Toward an Adapted Neurofeedback for Post-stroke Motor Rehabilitation: State of the Art and Perspectives.

Stroke is a severe health issue, and motor recovery after stroke remains an important challenge in the rehabilitation field. Neurofeedback (NFB), as part of a brain-computer interface, is a technique for modulating brain activity using on-line feedback that has proved to be useful in motor rehabilitation for the chronic stroke population in addition to traditional therapies. Nevertheless, its use and applications in the field still leave unresolved questions. The brain pathophysiological mechanisms after stroke remain partly unknown, and the possibilities for intervention on these mechanisms to promote cerebral plasticity are limited in clinical practice. In NFB motor rehabilitation, the aim is to adapt the therapy to the patient's clinical context using brain imaging, considering the time after stroke, the localization of brain lesions, and their clinical impact, while taking into account currently used biomarkers and technical limitations. These modern techniques also allow a better understanding of the physiopathology and neuroplasticity of the brain after stroke. We conducted a narrative literature review of studies using NFB for post-stroke motor rehabilitation. The main goal was to decompose all the elements that can be modified in NFB therapies, which can lead to their adaptation according to the patient's context and according to the current technological limits. Adaptation and individualization of care could derive from this analysis to better meet the patients' needs. We focused on and highlighted the various clinical and technological components considering the most recent experiments. The second goal was to propose general recommendations and enhance the limits and perspectives to improve our general knowledge in the field and allow clinical applications. We highlighted the multidisciplinary approach of this work by combining engineering abilities and medical experience. Engineering development is essential for the available technological tools and aims to increase neuroscience knowledge in the NFB topic. This technological development was born out of the real clinical need to provide complementary therapeutic solutions to a public health problem, considering the actual clinical context of the post-stroke patient and the practical limits resulting from it.

Optimizing Motor Intention Detection With Deep Learning: Towards Management of Intraoperative Awareness.

OBJECTIVE: This article shows the interest in deep learning techniques to detect motor imagery (MI) from raw electroencephalographic (EEG) signals when a functional electrical stimulation is added or not. Impacts of electrode montages and bandwidth are also reported. The perspective of this work is to improve the detection of intraoperative awareness during general anesthesia. METHODS: Various architectures of EEGNet were investigated to optimize MI detection. They have been compared to the state-of-the-art classifiers in Brain-Computer Interfaces (based on Riemannian geometry, linear discriminant analysis), and other deep learning architectures (deep convolution network, shallow convolutional network). EEG data were measured from 22 participants performing motor imagery with and without median nerve stimulation. RESULTS: The proposed architecture of EEGNet reaches the best classification accuracy (83.2%) and false-positive rate (FPR 19.0%) for a setup with only six electrodes over the motor cortex and frontal lobe and for an extended 4-38 Hz EEG frequency range while the subject is being stimulated via a median nerve. Configurations with a larger number of electrodes result in higher accuracy (94.5%) and FPR (6.1%) for 128 electrodes (and respectively 88.0% and 12.9% for 13 electrodes). CONCLUSION: The present work demonstrates that using an extended EEG frequency band and a modified EEGNet deep neural network increases the accuracy of MI detection when used with as few as 6 electrodes which include frontal channels. SIGNIFICANCE: The proposed method contributes to the development of Brain-Computer Interface systems based on MI detection from EEG.

Multiclass Classification Based on Combined Motor Imageries.

