Towards a BCI Based on Action Observation and Kinesthetic Motor Imagery for Neurorehabilitation

This work introduces the proposal of an EEG signal acquisition protocol for the creation of a Brain-Computer Interface (BCI) based on Kinesthetic Motor Imagery (KMI) and Action Observation (AO) for neurorehabilitation of post-stroke patients, as a strategy to support the challenge of improving the effectiveness of BCI systems. The protocol considers physical rehabilitation strategies, a Robotic Hand Orthosis (RHO), and a low-cost EEG signal acquisition platform composed of an EEG cap with electrodes placed on the most significant brain locations for discriminating Motor Imagery (MI) tasks, specifically Open Hand MI (OH-MI) and Close Hand-MI (CH-MI). Our initial findings indicate that OH-MI and CH-MI can evoke discriminant changes in ERD values in Mu band, specifically in CP3 and C3 brain locations. Also, OH-MI class shows more intensity of brain activation in this band in comparison with the CH-MI class. Based on this, we can consider that this protocol could be taken into account for rehabilitation of post-stroke patients using BCI systems and for neuroprothesis training, as the resulting topographic maps related to the Event-Related Dessynchronization/Synchronization (ERD/S) pattterns show the feasibility to differentiate between both classes of hand movements. Thus, these results hold significant promise for applications in post-stroke rehabilitation as well as neuroprothesis training, as the ability to control hand movements through this technology has the potential to greatly benefit individuals in regaining motor function.