Multi-class Motor Imagery classification techniques based on different Deep Learning algorithms

Ischaemic or haemorrhagic strokes and injuries are the main cause of motor or cognitive disability. Patients become incapable to performing the most common daily activities. Treatment and therapeutic aids are an expensive solution. Rehabilitation therapies based on the growing use of BCIs have proven to be effective. BCIs of this type apply a repetitive mechanical or electrical stimulus to a impaired body part. The stimulus must be applied within a time range from the imagined movement of the dysfunctional limb. The aim is to restore a connection between the central nervous system and the area of the body that has lost its function. The application of a stimulus, mechanical or electrical, within a time window, facilitates motor recovery. Because the stimulus to be applied, a BCI must be able to recognise the kind of movement. The aim of this thesis is the classification of two imaginary movement using different systems based on the most innovative deep learning techniques.