Motor imagery (MI) allows the design of self-paced brain-computer interfaces (BCIs), which can potentially afford an intuitive and continuous interaction. However, the implementation of non-invasive MI-based BCIs with more than three commands is still a difficult task. First, the number of MIs for decoding different actions is limited by the constraint of maintaining an adequate spacing among the corresponding sources, since the electroencephalography (EEG) activity from near regions may add up. Second, EEG generates a rather noisy image of brain activity, which results in a poor classification performance. Here, we propose a solution to address the limitation of identifiable motor activities by using combined MIs (i.e., MIs involving 2 or more body parts at the same time). And we propose two new multilabel uses of the Common Spatial Pattern (CSP) algorithm to optimize the signal-to-noise ratio, namely MC2CMI and MC2SMI approaches. We recorded EEG signals from seven healthy subjects during an 8-class EEG experiment including the rest condition and all possible combinations using the left hand, right hand, and feet. The proposed multilabel approaches convert the original 8-class problem into a set of three binary problems to facilitate the use of the CSP algorithm. In the case of the MC2CMI method, each binary problem groups together in one class all the MIs engaging one of the three selected body parts, while the rest of MIs that do not engage the same body part are grouped together in the second class. In this way, for each binary problem, the CSP algorithm produces features to determine if the specific body part is engaged in the task or not. Finally, three sets of features are merged together to predict the user intention by applying an 8-class linear discriminant analysis. The MC2SMI method is quite similar, the only difference is that any of the combined MIs is considered during the training phase, which drastically accelerates the calibration time. For all subjects, both the MC2CMI and the MC2SMI approaches reached a higher accuracy than the classic pair-wise (PW) and one-vs.-all (OVA) methods. Our results show that, when brain activity is properly modulated, multilabel approaches represent a very interesting solution to increase the number of commands, and thus to provide a better interaction.

Deep Learning Techniques to Improve Intraoperative Awareness Detection from Electroencephalographic Signals.

Every year, millions of patients regain conscious- ness during surgery and can potentially suffer from post-traumatic disorders. We recently showed that the detection of motor activity during a median nerve stimulation from electroencephalographic (EEG) signals could be used to alert the medical staff that a patient is waking up and trying to move under general anesthesia [1], [2]. In this work, we measure the accuracy and false positive rate in detecting motor imagery of several deep learning models (EEGNet, deep convolutional network and shallow convolutional network) directly trained on filtered EEG data. We compare them with efficient non-deep approaches, namely, a linear discriminant analysis based on common spatial patterns, the minimum distance to Riemannian mean algorithm applied to covariance matrices, a logistic regression based on a tangent space projection of covariance matrices (TS+LR). The EEGNet improves significantly the classification performance comparing to other classifiers (p- value <; 0.01); moreover it outperforms the best non-deep classifier (TS+LR) for 7.2% of accuracy. This approach promises to improve intraoperative awareness detection during general anesthesia.

Hypnotic State Modulates Sensorimotor Beta Rhythms During Real Movement and Motor Imagery.

Hypnosis techniques are currently used in the medical field and directly influences the patient's state of relaxation, perception of the body, and its visual imagination. There is evidence to suggest that a hypnotic state may help patients to better achieve tasks of motor imagination, which is central in the rehabilitation protocols after a stroke. However, the hypnosis techniques could also alter activity in the motor cortex. To the best of our knowledge, the impact of hypnosis on the EEG signal during a movement or an imagined movement is poorly investigated. In particular, how event-related desynchronization (ERD) and event-related synchronization (ERS) patterns would be modulated for different motor tasks may provide a better understanding of the potential benefits of hypnosis for stroke rehabilitation. To investigate this purpose, we recorded EEG signals from 23 healthy volunteers who performed real movements and motor imageries in a closed eye condition. Our results suggest that the state of hypnosis changes the sensorimotor beta rhythm during the ERD phase but maintains the ERS phase in the mu and beta frequency band, suggesting a different activation of the motor cortex in a hypnotized state.

Wavelet-Based Semblance Methods to Enhance the Single-Trial Detection of Event-Related Potentials for a BCI Spelling System.

Based on similarity measures in the wavelet domain under a multichannel EEG setting, two new methods are developed for single-trial event-related potential (ERP) detection. The first method, named "multichannel EEG thresholding by similarity" (METS), simultaneously denoises all of the information recorded by the channels. The second approach, named "semblance-based ERP window selection" (SEWS), presents two versions to automatically localize the ERP in time for each subject to reduce the time window to be analysed by removing useless features. We empirically show that when these methods are used independently, they are suitable for ERP denoising and feature extraction. Meanwhile, the combination of both methods obtains better results compared to using them independently. The denoising algorithm was compared with classic thresholding methods based on wavelets and was found to obtain better results, which shows its suitability for ERP processing. The combination of the two algorithms for denoising the signals and selecting the time window has been compared to xDAWN, which is an efficient algorithm to enhance ERPs. We conclude that our wavelet-based semblance method performs better than xDAWN for single-trial detection in the presence of artifacts or noise.

