Dynamics of motor imagery skill formation using a brain – computer interface with multimodal feedback

Objective: to assess the dynamic of motor imagery skill formation while using brain-computer interface (BCI) with visual and vibrotactile biofeedback. Material and methods. The pilot study included 10 healthy volunteers aged 29–34 years. All participants were right-handed. Over a five-day protocol, two feedback modalities were employed: visual and vibrotactile. Detection of motor imagery events occurred in real-time using a classifier based on the analysis of electroencephalographic (EEG) activity patterns. The accuracy (proportion) of recognizing motor imagery for both the left and right hands was evaluated across five experimental sessions. Results. It was demonstrated that learning followed a U-shaped trajectory: following an initial decline in recognition rates during the midpoint of the training sessions (dropping from baseline levels of 76% and 80% to 66% and 71% for the right and left hands, respectively), there was significant improvement by the fourth or fifth session, achieving peak values throughout the experiment (approximately 85% and 90%, respectively). This indicated the establishment of stable neural representations. Conclusion. Acquisition of motor imagery skills through a brain-computer interface incorporating visual and vibrotactile feedback exhibits a U-shaped learning curve, culminating at its highest point on the fifth day of training.