Detection of Cognitive Fatigue Based on Mathematical and Auditory Tasks using Gamma Band of EEG Signals

Abstract

Cognitive fatigue is a discontinuous inability to maintain the existing cognitive performance and is a psycho-biological condition that occurs due to prolonged activities or working under stress. Cognitive fatigue causes an increase in errors, the emergence of various security vulnerabilities, and a decrease in performance. In this study, cognitive fatigue was tried to be determined by using EEG signals, which provide advantages in terms of use-transportation. Experiments were carried out with a total of 8 participants using the paradigm created for the detection of cognitive fatigue and EEG signals were recorded. Using the recorded EEG signals, the effects of different brain regions, different frequency bands, and different EEG lengths on the classification of cognitive workload were investigated. In addition, band power of EEG signals in situations with resting and cognitive workload were compared graphically. With the artificial neural network algorithm, the highest 99.49% classification accuracy was obtained by using the band power of the gamma frequency of all electrodes and the 5-second-long EEG segments.