Neurophysiological correlates of user engagement in mixed reality exergames

<i>Objective.</i>This study investigates the neurophysiological correlates of user engagement (UE) during upper-limb exercises in mixed reality (MR)-based exergames. It explores whether MR environments modulate brain activity differently than traditional exercises and whether physical object interaction affects engagement.<i>Methods.</i>High-density electroencephalography (EEG) was recorded from fifteen healthy adults performing bilateral arm exercises under four conditions-with and without MR and physical object interaction. Engagement indices were derived from spectral features (theta, alpha, and their ratios) at both source- and sensor-levels across predefined cortical regions. Time-course dynamics and subjective user experience were also evaluated.<i>Main results.</i>MR-based exergames significantly increased theta power and decreased alpha power compared to non-MR conditions, indicating greater cognitive and attentional engagement. These effects were consistent across frontal, sensorimotor, and parietal regions and remained stable over time. Physical object interaction had no significant effect on engagement. Subjectively, MR conditions were rated higher in stimulation, efficiency, and novelty.<i>Significance.</i>The effectiveness of MR in rehabilitation depends on sustained UE, which is typically assessed through subjective questionnaires. This study demonstrates that EEG-derived metrics-particularly theta and alpha activity-objectively reflect engagement. These results support EEG as a real-time monitoring tool for developing adaptive, brain-responsive rehabilitation technologies.