Guest Lecture by Fabien Lotte (Inria Bordeaux Sud-Ouest/LaBRI, France)Understanding and Modeling User Training in Mental-Imagery-based Brain-Computer Interfaces
Brain-Computer Interfaces (BCIs) are systems that can translate the brain activity patterns of a user into messages or commands for an interactive application. Such brain activity is typically measured using Electroencephalography (EEG), before being processed and classified by the system. A prominent type of BCI is Mental Imagery-based BCI (MI-BCI), with which users send commands by performing mental imagery tasks, e.g., imagined movements or mental subtractions, which are recognized by the BCI. MI-BCIs have proven promising for a wide range of applications ranging from communication and control for motor impaired users, to gaming targeted at the general public and stroke rehabilitation, to name a few. Despite this promising potential, MI-BCIs are still scarcely used outside laboratories for practical applications. The main reason preventing MI-BCIs from being widely used is arguably their poor usability, which is notably due to their low robustness and reliability. There is thus a need to make BCI more reliable.
Currently, the majority of BCI research aims at improving BCI reliability by improving EEG signal processing and classification algorithms. However, another key element of the BCI loop should be considered to improve their reliability: the users themselves. Indeed, MI-BCI control is known to be a skill that needs to be learnt. The more users practice to control the MI-BCIs, the better they can learn to control them and thus the better the reliability of the system. An additional promising direction to improve the reliability of MI-BCIs is thus to improve user training. Unfortunately, why some users managed to learn to control MI-BCIs whereas some other do not or how to favor this learning, is rather poorly understood. Thus, it appears as necessary to understand and to model MI-BCI user learning processes, in order to be able to improve them subsequently, based on these models. In this talk, I will thus present our works aiming at understanding and modeling, both at the theoretical, conceptual and computational levels, MI-BCI user training and performances. I will then illustrate how we can leverage the knowledge gain from these models to improve MI-BCI user training, by providing better feedbacks or training tasks, adapted to the users.
About Fabien Lotte: Dr. Lotte is a research group leader of the team Potioc (http://team.inria.fr/potioc) at the Inria Bordeaux Sud-Ouest/LaBRI, France. He has substantially contributed both, to machine learning methods for brain-computer interfaces and to the question, how existing BCI systems can be improved beyond decoding algorithms in order to better meet the BCI users' requirements.
(Copy and paste into your host application - further infos here)