Search

International audience; We present and share a large database containing electroencephalographic signals from 87 human participants, collected during a single day of brain-computer interface (BCI) experiments, organized into 3 datasets (A, B, and C) that were all recorded using the same protocol: right and left hand motor imagery (MI). Each session contains 240 trials (120 per class), which in total represents more than 20,800 trials, or approximately 70 hours of recording time.It includes the performance of the associated BCI users, detailed information about the demographics, personality profile as well as some cognitive traits and the experimental instructions and codes (executed in the open-source platform OpenViBE).Such database could prove useful for various studies, including but not limited to: 1) studying the relationships between BCI users' profiles and their BCI performances, 2) studying how EEG signals properties varies for different users' profiles and MI tasks, 3) using the large number of participants to design cross-user BCI machine learning algorithms or 4) incorporating users' profile information into the design of EEG signal classification algorithms.

International audience

International audience; Introduction:While several meta-analyses have demonstrated the relevance of BCIs for improving motor recovery after stroke, these technologies are still barely used in clinical practice. We hypothesise that optimising the acceptability of BCIs could be used as a lever to increase the efficacy of BCI-based motor rehabilitation procedures, and thereby their clinical use. Indeed, a better acceptability will enable a reduction of anxiety levels and an enhancement of motivation and engagement in the procedure. Both will in turn allow for more cognitive resources to be allocated to the task, thus favouring learning and, ultimately, motor recovery. To the best of our knowledge, so far only one study has assessed the relevance of a BCI-based stroke rehabilitation procedure using acceptability measures among their primary criteria of efficacy. Most often, acceptability is only considered as an attribute of user satisfaction, itself being a dimension of user experience. Yet, given its potential impact on BCI use, it seems important to study acceptability as an integral component. Thus, our objective is to rely on validated models depicted in other literature in order to design a first thorough model of acceptability specifically dedicated to BCI-based procedures for motor rehabilitation after stroke. Material, Methods and Results: We designed a model of BCI acceptability based on the technology acceptance model 3 (TAM3) and on the unified theory of acceptance and use of technology 2 (UTAUT2) that assess acceptability in terms of perceived usefulness (PU), perceived ease of use (PEOU) and behavioural intention (BI). Then, we created a questionnaire based on our model in order to estimate its validity and to quantify the influence that each factor had on BCIacceptability. We collected a data sample of 753 respondents representative of the general public in France. We targeted the general public for two main reasons. (i) Due to the high prevalence of stroke, many of us are concerned, more or less directly, by stroke. (ii) Patients’ acceptability is likely to be influenced by the opinions and attitudes of their close relatives (who are part of the general public). Descriptive analyses: Our results suggest that BCIs are associated with high levels of acceptability: PU (8.28/10 ±1.57), PEOU (7.17/10 ±1.57), BI (8.23/10 ±1.69). Validity analyses: Cronbach’s α coefficient analyses revealed a satisfactory internal consistency of our questionnaire, i.e., the items used to measure each factor are mostly consistent and not too redundant (12/17 factors in [0.72;0.95], 4/17 in [0.50;0.62] and 1 = 0.97). Furthermore, a confirmatory factor analysis showed that the structure of the model is adequate. Quantification analyses: Regression analyses based on random forest algorithm revealed that BI is mainly driven by the PU of the system and by the perceived benefits on risk ratio associated with the technology. PU itself is mainly determined by the scientific relevance of BCIs and by PEOU. The main determinants of PEOU are ease of learning and playfulness.Discussion: Our results suggest that beyond the subjective norm (i.e., perceived opinions of our close ones), several factors impact significantly BCI acceptability. Given the weight of the scientific relevance and benefits/risk ratio, it is of utmost importance to clearly inform the population. In addition, the impact of playfulness and ease of learning should encourage us to adapt the rehabilitation procedures to each patient. In order to refine this model, additional data should be collected, in particular with i) patients and caregivers; ii) persons from different cultures; and iii) in different contexts, i.e. for other use cases.

