) Visualizing 2D and 3D brain networks. The software is carefully designed to enclose these different phases. The first version of EEGNET is easy to use, flexible and user friendly. Moreover EEGNET is an open source tool and can be freely downloaded from the internet soon.

Keywords: Brain functional imaging - EEG, Neural signal processing
Abstract: measuring similarity among graphs is a challenging issue in many disciplines including neuroscience. Several algorithms, mainly based on vertices or edges properties, were proposed to address this issue. Most of them ignore the physical location of the vertices, which is a crucial factor in the analysis of brain networks. Indeed, functional brain networks are usually represented as graphs composed of vertices (brain regions) connected by edges (functional connectivity). In this paper, we propose a novel algorithm to measure a similarity between graphs. The novelty of our approach is to account for vertices, edges and spatiality at the same time. The proposed algorithm is evaluated using synthetic graphs. It shows high ability to detect and measure similarity between graphs. An application to real functional brain networks is then described. The algorithm allows for quantification of the intersubjects variability during a picture naming task.

Keywords: Brain functional imaging - EEG, Neurological disorders - Diagnostic and evaluation techniques, Neuromuscular systems - Locomotion, posture and balance
Abstract: Freezing of gait (FOG) is a common disabling gait disorder in late stage Parkinson’s disease (PD), which can lead to falls and loss of independence. To date, the mechanisms causing FOG are still unknown and no treatment has proven to be effective. In this study, sixteen PD participants with and without clinically confirmed FOG symptoms were recruited, referred to as FOG+ and FOG-, respectively. All participants navigated through a customized virtual reality (VR) corridor by stepping in place (SIP) while electroencephalography (EEG) data was recorded. Additionally, a cognitive, visual two-stimulus oddball response task was conducted, which was repeated for a seated condition. The VR environment proved to be a reliable tool to elicit FOG like symptoms in a clinical test environment. EEG and behavioral results were compared between groups and conditions for qualitative effects. In the seated condition, the FOG+ group showed similar behavioral performance to the FOG- group, however, in the SIP condition the FOG+ group showed significantly decreased performance with longer reaction times and more target misses. Analysis of the EEG data revealed consistent visual responses to the stimuli, but an absence of the P3b component in stimulus-locked brain responses for FOG+ participants. However, if data is response-locked, the P3b component is clearly visible, supporting the theory that components related to decision making and motor preparation are present, but with variable delays.

Keywords: Brain functional imaging
Abstract: Comparing serially acquired fMRI scans is a typical way to detect functional brain changes in different conditions. However, this approach introduces additional variation on physical and physiological conditions, which results in substantial noise. To improve sensitivity and accuracy of signal detection in such highly noisy fMRI data, potentially important information should be incorporated. Here we propose a new significance indicator, the critical regularization value (CR-value), which detects significantly changed voxels by taking both the magnitude of the voxel-wise signal variation and spatial smoothness into account. The CR-value allows voxels that survive in a stronger sparse constraint to be considered as more significant. We demonstrate our method using a simulation dataset and a real fMRI dataset collected from the previous study. The results show that CR-value more accurately detects the true activation than GLM P-value, Posterior Probability Maps (PPM) and the Threshold Free Cluster Enhancement (TFCE) in noisy datasets.

Keywords: Brain functional imaging - EEG, Brain functional imaging - Source localization, Brain physiology and modeling - Neuron modeling and simulation
Abstract: In order to reconstruct the electrical brain activities, both the inverse and the forward problems have to be solved. The inverse problem consists of finding brain activities (sources) which are responsible for the measured signals. The forward problem is calculating the potentials at the sensor electrodes placed on the scalp from a given source inside the brain. Its resolution involves models of head and models of sources. The finite-element method (FEM) is investigated in this paper for its ability to model realistic geometry and physical proprieties. However, using this method, the point source introduces singularity in the calculation and impacts the accuracy. Different techniques have been developed to avoid this singularity but the accuracy of the most techniques decrease as the current dipole approaches the interface of different tissues with different conductivities. In this paper, the forward problem is solved by the FEM using a spherical head model and the saint Venant’s principle is used to avoid the source’s singularity. Near the interface of different brain tissues, a modification on the saint Venant’s method is introduced. Simulation results show that this new modification gives good accuracy as compared with the traditional approaches.

