Jogging imagery congruent with avatar gait is associated with enhanced avatar ownership during multimodal BCI control

Abstract

Objective. Multimodal brain–computer interfaces (BCIs) integrate neural and non-neural signals to enable intuitive interaction with external devices. However, BCI controllability is still predominantly evaluated based on neural signal strength, leaving the role of embodied self-experience largely unclear. Here, we investigated how kinesthetic congruence between motor imagery (MI) and avatar gait under BCI control contributes to subjective embodiment. Approach. Participants navigated a virtual avatar with a jogging gait through a curved course using a multimodal BCI. Locomotion was controlled by MI-related sensorimotor rhythm event-related desynchronization (SMR-ERD) derived from scalp electroencephalography, whereas movement direction was controlled by eye gaze. To manipulate kinesthetic congruence between MI and avatar gait, participants performed imagery of jogging (congruent condition) or sustained right-hand opening (incongruent condition). Subjective embodiment toward the controlled avatar was assessed using questionnaires. Main results. Participants reported significantly greater ownership in the congruent condition than in the incongruent condition. Although jogging imagery elicited weaker MI-related SMR-ERD than hand-opening imagery, no statistically significant condition differences were detected in course completion rate or time remaining upon completion. Significance. These findings suggest that, in multimodal BCIs, neural signal strength and embodied experience may dissociate, and that imagery–action congruence should be considered alongside conventional decoding metrics. By showing that action-congruent imagery enhances avatar ownership despite weaker neural signatures, this study motivates a shift from signal-centric toward context-aware BCI design for more intuitive and embodied interaction.