Dry versus Wet EEG electrode systems in Motor Imagery Classification

Abstract

Inês Domingos , Fani Deligianni 1 and Guang-Zhong Yang 1 The Hamlyn Centre for Robotic Surgery, Imperial College London, UK Abstract Motor imagery (MI) classification performance is important in developing robust brain computer interface environments for neuro-rehabilitation of patients and robotic prosthesis control. To bring this technology to everyday use relatively new EEG acquisition systems have been developed. These systems are highly portable, wireless and they are based on dry, active electrodes, which does not require the use of conductive gel. As a result they are more prone to interference via noise sources that are commonly around and their signal-to-noise ratio may be low. Here, we device a number of motor imagery tasks along with actual movements of the limbs and compare the classification performance of a dry 16-channel and a wet, 32channel, wireless EEG system. Our results demonstrate the feasibility of home use of dry electrode systems with a small number of sensors.