Classification of cognitive fatigue with eeg signs

Cognitive fatigue is caused by excessive activity of the brain. It can be caused by other factors such as prolonged, high-risk stress work and insomnia. Cognitive fatigue leads to decreased productivity and increased safety risks. In this thesis, it was aimed to determine cognitive fatigue quickly and accurately without depending on subjective data. For this, CogBeacon data set was used. This dataset was renewed from 19 female and male participants, including 76 cognitive tasks in two sessions, with a 4-electrode MUSE EEG headset. The raw EEG collected was randomly assigned. The classification process was made by calculating the 12 different coefficients generally used in the analysis of audio signals. Root mean square, log detector, mean absolute value, modified mean absolute value 1, modified mean absolute value 2 and enhanced mean absolute value have been tested as attributes. The best results were obtained as a result of the classification processes made with KNN and SVM algorithms after calculating MFCC and GTCC. In the classification made after the calculation of the MFCC, the classification accuracy with KNN is 78.72% and SVM classification accuracy as 78.43%. In the classification made after the calculation of GTCC, KNN classification accuracy was calculated as 78.08% and SVM classification accuracy as 76.61%. Keywords: Cognitive Fatigue, Electroencephalography (EEG), Mel Frequency Cepstral Coefficient (MFCC), Gammatone Cepstral Coefficient (GTCC