International audience; Introduction:Motor imagery (MI) can be defined as a “dynamic state during which one simulates an action mentally without any body movement” [1]. The aim of MI is to optimise learning (e.g., in athletic training) or re-learning (e.g., in motor rehabilitation after stroke) by mastering the technique of new motor skills, but also through attentional focus [2] thanks to brain plasticity mechanisms. Indeed, similarities exist between MI and motor execution with regards to the solicitation of certain brain networks and regions, including premotor, parietal, and somatosensory regions [3]. Current BCI protocols targeting MI consist in positively reinforcing the maximum modulation of sensorimotor rhythms (SMRs) from baseline levels. This suggests that we consider that the growing expertise in the MI task will be associated with a higher desynchronisation of neurons in the sensorimotor cortices [4]. Yet, experiments investigating the neural efficiency hypothesis have shown that experts happen to have a reduced modulation of neural activity in comparison to novices [5], which can be attributed to a more efficient resource distribution. This efficiency would take form of reinforced temporal and spatial stability during MI tasks [6,7]. Thus, our questions are as follows: Q1. Does expertise modify the brain patterns associated with a MI task? Q2. If so, are those modifications elicited exclusively during MI of mastered movements? Or do these modifications reflect the acquisition of a generic skill (whatever the imagined movement)? Q3. Is maximising the percentage of desynchronisation a relevant objective? If not, what metrics of performance should be used in order to optimise the training of users/patients to self-regulate their brain activity? In order to investigate those different questions, we will recruit athletes who can be considered as an expert population in MI because of their frequent mental training use (to prevent overtraining or during rehabilitation but also as warm-up routines and rehearsal technique). Material and Methods:We will recruit 48 participants who will be divided into three groups: “basketball experts” (G1exp), “dance experts” (G2exp) and “novices” (G3) [16 participants per group, 8 men, 8 women]. All participants will perform 20 MI trials lasting 10s each for all four of the following movements: a simple reaching action (T1 – for which all participants are experts), a complex novel drawing task (T2 – for which none are expert), a basketball free throw (T3 – expertise of G1exp only) and a short pre-defined dancing choreography (T4 – expertise of G2exp only). EEG activity will be recorded during trials. This paradigm will enable us to compare MI-related brain patterns between experts and novices (G1exp/G2exp vs. G3) (Q1). It will also enable us to assess the extent to which the potential modifications of brain patterns due to expertise are specific to expert movements or generic (for G1exp and G2exp: T3 vs. T4; for all, control: T1 vs. T2) (Q2). From those results, we will investigate which performance metrics seem the most relevant to use in MI-based BCI/NF paradigms (Q3).Discussion:This study will enable us to acquire new knowledge regarding the neural efficiency hypothesis. If our results comfort this hypothesis, it will be important to identify, implement and evaluate new metrics of performance to guide BCI/NF users during their training (i.e., instead of ERD%). Following the neural efficiency hypothesis, we expect experts to show brain patterns that are more spatially and temporally stable than those of novices [6,7]. In other terms, in experts, we expect the modulations of brain activity during MI to be circumscribed to sensorimotor cortices (provided that they perform kinaesthetic MI) and to be highly stable across trials in terms of location and frequency. In addition, we hypothesise that those modulations will represent a general skill. In other terms, we expect the same patterns to be elicited for experts when doing MI of a mastered technique from their discipline, but also when doing MI of a novel movement of similar nature (in our case a different physical activity). Indeed, we believe that a transfer of neural efficiency exists. However, this phenomenon might only happen to a certain extent and not be identifiable for MI of a complex novel task or of a simple everyday life action. To our knowledge, this last hypothesis hasn’t been tested in the current literature and will therefore require an exploratory approach.References:[1] Guillot, A. and Collet, C. (2008). Construction of the Motor Imagery Integrative Model in Sport: A review and theoretical investigation of motor imagery use. IRSEP 1, 31–44[2] Budnik-Przybylska, D. et al. (2021). Neural Oscillation During Mental Imagery in Sport: An Olympic Sailor Case Study. Frontiers in Human Neuroscience 15, 669422.[3] Hardwick, R. M. et al. (2018). Neural Correlates of Action: Comparing Meta-Analyses of Imagery, Observation, and Execution. Neuroscience & Biobehavioral Reviews 94, 31‑44.[4] Ono, T. et al. (2013). Daily Training with Realistic Visual Feedback Improves Reproducibility of Event-Related Desynchronisation Following Hand Motor Imagery. Clinical Neurophysiology 124(9), 1779 86.[5] Li, L. and Smith, D. M. (2021). Neural Efficiency in Athletes: A Systematic Review. Frontiers in Behavioral Neuroscience 15.[6] Del Percio, C. et al. (2009). “Neural Efficiency” of Athletes’ Brain for Upright Standing: A High-Resolution EEG Study. Brain Research Bulletin 79(3), 193‑200.[7] Kraeutner, S. N. et al. (2018). Experience Modulates Motor Imagery-Based Brain Activity. EJN 47, no. 10: 1221–29.

