Non-immersive Versus Immersive Extended Reality for Motor Imagery Neurofeedback Within a Brain-Computer Interfaces

A sensory feedback was employed for the present work to remap brain signals into sensory information. In particular, sensorimotor rhythms associated with motor imagery were measured as a mean to interact with an extended reality (XR) environment. The aim for such a neurofeedback was to let the user become aware of his/her ability to imagine a movement. A brain-computer interface based on motor imagery was thus implemented by using a consumer-grade electroencephalograph and by taking into account wearable and portable feedback actuators. Visual and vibrotactile sensory feedback modalities were used simultaneously to provide an engaging multimodal feedback in XR. Both a non-immersive and an immersive version of the system were considered and compared. Preliminary validation was carried out with four healthy subjects participating in a total of four sessions on different days. Experiments were conducted according to a wide-spread synchronous paradigm in which an application provides the timing for the motor imagery tasks. Performance was compared in terms of classification accuracy. Overall, subjects preferred the immersive neurofeedback because it allowed higher concentration during experiments, but there was not enough evidence to prove its actual effectiveness and mean classification accuracy resulted about 65%. Meanwhile, classification accuracy resulted higher with the non-immersive neurofeedback, notably it reached about 75%. Future experiments could extend this comparison to more subjects and more sessions, due to the relevance of possible applications in rehabilitation. Moreover, the immersive XR implementation could be improved to provide a greater sense of embodiment.