Keywords: Brain functional imaging - EEG, Neural signal processing, Sensory neuroprostheses - Signal and vision processing
Abstract: The purpose of this study was to explore the effects of stereoscopic 3D (S3D) display technology on event-related brain potentials (ERPs). A sample of thirty-four healthy participants was subjected to an oddball paradigm after being exposed to stereoscopic 3D contents with passive polarized display or traditional 2D display. The participants were randomly assigned to two groups―2D group and S3D group; in such a way that their intelligence ability and age were controlled between the groups. The behavioral and ERP results did not show any significant differences between S3D and 2D groups for either ERP components (amplitude and latency) or accuracy and response time of the target detection. These results suggest that passive polarized S3D display technology may not induce any effects (cognitive or visual fatigue) which may disturb the ERP components.

Keywords: Brain functional imaging - EEG, Clinical neurophysiology, Human performance - Cognition
Abstract: As the brain ages, it suffers several neurochemical, structural and functional changes. These deficits are primarily reflected on daily memory tasks. In the present work, we use the Wisconsin Card Sorting Task (WCST) task to assess P300 component of the event-related potential (ERP) as a marker for aging. Considering age-related effects, WCST was applied to young, middle-aged and elder participants. Early-late trial analysis tested the attentional set shifting and working memory updating hypothesis for the mechanisms behind WCST. The results suggest that, as people age, P3b peak latency highly correlates with age on both early and late trials, increasing as people gets older. P3b peak amplitude significantly distinguished between early and late trials regardless the subjects’ age, while there were no differences on P300 peak latency.

Keywords: Brain functional imaging - EEG, Brain physiology and modeling - Neural dynamics and computation
Abstract: This paper studies the effect of subdermal EEG lead placement on measurement sensitivity distributions, and compares them with the sensitivity distributions obtained using surface EEG leads. A five-layered isotropic head model was constructed based on magnetic resonance imaging (MRI) data. The surface electrodes were placed on the scalp of the model according to the traditional 10-20 EEG system. The subdermal electrodes were arranged in 5 × 5 grids and placed on the skull in seven reference locations: Fz, Cz, Oz, T3, T4, P3, and P4. The effects on the measurement sensitivity were studied by means of the half-sensitivity volume (HSV). For the surface measurements, the size of the HSV varies around 1 cm3, while the subdermal leads can concentrate the measurement in regions ten times smaller. The results indicate that the EEG measurement may benefit from subdermal implantation since the subdermal measurements are more accurate and specific than the surface measurements. Nevertheless, the improvement was registered only for the subdermal grids centred on Cz, T3 and T4 locations. This suggests that the subdermal electrode performance highly depends on the thickness of the underlying matter, such as the skull and cerebrospinal fluid (CSF).

Keywords: Brain functional imaging - EEG, Human performance - Modelling and prediction, Human performance - Cognition
Abstract: Electroencephalography (EEG) has been widely adopted for investigating brain behavior in different cognitive tasks e.g. learning and memory. In this paper, we propose a pattern recognition system for discriminating the true and false memories in case of short-term memory (STM) for 3D and 2D educational contents by analyzing EEG signals. The EEG signals are converted to scalp-maps (topomaps) and city-block distance is applied to reduce the redundancy and select the most discriminative topomaps. Finally, statistical features are extracted from selected topomaps and passed to Support Vector Machine (SVM) to predict brain states corresponding to true and false memories. A sample of thirty four healthy subjects participated in the experiments, which consist of two tasks: learning and memory recall. In the learning task, half of the participants watched 2D educational contents and half of them watched the same contents in 3D mode. After 30 minutes of retention, they were asked to perform memory recall task, in which EEG signals were recorded. The classification accuracy of 97.5% was achieved for 3D as compared to 96.5% for 2D. The statistical analysis of the results suggest that there is no significant difference between 2D and 3D educational contents on STM in terms of true and false memory assessment.

