The effect of multimodal and enriched feedback on SMR-BCI performance

•Uncertainty is not static and can vary substantially over time, possibly rendering BCI frustrating for the end-user.•The multimodal feedback gives direct continuous feedback about the quality of motor imagery classification.•Participants were able to control the BCI with the funnel feedback with better performance during the initial session and less frustration compared to the conventional cursorbar feedback.

ObjectiveThis study investigated the effect of multimodal (visual and auditory) continuous feedback with information about the uncertainty of the input signal on motor imagery based BCI performance. A liquid floating through a visualization of a funnel (funnel feedback) provided enriched visual or enriched multimodal feedback.MethodsIn a between subject design 30 healthy SMR-BCI naive participants were provided with either conventional bar feedback (CB), or visual funnel feedback (UF), or multimodal (visual and auditory) funnel feedback (MF). Subjects were required to imagine left and right hand movement and were trained to control the SMR based BCI for five sessions on separate days.ResultsFeedback accuracy varied largely between participants. The MF feedback lead to a significantly better performance in session 1 as compared to the CB feedback and could significantly enhance motivation and minimize frustration in BCI use across the five training sessions.ConclusionThe present study demonstrates that the BCI funnel feedback allows participants to modulate sensorimotor EEG rhythms. Participants were able to control the BCI with the funnel feedback with better performance during the initial session and less frustration compared to the CB feedback.SignificanceThe multimodal funnel feedback provides an alternative to the conventional cursorbar feedback for training subjects to modulate their sensorimotor rhythms.