Abstract
Prolonged use of brain-computer interfaces (BCIs)
with virtual reality (VR) via head-mounted displays (HMDs)
induces mental fatigue, potentially impairing
neurorehabilitation. This study examines EEG-based fatigue
markers in healthy participants during extended BCI-HMD
sessions. Fatigue was classified using N-way Partial Least
Squares (N-PLS) with linear discriminant analysis, achieving
82.42% (± 7.5) accuracy. N-PLS components revealed spatial-
spectral patterns in occipital and sensorimotor alpha activity.
Temporal trajectories indicated progressive fatigue
accumulation during sessions. Results demonstrate the
feasibility of EEG-based fatigue monitoring for optimizing
BCI-HMD post-stroke neurorehabilitation.
Go back