Keywords: Brain functional imaging - EEG, Sensory neuroprostheses - Auditory, Human performance - Cognition
Abstract: In complex listening environments, people with hearing loss have an increased listening effort to achieve a similar speech understanding level as normal hearing people. However, a standardized method to estimate this listening effort does not exist. Recently, we have shown a possible way to determine listening effort objectively. This method is based on the phase distribution of the ongoing oscillatory EEG activity. The aim of the current study was to assess, whether such objective methods can also be used to extract dynamical listening effort profiles. Hearing aids were fitted using a directional microphone (DM) configuration, a new binaurally coupled hearing aid technique (BHA) and a fitting using omnidirectional microphones (ODM). Furthermore, a temporal resolution of the measure was obtained to unveil underlying processes such as fatigue effects or a cessation to spend attentional effort. The results indicate that time-varying listening effort profiles related to effects like fatigue or a cessation to solve the auditory task can be detected by the proposed objective measures.

Keywords: Brain functional imaging, Human performance - Sensory-motor, Neural interfaces - Sensors and body interfaces
Abstract: Functional neuroimaging plays a key role in addressing open questions in systems and motor neuroscience directly applicable to brain machine interfaces. Building on our low-cost motion capture technology (fMOVE), we developed f2MOVE, an fMRI-compatible system for 6DOF goal-directed hand and wrist movements of human subjects enabling closed-loop sensorimotor haptic experiments with simultaneous neuroimaging. f2MOVE uses a high-zoom lens high frame rate camera and a motion tracking algorithm that tracks in real-time the position of special markers attached to a hand-held object in a novel customized haptic interface. The system operates with high update rate (120~Hz) and sufficiently low time delays (<20~ms) to enable visual feedback while complex, goal-oriented movements are recorded. We present here both the accuracy of our motion tracking against a reference signal and the efficacy of the system to evoke motor control specific brain activations in healthy subjects. Our technology and approach thus support the real-time, closed-loop study of the neural foundations of complex haptic motor tasks using neuroimaging.

Keywords: Brain functional imaging - EEG, Neural signal processing - Blind source separation, Brain-computer/machine interface
Abstract: Electroencephalogram (EEG) is susceptible to various non-neural physiological artifacts. Automatic artifact removal from EEG remains a great challenge for extracting relevant information from brain activities. In order to adapt to variable subjects and EEG acquisition environments, this paper presents a novel automatic artifact removal method based on priori artifact information. First, the wavelet-ICA algorithm, which combines of discrete wavelet transform (DWT) and independent component analysis (ICA), is utilized to separate artifact components. Then the artifact components are automatically identified using the priori artifact information, which is acquired in advance. Subsequently, signal reconstruction is performed without the identified artifact components to obtain the artifact free signals. At last, the method is validated by the improvements of the classification accuracies in a motor imagery experiment.

Keywords: Brain functional imaging - EEG, Brain-computer/machine interface, Neurological disorders - Stroke
Abstract: Rehabilitation of motor impairment after stroke has an important medical value. However, the present stroke rehabilitation mainly are passive, its efficacy is limited. The BCI-FES system focused on the limit of passive stroke rehabilitation, combined with motor imagery (MI) and functional electrical stimulation (FES), through the pattern recognition of motor imagery mode of patients with EEG signal, generated FES signal. The BCI-FES system achieved the active rehabilitation of patients with mind-control, stimulated brain plasticity and improved rehabilitation efficacy. In this study, the BCI-FES system and clinical rehabilitation evaluation of post-stroke hemiplegia patients was investigated, and the rehabilitation efficacy showed, affected motor related cortex of patient subject was activated significantly, and motor function was further enhanced, by brain plasticity guidance of BCI-FES.

Keywords: Brain functional imaging, Neural signal processing, Neural signal processing - Time frequency analysis
Abstract: Magnetoencephalography (MEG) is used for functional connectivity analysis, and can record brain signals from deep sources non-invasively. Modern MEG systems measure signals at a temporal resolution of milliseconds and at millimeter precision. However, there is a lack of standardization in the position and orientation of sensors, unlike the electroencephalography (EEG) that follows sensor positioning guidelines defined by international 10-20 10-10 or 10-5 systems. Mapping MEG sensor positioning to EEG’s is essential to enable data fusion and comparison of both modalities. This paper reports the development of a novel framework for MEG data visualization and analysis. The strength of the proposed framework is demonstrated through inputting sizable data from multiple healthy subjects and generating default mode connectivity visualization from the most common and significantly active coherent brain regions.