MOTANA: study protocol to investigate motor cerebral activity during a propofol sedation.

BACKGROUND: Accidental Accidental awareness during general anesthesia (AAGA) occurs in 1-2% of high-risk practice patients and is a cause of severe psychological trauma, termed post-traumatic stress disorder (PTSD). However, no monitoring techniques can accurately predict or detect AAGA. Since the first reflex for a patient during AAGA is to move, a passive brain-computer interface (BCI) based on the detection of an intention of movement would be conceivable to alert the anesthetist. However, the way in which propofol (i.e., an anesthetic commonly used for the general anesthesia induction) affects motor brain activity within the electroencephalographic (EEG) signal has been poorly investigated and is not clearly understood. For this reason, a detailed study of the motor activity behavior with a step-wise increasing dose of propofol is required and would provide a proof of concept for such an innovative BCI. The main goal of this study is to highlight the occurrence of movement attempt patterns, mainly changes in oscillations called event-related desynchronization (ERD) and event-related synchronization (ERS), in the EEG signal over the motor cortex, in healthy subjects, without and under propofol sedation, during four different motor tasks. METHODS: MOTANA is an interventional, prospective, exploratory, physiological, monocentric, and randomized study conducted in healthy volunteers under light anesthesia, involving EEG measurements before and after target-controlled infusion of propofol at three different effect-site concentrations (0 µg.ml -1, 0.5 µg.ml -1, and 1.0 µg.ml -1). In this exploratory study, 30 healthy volunteers will perform 50 trials for the four motor tasks (real movement, motor imagery, motor imagery with median nerve stimulation, and median nerve stimulation alone) in a randomized sequence. In each conditions and for each trial, we will observe changes in terms of ERD and ERS according to the three propofol concentrations. Pre- and post-injection comparisons of propofol will be performed by paired series tests. DISCUSSION: MOTANA is an exploratory study aimed at designing an innovative BCI based on EEG-motor brain activity that would detect an attempt to move by a patient under anesthesia. This would be of interest in the prevention of AAGA. TRIAL REGISTRATION: Agence Nationale de Sécurité du Médicament (EUDRACT 2017-004198-1), NCT03362775. Registered on 29 August 2018. https://clinicaltrials.gov/ct2/show/NCT03362775?term=03362775&rank=1.

Median Nerve Stimulation Based BCI: A New Approach to Detect Intraoperative Awareness During General Anesthesia.

Hundreds of millions of general anesthesia are performed each year on patients all over the world. Among these patients, 0.1-0.2% are victims of Accidental Awareness during General Anesthesia (AAGA), i.e., an unexpected awakening during a surgical procedure under general anesthesia. Although anesthesiologists try to closely monitor patients using various techniques to prevent this terrifying phenomenon, there is currently no efficient solution to accurately detect its occurrence. We propose the conception of an innovative passive brain-computer interface (BCI) based on an intention of movement to prevent AAGA. Indeed, patients typically try to move to alert the medical staff during an AAGA, only to discover that they are unable to. First, we examine the challenges of such a BCI, i.e., the lack of a trigger to facilitate when to look for an intention to move, as well as the necessity for a high classification accuracy. Then, we present a solution that incorporates Median Nerve Stimulation (MNS). We investigate the specific modulations that MNS causes in the motor cortex and confirm that they can be altered by an intention of movement. Finally, we perform experiments on 16 healthy participants to assess whether an MI-based BCI using MNS is able to generate high classification accuracies. Our results show that MNS may provide a foundation for an innovative BCI that would allow the detection of AAGA.

Electroencephalographic modulations during an open- or closed-eyes motor task.