International audience; Introduction: Stroke leaves around 40% of surviving patients dependent in their activities, notably due to severe motor disabilities[1]. BCIs have been shown to favour motor recovery after stroke [2], but this efficiency has not reached yet the level required to achieve a clinical usage. We hypothesise that improving BCI acceptability, notably by personalising BCI-based rehabilitation procedures to each patient, will reduce anxiety and favour engagement in the rehabilitation process, thereby increasing the efficiency of those procedures. To test this hypothesis, we need to understand how to adapt BCI procedures to each patient depending on their profile. Thus, we constructed a model of BCI acceptability based on the literature [3], adapted it in a questionnaire, and distributed the latter to post-stroke patients (N=140). Methods: The questionnaire consisted of i) 3 target factors used as a proxy of BCI acceptability, namely the perceived usefulness (PU), perceived ease of use (PEoU) intention to use (IU) and ii) 23 explanatory factors that could influence acceptability. First, k-mean clustering analyses were performed to identify different profiles of patients. Then, for each cluster, elastic net regressions were used to identify the explanatory factors that predicted PU, PEoU and IU the best, i.e., to identify the factors that are the most important to personalise for each patient. Results: Five clusters (c1 to c5) were identified. The regression analyses indicated that the following factors had to be considered: (c1 & c5) "scientific relevance" & "ease of learning"; (c2) "benefits/risks ratio", "ease of learning", "visual aesthetic" & "result demonstrability"; (c3) "scientific relevance" & "benefits/risks ratio";(c4) none. Perspectives: We will use those results in a clinical study to personalise the BCI procedures to each patient. We expect lower anxiety and better motivation, acceptability and motor recovery with this personalised setting than with a standard one. Sources [1] Inserm, 2019 [2] Cervera, María A., et al. "Brain-computer interfaces for post-stroke motor rehabilitation: a meta-analysis."

National audience; Virtual reality (VR) relies heavily on the sense of presence - the feeling of being physically present in a virtual environment (VE). Currently, presence is mainly assessed through subjective questionnaires, which cannot be completed while users experience VR as it would disrupt their immersion. To address this limitation, we propose to explore the feasibility of using electroencephalography (EEG) to monitor users' level of presence and classify it based on the reaction to oddball stimuli.

