Our partner institutions currently offer two courses that have a strong focus on robotics at the master level, although it is worth noting that students with a wide variety of… Read more
Looking for publications? You might want to consider searching on the EPFL Infoscience site which provides advanced publication search capabilities.
One of the challenges of using brain-computer interfaces (BCIs) over extended periods of time is the variation of the users’ performance across different experimental days. The goal of the current study is to propose a performance estimator for an electroencephalography-based motor imagery BCI by assessing the reliability of a command (i.e., predicting a ‘short’ or ‘long’ command delivery time, CDT). Using a short time window (< 1.5 s, shorter than the delivery time) of the mental task execution and a linear discriminant analysis classifier, we could reliably differentiate between long and short CDT (AUC around 0.8) for 9 healthy subjects. Moreover, we assessed the feasibility of providing online adaptive assistance using the performance estimator in a BCI game, comparing two conditions: (i) allowing a ‘fixed timeout’ to deliver each command or (ii) providing ‘adaptive assistance’ by giving more time if the performance estimator detects a long CDT. The results revealed that providing adaptive assistance increases the ratio of correct commands significantly (p < 0.01). Moreover, the task load index (measured via the NASA TLX questionnaire) shows a significantly higher user acceptance in case of providing adaptive assistance (p < 0.01). Furthermore, the results obtained in this study were used to simulate a robotic navigation scenario, which showed how adaptive assistance improved performance.
In order for brain-computer interfaces (BCIs) to be used reliably for extended periods of time, they must be able to adapt to the users evolving needs. This adaptation should not only be a function of the environmental (external) context, but should also consider the internal context, such as cognitive states and brain signal reliability. In this work, we propose three different shared control frameworks that have been used for BCI applications: contextual fusion, contextual gating, and contextual regulation. We review recently published results in the light of these three context-awareness frameworks. Then, we discuss important issues to consider when designing a shared controller for BCI.