A Setup for Lower-Limb Post-stroke Rehabilitation Based on Motor Imagery and Motorized Pedal

This work presents a low-cost solution for a neuro-rehabilitation system of stroke patients through a motorized pedal, as pedaling has the potential to provide a high number of flexion and extension repetitions of the lower limb. Stand-alone hardware and software were developed to control the motorized pedal, and the motor imagery recognition is done through a BCI, which analyzes EEG patterns on the motor region, related to feet movements and rest state. From this BCI, an avatar is triggered into an immersive virtual reality environment together the motorized pedal. sEMG signals collected on Rectus femoris (RF), Biceps femoris (BF), Tibalis Anterior (TA) and Gastrocnemius muscles (GM) allow identifying the muscle onset and offset when a force is exerted on the pedals, and force sensors are used to generate biofeedback for the subject. As a first result of this setup, the maximum accuracy reached with our BCI was 94.41%.