EEG-based monitoring of mental fatigue during virtual-reality motor imagery tasks.

Abstract

Prolonged motor-imagery training in immersive virtual-reality environments can induce mental fatigue, reducing engagement and potentially limiting the effectiveness of neurorehabilitation. This study investigated neural markers of mental fatigue by recording electroencephalography (EEG) from healthy participants during extended motor-imagery and control sessions in a head-mounted display setup. Multidimensional analysis was applied to extract spectral, spatial, and temporal features while using a novel deflation step for removing task-related motor components to isolate fatigue-specific activity. Evidence of mental fatigue was consistently seen in parieto-occipital alpha-band modulation, with increases in alpha power corresponding to subjective reports and EEG-based measures of mental fatigue. The derived models were robust to common EEG artifacts and demonstrated consistent fatigue estimation across tasks and sessions. These findings suggest that individualized neural markers can enable real-time monitoring of fatigue (with an accuracy of 83.49 ± 6.34%), allowing adaptive adjustments of task difficulty or pacing in brain-computer interface systems. This work advances understanding of the neurophysiological dynamics of mental fatigue during immersive motor-imagery tasks and provides a foundation for designing more effective, personalized neurorehabilitation protocols.