Abstract:
Brain-computer interfaces (BCI), in the last decade have emerged as a promising tool to predict
the user’s intent accurately. One of the most important application domains for the BCI systems
is their usage for the rehabilitation of stroke patients. Stroke is caused due to the disturbance in
blood flow to the brain resulting in the loss of brain functionalities. As a consequence the
affected parts of brain are not able to perform the normal functionality that might result in the
inability to move one or more limbs. Human brain provides the command for limb movements
and completes the process when it receives afferent feedback from muscles after movement. In
stroke patient this feedback is missing due to affected muscles. Stroke rehabilitation is the
process of retraining the brain and muscles to perform normal activities. The movement
intentions can be detected from the stroke patients using brain signal via electroencephalography
(EEG). The EEG signal contains movement related cortical potential signals (MRCPs) which is
used to detect the movement or intention to move and according to the intent an external stimuli
can be provided to the limb using the actuators to help the patient perform the desired task. In
our project, we are going to use the EEG signals to detect the intention and intensity of the
movements of the stroke patients to be used for rehabilitation purposes. The signals are acquired
through Emotiv and processed on ‘Raspberry pi’, according to the classified result the output
signal will be provided to a dc motor connected to the prototype of a bionic limb.
