A Novel Methodology to study the Cognitive Load Induced EEG Complexity Changes: Chaos, Fractal and Entropy based approach

• Published in: Biomedical signal processing and control Vol. 64; p. 102277
• Main Authors: Parbat, Debanjan, Chakraborty, Monisha
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.02.2021
• Subjects: Cognitive Neurodynamics; Electroencephalography; Fractal Dimension; Largest Lyapunov Exponent; Non-linear Dynamical Systems; Sample Entropy; Wavelet Denoising; Sample Entropy; Wavelet Denoising; Largest Lyapunov Exponent; Fractal Dimension; Cognitive Neurodynamics; Electroencephalography; Non-linear Dynamical Systems
• ISSN: 1746-8094; 1746-8108
• DOI: 10.1016/j.bspc.2020.102277

•Application of Dynamical Systems Theory to capture the complexity dynamics of the human brain while cognitive processing.•Chaos, Complexity and Entropy has been estimated over a customizable fixed time sliding window to represent the changing dynamics over a joint Time-Space topographical map.•The evolution of the states of the brain processes with and without cognitive loading has been appreciably distinguished viz. pre, during and post application of the task.•Classification of various cognitive states using SVM, Random Forest & Decision Tree classification Dynamic Systems Theory (DST) can provide both the conceptual framework and literal description of the underlying complexity dynamics associated with human cognition, specifically during information processing of the brain under the effect of an external stimuli. To study the complexity changes during cognitive loading of the brain using Largest Lyapunov Exponent (LLE), Higuchi Fractal Dimension (HFD) and Sample Entropy (SampEn) as a multiparametric signature of cognitive processing. The proposed methodology demonstrates joint Time-Space representation of the various Brain Rhythms under four different classes of Cognitive Tasks (Emotion, Focus, Memory and Problem Solving) given to four subjects. The raw EEG signal is acquired using a 19 channel EEG machine, denoised using Wavelet packet decomposition technique. Brain waves are extracted using the scalogram plot. The parameters are calculated for each channel over a 2 min analysis window sliding through the whole length. These parameters were able to classify between different cognitive states, such as Emotion, Focus, Memory and Problem Solving with an accuracy of 99%. Previous works haven’t addressed complexity changes during cognitive processing using DST. Earlier studies explain average topographical map of the brain for a fixed time window where as, we have presented the topographical map over a customizable fixed time sliding window. The cubic representation of the brain map containing non-linear parameters can prove to be a significant visualization tool for monitoring effects of cognitive loading using DST proponents as biomarker.