Four State Brain Machine Interface Design using Functional Link Networks

The ability of an individual to control his EEG through imaginary motor tasks enables him to control devices through a brain machine interface [BMI]. BMI provides a direct link between the human brain and devices such as wheelchair and hand prosthesis bypassing the biological channels (peripheral nerves) for control. BMI are essentially designed to provide mobility to people with severe motor disabilities. This paper presents a four-state BMI design for controlling a power wheelchair. Electroencephalogram [EEG] signals acquired during motor imagery for left and right hand movements are used to classify the four controls. The BMI is designed using a Functional Link Neural Classifier [FLNN]. The performance of the four-state BMI is tested with three feature sets. From the results it is observed that the performance of the BMI is better for the FLNN model using MEIG features with an average efficiency of 93%.