Analysis Of Eeg Using Ica To Rehabilitate Completely Locked In Syndrome Patients
Abstract - Millions of people worldwide are permanently disabled either due to stroke or injury to the spinal cord.
Rehabilitation of the disabled is one of the major challenges in the present world. There are many undergoing researches in
the field of Brain Computer Interface (BCI) to establish an interaction between the human neural system and machine to
enable the disabled to communicate and control devices for better living. Establishment of muscular movement in humans
who have lost motor control due to injury or paralysis by the interpretation of brain rhythm is a very interesting and ongoing
front end research. This paper proposes the use of EEG signals for controlling the movement of muscles and subsequent
rehabilitation of the subject. Here the EEG signals from brain activity while imagining motor movements are captured. The
required features for classification of motor movements is done using Independent Component Analysis by Entropy Bound
Minimization (ICA_EBM), a method for finding underlying factors or components from multivariate (multi-dimensional)
statistical data that are both statistically independent and non-Gaussian. The processed motor imagery signals are classified
using feed forward back propagation Neural Networks (NN). The EEG processed and classified signals act as stimulus to the
BCI, developed using Arduino, providing rehabilitation for the disabled.
Index terms - Brain Compute interface, EEG, Artificial neural network, ICA