<p>Asynchronous brain-computer interfaces (aBCIs) using brain switches have demonstrated significant reliability and usability in discriminating a user’s control intention and resting state. However, there is a rare research demonstration of the BCI brain switch independent of visual feedback; the drawback of relying on visual channels might be disadvantageous in exploring the surrounding environment and inappropriate for some target users with severe disability. In this study, we proposed a vibrotactile-based brain switch using a virtual physical model to integrate the weak intention information to address the challenge without visual feedback. Specifically, two vibrators were worn on the left and right ankles to convey the system’s swing direction to the subject, and the subject completed the periodic motor imagery task according to the vibrotactile feedback direction until the intention information was integrated enough to trigger the brain switch. A control experiment with visual feedback and vibrotactile feedback was conducted, respectively. The results showed that the triggering time for visual feedback was 54±67 s, the triggering time for vibration feedback was 53±50 s, and the false triggering rate (or false positive rate, FPR) was 1.3±1.5 and 1.1±1.4 FP/h, respectively. These results show that the proposed brain switch framework based on virtual physical systems does not need to rely on the visual pathway and can provide good reliability and triggering speed. It sheds light on developing asynchronous visual independent brain-computer interface systems.</p>