EEG-induced Probabilistic Prediction of the Color-Pathways in the Brain using Dempster-Shafer Theory

• Published in: 2019 International Conference on Wireless Communications Signal Processing and Networking (WiSPNET) pp. 339 - 343
• Main Authors: Laha, Mousumi, Hazra, Priti, Konar, Amit, Rakshit, Pratyusha
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.03.2019
• Subjects: Dempster-Shafer theory; frame of discernment; Short term Fourier transform; Signaling color pathway
• DOI: 10.1109/WiSPNET45539.2019.9032841

The paper proposes a novel approach to predict the signaling pathways in the brain in response to color stimuli by the following four steps. First, the EEG response of different brain lobes to a (monochromatic) colored light stimulus is recorded, and the acquired EEG signals are filtered using an active band pass filter to eliminate the effect of artifacts. Second, short-term Fourier transform is taken on a moving window placed over the EEG signals to extract the power in each window with an ultimate aim to determine the time-point (centre of one specific window) on the EEG signal having the maximum power. Each selected brain lobe is labeled with a time-point corresponding to the maximum power in the EEG signal acquired from that lobe. Third, a brain signaling pathway of an individual subject in response to a color stimulus is determined from multiple experiments on the same subject with the same stimulus, and a color-pathway graph is developed where nodes and edges respectively indicate brain lobes and transition pathway from one lobe to the next. A probability estimate of each transition (edge) is determined from the results of multiple experiments. Finally, the color-pathway graphs obtained from different individuals for the same color stimuli are fused using Dempster-Shafer theory to have a generic consensus about the signaling pathway in the brain due to retinal response to monochromatic color stimulus. The predicted pathways for a specific color (red/green/blue) are found to be unique and time-invariant across 30 experimental subjects. This method is useful for the diagnosis in abnormal color pathways for people with psychological disorder.