EEG-Based Emotional Video Classification via Learning Connectivity Structure

Electroencephalography (EEG) is a useful way to implicitly monitor the user's perceptual state during multimedia consumption. One of primary challenges for practical use EEG-based monitoring achieve satisfactory level accuracy in EEG classification. Connectivity between different brain regions an important property classification EEG. However, how define connectivity structure given task still open problem, because there no ground truth about should be order maximize performance. In this paper, we propose end-to-end neural network model emotional video classification, which can extract appropriate multi-layer graph and signal features directly from set raw signals perform using them. Experimental results demonstrate that our method yields improved performance comparison existing approaches where manually defined structures are used. Furthermore, show extraction process reliable terms consistency, learned make much sense viewpoint perception occurring brain.