Treffer: Research on combining motor imagery and somatosensory attentional orientation to enhance BCI performance.

In this study, we propose a motor imagery(MI) method based on Somatosensory Attentional Orientation(SAO) to enhance the performance of MI based brain-computer interfaces (BCI). In this BCI system, participants perform unilateral hand MI tasks while maintaining attention to the corresponding hand, as if the wrist skin is actually receiving tactile stimulation(TS). A total of 44 participants were recruited and randomly divided into the experimental group(SAO and MI joint group, SMI group) and control group(MI group). The MI group performed right hand MI tasks, and two sessions were conducted, the content of the two experiments was identical. Each session was divided into two stages: the first stage including 1 run was the right hand MI mental task with TS on the right wrist, and the second stage including 6 runs was the right hand MI mental task without TS . For SAO group, first session was the same with the MI group. However, the second stage for SAO group was the right hand MI mental task with SAO. Compared with the first session, the performance in the first session was comparable between the MI group and SMI group, indicating similar MI abilities in both set of participants. For SAO group, A 6.5% performance enhancement was observed in the second session relative to the first session(p < 0.05). However, no significant improvement was observed in the MI group(p > 0.05), indicating no evidence of learning effect. EEG topographic mapping demonstrated robust bilateral hemispheric engagement when right hand MI mental task was performed for MI group. While in the SAO mental task, EEG exhibited clear hemispheric lateralization. This paradigm combining attention mechanisms with MI restructures the bilateral control modality inherent in conventional MI paradigms. As SAO paradigm engages endogenous cognitive processes, this approach augments corticomotor excitability during MI task, thereby improving BCI control performance.
(© 2026 IOP Publishing Ltd. All rights, including for text and data mining, AI training, and similar technologies, are reserved.)