Date of Award

Spring 2013

Document Type

Thesis

Degree Name

Master of Science in Biomedical Engineering - (M.S.)

Department

Biomedical Engineering

First Advisor

Richard A. Foulds

Second Advisor

Sergei Adamovich

Third Advisor

Mesut Sahin

Abstract

Neurorehabilitation has recently been augmented with the use of virtual reality and rehabilitation robotics. In many systems, some known volitional control must exist in order to synchronize the user intended movement with the therapeutic virtual or robotic movement. Brain Computer Interface (BCI) aims to open up a new rehabilitation option for clinical population having no residual movement due to disease or injury to the central or peripheral nervous system. Brain activity contains a wide variety of electrical signals which can be acquired using many invasive and non-invasive acquisition techniques and holds the potential to be used as an input to BCI. Electroencephalogram (EEG) is a non-invasive method of acquiring brain activity which then, with further processing and classification, can be used to predict various brain states such as an intended motor movement. EEG provides the temporal resolution required to obtain significant result which may not be provided by many other non-invasive techniques. Here, EEG is recorded using a commercially available EEG headset provided by Emotiv Inc. Data is collected and processed using BCI2000 software, and the difference in the Mu-rhythm due to Event Related Synchronization (ERS) and Desynchronization (ERD) is used to distinguish an intended motor movement and resting brain state, without the need for physical movement. The idea is to combine this user intent/free will with an assistive robot to achieve the user initiated, repetitive motor movements required to bring therapeutic changes in the targeted subject group, as per Hebbian type learning.

Share

COinS