|
|
Classification of mental workload using brain connectivity measure
Doležalová, Radka ; Kolářová, Jana (oponent) ; Ronzhina, Marina (vedoucí práce)
This thesis deals with possibilities of using EEG connectivity measures for automatic classification of mental workload levels. The theoretical principles of EEG recording and different measures of brain connectivity are discussed at the beginning. Two different levels of mental workload were evoked in healthy participants during real experiments. The course of experiment, processing of recorded EEG, as well as extraction of classification features from EEG based on some connectivity measures (such as cross-correlation, covariance, coherence and phase locking value), and automatic classification approaches (classification based on distance from average, 1-nearestneighbor searching and discriminant analysis) were then described. Obtained results were interpreted and discussed. The best classification accuracy (approx. 60,64%) was obtained using beta band of EEG recorded with 4 channels from different scalp, when features were classified with linear discriminant function.
|
|
|
Classification of mental workload using brain connectivity measure
Doležalová, Radka ; Kolářová, Jana (oponent) ; Ronzhina, Marina (vedoucí práce)
This thesis deals with possibilities of using EEG connectivity measures for automatic classification of mental workload levels. The theoretical principles of EEG recording and different measures of brain connectivity are discussed at the beginning. Two different levels of mental workload were evoked in healthy participants during real experiments. The course of experiment, processing of recorded EEG, as well as extraction of classification features from EEG based on some connectivity measures (such as cross-correlation, covariance, coherence and phase locking value), and automatic classification approaches (classification based on distance from average, 1-nearestneighbor searching and discriminant analysis) were then described. Obtained results were interpreted and discussed. The best classification accuracy (approx. 60,64%) was obtained using beta band of EEG recorded with 4 channels from different scalp, when features were classified with linear discriminant function.
|
host ::
RSS.