International audience; Motor imagery (MI) is widely used in sports, in particular for motor learning and anxiety management [1]. Indeed, athletes consistently report using MI for rehearsing technique, maintaining arousal or enhancing self-confidence [2]. However, optimal gains are limited by the fact that they receive no feedback and cannot assess their performance. This can have detrimental effects on their motivation to diligently practice MI. Therefore, using Neurofeedback (NF) appears to be an adapted solution, allowing athletes to know if they are correctly performing the task. This addition also makes it possible to objectify performance and quantify progression throughout the sessions. In fact, three recent meta-analysis testify that NF improves the ability to self-regulate brain activity and sport performance [3, 4, 5]. The number of papers using MI-NF protocols has increased in the past few years. However, their efficiency has been shown to depend on various factors such as expertise in the imagined task [6], personality traits [7], MI ability [8] or NF acceptability [9]. Some of these factors can be assessed using validated questionnaires [10,11,12] but items are generally focused on questioning the current use of MI or thoughts on NF. An issue, as we know that practice can sometimes differ from real needs and expectations. To our knowledge, no previous work has investigated sports field actors’ preferences with the aim of designing specific NF tools, perfectly fitted to their profile, goals and expectations. Therefore, based on pre-existing questionnaires, we will build and broadcast an online questionnaire for athletes where personality, MI use, NF acceptability and NF preferences will be assessed. This method will enable us to personalise MI-NF trainings to fit each athlete’s needs. A longitudinal study will then require that athletes follow a personalised procedure to investigate how it impacts their MI and NF performances as well as their acceptability levels.[1] Guillot, A., & Collet, C. (2008). Construction of the Motor Imagery Integrative Model in Sport : A review and theoretical investigation of motor imagery use. International Review of Sport and Exercise Psychology, 1.[2] Munroe, K. J., Giacobbi Jr., P. R., Hall, C., & Weinberg, R. (2000). The Four Ws of Imagery Use : Where, When, Why, and What. Sport Psychologist, 14(2), 119.[3] Mirifar, A., Beckmann, J., & Ehrlenspiel, F. (2017). Neurofeedback as supplementary training for optimizing athletes’ performance: A systematic review with implications for future research. Neuroscience & Biobehavioral Reviews, 75, 419 432. [4] Xiang, M.-Q., Hou, X.-H., Liao, B.-G., Liao, J.-W., & Hu, M. (2018). The effect of neurofeedback training for sport performance in athletes: A meta-analysis. Psychology of Sport and Exercise, 36, 114 122.[5] Gong, A., Gu, F., Nan, W., Qu, Y., Jiang, C., & Fu, Y. (2021). A Review of Neurofeedback Training for Improving Sport Performance From the Perspective of User Experience. Frontiers in Neuroscience, 15.[6] Kraeutner, S. N., McWhinney, S. R., Solomon, J. P., Dithurbide, L., & Boe, S. G. (2018). Experience modulates motor imagery-based brain activity. European Journal of Neuroscience, 47(10), 1221 1229.[7] Jeunet, C., N’Kaoua, B., Subramanian, S., Hachet, M., & Lotte, F. (2015). Predicting Mental Imagery-Based BCI Performance from Personality, Cognitive Profile and Neurophysiological Patterns. PLOS ONE, 10(12), e0143962.[8] Guillot, A., Collet, C., Nguyen, V. A., Malouin, F., Richards, C., & Doyon, J. (2008). Functional neuroanatomical networks associated with expertise in motor imagery. NeuroImage. [9] Morone, G., Pisotta, I., Pichiorri, F., Kleih, S., Paolucci, S., Molinari, M., Cincotti, F., Kübler, A., & Mattia, D. (2015). Proof of principle of a brain-computer interface approach to support poststroke arm rehabilitation in hospitalized patients: Design, acceptability, and usability. Archives of Physical Medicine and Rehabilitation, 96(3 Suppl), S71-78. [10] Hall, C. R., Rodgers, W. M., & Barr, K. A. (1990). The Use of Imagery by Athletes in Selected Sports. The Sport Psychologist, 4(1), 1 10.[11] Plaisant, O., Courtois, R., Réveillère, C., Mendelsohn, G. A., & John, O. P. (2010). Validation par analyse factorielle du Big Five Inventory français (BFI-Fr). Analyse convergente avec le NEO-PI-R. Annales Médico-psychologiques, revue psychiatrique, 168(2), 97 106. [12] Grevet, E., Forge, K., Tadiello, S., Izac, M., Amadieu, F., Brunel, L., Pillette, L., Py, J., Gasq, D., & Jeunet-Kelway, C. (2023). Modeling the acceptability of BCIs for motor rehabilitation after stroke: A large scale study on the general public. Frontiers in Neuroergonomics, 3.

International audience; During motor imagery-based brain computer interfaces (MI-BCI) training, users are most often instructed to perform kinesthetic motor imagery, i.e., imagine the sensations related to a movement [1]. However, there is a great variety of sensations associated with motor imagery that can either be interoceptive or exteroceptive [2]. Interoceptive sensations related to movement arise from within the body, such as muscle contraction, whereas exteroceptive sensations are sensed through the skin, such as touch or vibration. Among those different sensations, we do not know which to advise MI-BCI users to imagine [3].Thus, our experiment aims at studying the influence of imagining sensations on neurophysiological activity and user experience. It will consist in a two hours session during which participants (thirty expected) will be equipped with an electroencephalographic headset while performing five MI tasks: interoceptive, exteroceptive and both, with either a unique or multiple exteroceptive sensations, i.e., pressure and vibration (Fig. 1). After all MI conditions, participants will be asked to fill a user-experience questionnaire [4]. The kinesthetic and visual imagery abilities of participants will also be considered in our analyses [5].Based on previous results, we expect that imagining both exteroceptive and interoceptive sensations leads to a stronger desynchronization over the sensorimotor cortex [6] than imagining interoceptive sensations only [7, 8], in turn stronger than imagining exteroceptive sensations only. Imagining several exteroceptive sensations should also enable a stronger desynchronization than imagining only one [9]. We also expect a correlation between participants’ kinesthetic and visual imagery abilities, their neurophysiological results [10, 11] and their preferences.Our results should provide useful insights on which instructions to give to participants during MI-BCI user training and potentially how to adapt them to the participants’ profile.