There is fundamental knowledge that during the resting state cerebral activity recorded by electroencephalography (EEG) is strongly modulated by the eyes-closed condition compared to the eyes-open condition, especially in the occipital lobe. However, little research has demonstrated the influence of the eyes-closed condition on the motor cortex, particularly during a self-paced movement. This prompted the question: How does the motor cortex activity change between the eyes-closed and eyes-open conditions? To answer this question, we recorded EEG signals from 15 voluntary healthy subjects who performed a simple motor task (i.e., a voluntary isometric flexion of the right-hand index) under two conditions: eyes-closed and eyes-open. Our results confirmed strong modulation in the mu rhythm (7-13 Hz) with a large event-related desynchronisation. However, no significant differences have been observed in the beta band (15-30 Hz). Furthermore, evidence suggests that the eyes-closed condition influences the behaviour of subjects. This study gives us greater insight into the motor cortex and could also be useful in the brain-computer interface (BCI) domain.

Can a Subjective Questionnaire Be Used as Brain-Computer Interface Performance Predictor?

Predicting a subject's ability to use a Brain Computer Interface (BCI) is one of the major issues in the BCI domain. Relevant applications of forecasting BCI performance include the ability to adapt the BCI to the needs and expectations of the user, assessing the efficiency of BCI use in stroke rehabilitation, and finally, homogenizing a research population. A limited number of recent studies have proposed the use of subjective questionnaires, such as the Motor Imagery Questionnaire Revised-Second Edition (MIQ-RS). However, further research is necessary to confirm the effectiveness of this type of subjective questionnaire as a BCI performance estimation tool. In this study we aim to answer the following questions: can the MIQ-RS be used to estimate the performance of an MI-based BCI? If not, can we identify different markers that could be used as performance estimators? To answer these questions, we recorded EEG signals from 35 healthy volunteers during BCI use. The subjects had previously completed the MIQ-RS questionnaire. We conducted an offline analysis to assess the correlation between the questionnaire scores related to Kinesthetic and Motor imagery tasks and the performances of four classification methods. Our results showed no significant correlation between BCI performance and the MIQ-RS scores. However, we reveal that BCI performance is correlated to habits and frequency of practicing manual activities.

Comparison of sensorimotor rhythms in EEG signals during simple and combined motor imageries over the contra and ipsilateral hemispheres.

Imaginary motor tasks cause brain oscillations that can be detected through the analysis of electroencephalographic (EEG) recordings. This article aims at studying whether or not the characteristics of the brain activity induced by the combined motor imagery (MI) of both hands can be assumed as the superposition of the activity generated during simple hand MIs. After analyzing the sensorimotor rhythms in EEG signals of five healthy subjects, results show that the imagination of both hands movement generates in each brain hemisphere similar activity as the one produced by each simple hand MI in the contralateral side. Furthermore, during simple hand MIs, brain activity over the ipsilateral hemisphere presents similar characteristics as those observed during the rest condition. Thus, it is shown that the proposed scheme is valid and promising for brain-computer interfaces (BCI) control, allowing to easily detect patterns induced by combined MIs.

Decoding Finger Flexion from Band-Specific ECoG Signals in Humans.

This article presents the method that won the brain-computer interface (BCI) competition IV addressed to the prediction of the finger flexion from electrocorticogram (ECoG) signals. ECoG-based BCIs have recently drawn the attention from the community. Indeed, ECoG can provide higher spatial resolution and better signal quality than classical EEG recordings. It is also more suitable for long-term use. These characteristics allow to decode precise brain activities and to realize efficient ECoG-based neuroprostheses. Signal processing is a very important task in BCIs research for translating brain signals into commands. Here, we present a linear regression method based on the amplitude modulation of band-specific ECoG including a short-term memory for individual finger flexion prediction. The effectiveness of the method was proven by achieving the highest value of correlation coefficient between the predicted and recorded finger flexion values on data set 4 during the BCI competition IV.

Using neuronal states for transcribing cortical activity into muscular effort.

We study the relations between the activity of corticomotoneuronal (CM) cells and the forces exerted by fingers. The activity of CM cells, located in the primary motor cortex is recorded in the thumb and index fingers area of a monkey. The activity of the fingers is recorded as they press two levers. The main idea of this work is to establish and use a collection of neuronal states. At any time, the neuronal state is defined by the firing rates of the recorded neurons. We assume that any such neuronal state is related to a typical variation (or absence of variation) in the muscular effort. Our forecasting model uses a linear combination of the firing rates, some synchrony information between spike trains and averaged variations of the positions of the levers.