The setup was a multimodal BCI: locomotion was driven by motor imagery-related sensorimotor rhythm event-related desynchronization (SMR-ERD) from scalp electroencephalography, while direction was controlled by eye gaze. Participants steered an avatar with a jogging gait along a curved course. To manipulate kinesthetic congruence between imagery and gait, they performed either jogging imagery (the congruent condition) or sustained right-hand opening imagery (the incongruent condition), then rated their subjective embodiment toward the avatar on questionnaires.
Participants reported significantly greater ownership in the congruent condition than in the incongruent condition. Jogging imagery elicited weaker SMR-ERD than hand-opening imagery, but the two conditions showed no statistically significant differences in course completion rate or in time remaining when the course was finished.
The authors conclude that neural signal strength and embodied experience can dissociate in multimodal BCIs, and that imagery-action congruence should be weighed alongside conventional decoding metrics. On that basis, they argue for a shift from signal-centric toward context-aware BCI design aimed at more intuitive, more embodied interaction.