Date of Award
Master of Science
Electrical and Computer Engr
Gavin M. Bidelman
Cognitive load refers to the amount of used working memory resources, which is limited in both capacity and duration. Predicting cognitive load from raw electroencephalogram (EEG) recordings remains a challenge because of the high degree of noise due to technical variations in the recording process and the multi-factorial nature of the mapping between the EEG data and cognitive load. We present parameter-optimized deep convolutional neural network (CNN) models to predict four levels of cognitive load. We use eigenspace-based bootstrap sampling and Generative Adversarial Network (GAN) to address the issue of noise and a small number of samples in EEG. We transform time-series signals into a spatial-spectral representation called Topomap, which maintains both spatial and spectral information embedded in EEG recordings. We use two different EEG data representations for cognitive load prediction. First, we use power spectral densities of three individual frequency bands (Theta, Alpha, Beta) to create the topomap. Second, we combine all three bands to develop a composite representation. We performed empirical evaluations to determine the role of individual frequency bands in predicting cognitive load. The prediction accuracy of CNN models built using Theta, Alpha, Beta bands, and composite representation are 89%, 89%, 91%, and 92%, respectively. The results suggest that the Beta band has the most predictive power and composite representation produces higher accuracy than the individual frequency bands.
Dissertation or thesis originally submitted to the local University of Memphis Electronic Theses & dissertation (ETD) Repository.
Havugimana, Felix, "A Deep Generative and Discriminative Approach in Modelling Spatial-spectral Dynamics of Varying Cognitive Load from EEG Recordings" (2021). Electronic Theses and Dissertations. 2368.