Detecting Cognitive Features of Videos Using EEG Signal

Electroencephalography (EEG) emerged as a highly relevant signal to human emotion, brain diagnosing and brain–computer interfaces (BCI) applications. In this paper, the EEG signal is used to evaluate the cognitive response of subjects during watching test video clips. The measurements are performed with 25 subjects using eight channels while simultaneously running the video clips. The β and γ waves of the EEG signal are used to extract the features that represent the evoked activity in each group of frames using the Peak-Over-Threshold (POT) technique. Significant EEG patterns are derived from the time-correlated measurements, which can be related to the subjects’ interests. In addition, the conjunctions that represent the occurrence of segments-of-interest in more than one channel are determined. The results show that ~15% of the segments attracted the attention of the viewers in each test video clip. Such a technique can potentially be implemented in neuromarketing analysis or to develop a new video compression technique that depends on the human cognitive system.