International audience; The adoption of a technology can be defined as the decision made by a person or a society to accept and use it. It is based on two phases during which the willingness to use that technology is assessed. The first phase corresponds to the evaluation of the acceptability of the technology and takes place before the use while the second phase, the technology acceptance, comes afterwards. Assessing the acceptability of BCIs/NF is of utmost importance as it could provide i) insights on the propensity of people to use these technologies and ii) leads to facilitate their adoption.We reviewed the literature dealing with the acceptability of BCIs/NF and using validated questionnaires and models. Through a reproducible search method, we retrieved a list of 55 main references of acceptability questionnaires and models. Then, we searched Scopus, PubMed, and Web of Science to identify the articles in the fields of BCIs/NF that cited these references. We included in our review all the articles reporting experimental results on the acceptability of BCIs/NF that were based on one or several validated acceptability questionnaire(s).Four articles were included in our review. Two of them assessed the acceptability while the others assessed the acceptance of the BCIs. While there are too few articles included to conclude on the acceptability of those technologies, the results are encouraging. BCIs were perceived as useful at 78% by motor impaired individuals who used BCIs for pain management and at 84% by neurotypical individuals who used BCIs to browse an instruction manual. The perceived ease of use varied from 30% to 80%, which could be related to differences in the protocols. Among other factors, the presence of social support appeared to influence the acceptability.

International audience; A progressive desynchronization of beta-band power (β-desync; 13-30 Hz) has been consistently observed before movement initiation. Although mounting evidence demonstrate a significant relationship between this β-desync and subsequent motor vigor, a neurophysiological foundation to explain this finding is still lacking. In the present study, we proposed that this association might come from changes in motor motivation and perceived effort. This hypothesis will be tested by recording the signal from electroencephalography and using it as neurofeedback to decrease or increase β-desync in two distinct sessions, in an intra-subject and double-blinded design. First, participants will be screened based on their ability to up- and down-regulate their β-desync. The selected participants will then perform the experimental test, in which each trial comprised a neurofeedback phase directly followed by an effort phase. The effort will consist in squeezing a dynamometer with the right hand during an allocated time of 10 seconds. Participants will be asked to apply enough force to cross a defined threshold, and that the more time the exerted force would exceed the threshold, the more money they would earn. Subjective perception of effort will be evaluated afterwards using an analog scale. Because effort exertion is associated with increased pupil dilation, which is itself modulated by noradrenergic activity, we will further investigate the influence of noradrenergic activity on changes in β-desync due to neurofeedback and effort exertion. Besides the contribution to unraveling the neurophysiological and functional bases of beta activity, the present work is of special interest to the study of Parkinson’s disease by helping to better understand the association between beta activity and motor symptoms.

International audience; Motor Imagery-based BCI (MI-BCI) performances remain highly variable, with up to 30% of users unable to achieve control. Operating a BCI system requires users to learn to modulate their neural activity. This is a complex task, thus the training procedure is crucial. Our study draws on task-feedback transparency to improve the feedback given during training. MI-BCI protocols often feature abstract feedback (e.g., visual gauge) unrelated to the imagined movement. Transparent feedback, directly related to the task, may reinforce agency-the ability to self-attribute actions, such as feedback movements-and embodiment, facilitating kinesthetic MI. This could lessen complexity in BCI control and ultimately improve BCI performance.We tested this hypothesis in a beta-based MI neurofeedback task with electroencephalography (EEG). Twenty-three subjects participated in a single MIneurofeedback session that consisted in imagining clenching their right hand rhythmically. Feedback was based on online laplacian-filtered 8-30Hz power on C3 electrode computed with OpenViBE. There were 3 feedback conditions of expected increasing transparency: pendulum; virtual hand; virtual hand with additional motor illusion induced by vibrations to the participant's hand. The amplitude of the pendulum and virtual hand movements was proportional to beta desynchronisation. Each participant performed two runs of five 24s-trials per condition. Agency, embodiment, and user experience questionnaires were included. Repeated-measures ANOVA showed better BCI performance when using the virtual hand than the pendulum. However, participants had lower performance in the visuotactile condition compared to the virtual hand-only feedback. Agency showed a similar pattern.Feeling of success in the task did not vary with feedback conditions, but the virtual hand feedback felt easier to use. BCI performance correlated positively with agency for all conditions and with embodiment in conditions with a virtual hand.Thus, visual transparency—but not tactile feedback addition—increased BCI performance and this was associated with better agency and embodiment.