Multi class sensory motor imagery EEG classification for brain computer interfacing

Abstract

The few decades old field of electrophysiology on neural signal events in single cell recording can be acknowledged that there is production of far field potentials containing near synchronous field patterns might reach at the scalp electrodes. These neural signals then can be analyzed and converted into the control signals for computers and other electronic devices. Electroencephalography (EEG) is a method to acquire these neural signals from the scalp of human brain. EEG signals are simple, economical and have high temporal resolution properties. These properties make it advantageous to use it in wide range of diverse applications. The event related synchronization and event related desynchronization (ERS/ERD) pattern present in EEG during sensory motor imagery (SMI) process over the cortical area is an important feature to take Brain Computer Interface towards realistic approach. The contribution of classification accuracy for brain computer interfacing is strongly dependent on the orchestration of the BCI sub- components like signal processing, feature extraction and selection, and classification of EEG signals. The brain computerinterfacing involves multidisciplinary concepts which bring great technological challenges like less accuracy, non-stationarity in brain signals, removal of artifacts, poor signal-to-noise ratio etc. A lot of work has been done by several means but further improvement is required in understanding the instructions hiding inside the electroencephalogram (EEG) signals and converting them into control commands with better efficiency. This research aims to analyze EEG signals to detect cognitive activity and to provide novel solution in developing EEG based brain computer interface with high accuracy. To achieve this goal, this research put forward three phases in proposing a novel framework for more accurate EEG based brain computer interfacing. To validate the proposed approach, it is compared with the single classifier approach which includes kNN, SVM, SVM (RBF) and Decision tree to kn