Identification of Cognitive Workload via EEG Based Brain Mapping in Real Time

Abstract

Electroencephalography (EEG) is emerged as a powerful tool in examining the cognitive workload of human. In earlier times, high-cost EEG systems were being used to acquire EEG signals which limit the scope of EEG but in recent times less costly EEG acquisition device has been accessible to acquire EEG signals which make it possible to reach to the masses for numerous applications such as assessing cognitive workload. The real time EEG signals have been acquired using external electrodes of RMS Super Spec 32 Channels EEG acquisition device. The cognitive workload has been induced by asking the subjects to perform mental arithmetic task of varying difficulty and solve audio puzzle. The acquired EEG signals have been preprocessed using Kaiser window based FIR filter to remove baseline noise. The pre-processed EEG signals have been further analyzed to differentiate between task difficulties over all participants by extracting time and frequency domain feature set. In time domain analysis, event relative potential distribution parameters across scalp region have been calculated. In frequency domain analysis, power spectral variations have been analyzed in MATLAB and EDF Browser Tool. The detailed analysis performed in EDF Browser software application as well as in MATLAB indicates that there is a decrease in frontal delta power (0.5-4Hz) during cognitive workload state as compared to relaxed state. Hence, it is observed that the proposed algorithm possesses the potential to identify cognitive workload levels by monitoring and characterizing brain signal variations via EEG.

Keywords: Electroencephalogram, Cognitive workload, event-related potential (ERP), spectral power, Real time, Brain Mapping.