Neurorehabilitation Platform Based on EEG, sEMG and Virtual Reality Using Robotic Monocycle

A recent study in the literature showed that eight paraplegic patients with chronic spinal cord injury, who underwent 12 months of training in brain-machine interface (BMI), based on neurorehabilitation using a virtual system and a very high cost exoskeleton, experienced neurological enhancements in somatic sensation, as well as motor improvements. A possible low-cost solution is to use a robotic monocycle instead of an exoskeleton, since the exercise of pedaling a monocycle has the potential to provide a high number of flexion and extension repetitions of the lower limb in reasonable therapeutic time periods. The objective of this work is to develop a neurorehabilitation platform based on electroencephalography (EEG), surface electromyography (sEMG) and immersive virtual reality (IVR), and using a robotic monocycle to move the user’s legs. The monocycle is instrumented with inertial sensors placed on the pedals, which is used to measure the cadence developed by the user while pedaling, and a customized electronic board to control the monocycle according to the user’s motor imagery detected through a Brain-Computer Interface (BCI) using EEG. On the other hand, sEMG signals are collected from the rectus femoris, biceps femoris, tibialis anterior and trochanteric muscles, in order to allow the identification of their onset and offset. In addition, a serious game was designed to be used as part of the rehabilitation platform. As preliminary results, the developed BCI is able recognize motor imagery patterns related to feet movements and resting state with an average accuracy higher than 80%.