Abstract
ABSTRACTFace cognition plays a significant role in social interaction. The typical stimulus used to study face cognition mechanisms is a rapid serial visual presentation (RSVP). During the RSVP task, the brain response called event-related potential (ERP) is evoked when a person recognizes a target image. Many trials are required to average and obtain a clean ERP to interpret the cognitive mechanism behind the ERP response. However, increasing the trial number can cause fatigue and affect evoked ERP amplitude. This paper adopts a different perspective; machine learning might extract a meaningful cognitive result that reveals the face cognition mechanism without directly focusing on the characteristic of the ERP. We implemented an xDAWN covariance matrix method to enhance the data quality and a support vector machine classification model to predict the participant’s event of interest using ERP components evoked in the partial face cognition task. The effect of face components and the physical response was also investigated to explore the role of each component and find the possibility of reducing fatigue caused during the experiment. We found that the eyes were the most effective component. Similar statistical results were obtained from full face and partial face with eyes visible in both behavioral response and classification performance. From these results, the eye component could be the most crucial in face cognition. So, there could be some similarities in the face cognition mechanism of the full face and the partial face with eyes visible, which should be further investigated using ERP characteristics.
Publisher
Cold Spring Harbor Laboratory