Section: New Results

Brain Computer Interfaces

Augmenting Motor Imagery Learning for Brain–Computer Interfacing Using Electrical Stimulation as Feedback

Participants : Saugat Bhattacharyya [School of Bio-Science and Engineering, Calcutta] , Mitsuhiro Hayashibe [Tohoku University, Sendai] , Maureen Clerc.

Brain-computer Interfaces (BCI) and Functional electrical stimulation (FES) contribute significantly to induce cortical learning and to elicit peripheral neuronal activation processes and thus, are highly effective to promote motor recovery. This study aims at understanding the effect of FES as a neural feedback and its influence on the learning process for motor imagery tasks while comparing its performance with a classical visual feedback protocol. The participants were randomly separated into two groups: one group was provided with visual feedback (VIS) while the other received electrical stimulation (FES) as feedback. Both groups performed various motor imagery tasks while feedback was provided in form of a bi-directional bar for VIS group and targeted electrical stimulation on the upper and lower limbs for FES group. The results shown in this paper suggest that the FES based feedback is more intuitive to the participants, hence, the superior results as compared to the visual feedback. The results suggest that the convergence of BCI with FES modality could improve the learning of the patients both in terms of accuracy and speed and provide a practical solution to the BCI learning process in rehabilitation.

This work, obtained in the context of the BCI-LIFT IPL, has been published in [9].

Adaptive parameter setting in a code modulated visual evoked potentials BCI

Participants : Federica Turi, Maureen Clerc.

Code-modulated visual evoked potentials (c-VEPs) BCI are designed for high-speed communication. The setting of stimulus parameters is fundamental for this type of BCI, because stimulus parameters have an influence on the performance of the system. In this work we design a c-VEP BCI for word spelling, in which it is possible to find the optimal stimulus presentation rate per each subject thanks to an adaptive setting parameter phase. This phase takes place at the beginning of each session and allows to define the stimulus parameters that are used during the spelling phase. The different stimuli are modulated by a binary m-sequence circular-shifted by a different time lag and a template matching method is applied for the target detection. We acquired data from 4 subjects in two sessions. The results obtained for the offline spelling show the variability between subjects and therefore the importance of subject-dependent adaptation of c-VEP BCI.

This work has been published in [32].

Participation to the Cybathlon BCI Series

Participants : Karine Leclerc [Centre René Labreuille, Le Cannet] , Magali Mambrucchi [Centre René Labreuille, Le Cannet] , Amandine Audino, Pierre Giacalone, Federica Turi, Maureen Clerc, Théodore Papadopoulo.

The CYBATHLON is a unique championship in which people with physical disabilities compete against each other to complete everyday tasks using state-of-the-art technical assistance systems. Athena participated in the CYBATHLON BCI Series that took place on September 8th, 2019 as a satellite event of the Graz Brain–Computer Interface Conference. Athena was part of a bigger Inria team which encompassed also the Inria Bordeaux Sud-Ouest Potioc team (participants from Bordeaux are not listed). For both Inria sub-teams, it was a first participation to such a competition : we learned a lot about the practical issues of working with people with physical disabilities and on all the practical issues that can encounter a BCI user out of the lab. The actual competition consisted of driving a car on a track by issuing three types of commands (Left, Right, Lights) using mental imagery. Even though our pilot finished last, she was for each run leading the race till a few seconds before its end. A great satisfaction was to see that the software that we built worked reliably out of the lab (many teams have had troubles in issuing commands and had to redo a race). Yet, this required a lot of last minute work to integrate smoothly in the competition system: we learned a lot in this respect. The poster [36] summarises this effort.

BCI Performance prediction

Participants : Maureen Clerc, Nathalie Gayraud, Laurent Bougrain [NeuroSys Project-Team] , Sébastien Rimbert [NeuroSys Project-Team] , Stéphanie Fleck [Perseus] .

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). 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. This work is an outcome of the BCI-LIFT IPL and was published in [18]. Another joint publication [19] investigated median nerve stimulation as a new approach to detect intraoperative awareness during General Anesthesia.

EEG Classification of Auditory Attention

Participants : Joan Belo, Johann Benerradi, Maureen Clerc, Michel Pascal [Nice Music Conservatory] , Daniele Schön [Institut de Neurosciences des Systèmes] .

In a Master's thesis [33] in collaboration with Nice Music Conservatory and Institut de Neurosciences des Systèmes, we focused on analyzing auditory attention of human participants who are presented two auditory streams, simultaneously on left and right. By analyzing the EEG signals measured, the problem is to detect to which stream the participant is attending. Auditory Attention is also the topic of the PhD thesis of Joan Belo, funded by a CIFRE with Oticon Medical.

Innovative Brain-Computer Interface based on motor cortex activity to detect accidental awareness during general anesthesia

Participants : Sébastien Rimbert, Philippe Guerci, Nathalie Gayraud, Claude Meistelman, Laurent Bougrain.

Accidental Awareness during General Anesthesia (AAGA) occurs in 1-2% of high-risk practice patients and is responsible for severe psychological trauma, termed post-traumatic stress disorder (PTSD). Currently, monitoring techniques have limited accuracy in predicting or detecting AAGA. Since the first reflex of a patient experiencing 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 and prevent this phenomenon. However, the way in which the propofol (an anesthetic drug commonly used for inducing and maintaining general anesthesia) affects the motor brain activity and is reflected by the electroencephalo-graphic (EEG) signal has been poorly investigated and is not clearly understood. The goal of this forward-looking study is to investigate the motor activity behavior with step-wise increase of propofol doses in 4 healthy subjects and provide a proof of concept for such an innovative BCI.

This work has been published in [26].