Individual differences in eye movements are associated with individual differences in face perception : evidence from behavioral performance and neural responses

Abstract

Although all faces have similar configurations of features, individuals show different patterns of eye movements when viewing faces. Some people have a preference to scan the eyes, while others avoid looking at the eyes but fixate on the nose and mouth. However, it is poorly understood whether individual differences in eye movements indicate (or are indicators of) different abilities and neural mechanisms to process faces. In the thesis, I compared individual differences of eye movements in multiple aspects of behavioral performance and neural responses. In Study 1, I compared the holistic face processing between those two eye-movement groups. Results showed that the upper-focused group (preferring the eyes) showed a larger magnitude composite face effect when they were asked to match upper halves of faces than to match lower halves of faces. In contrast, the lower-focused group (preferring the nose and/or mouth) showed a comparable magnitude composite face effect between matching upper halves of faces and lower halves of faces. These findings suggest that eye-movement patterns tune the holistic face processing as measured by the composite face effect. Since eye-movement patterns modulated holistic face processing as a function of the target half of faces, and holistic processing has been considered as the underlying mechanism of face perception, in Study 2 and Study 3, I explored whether eye-movement patterns modulated face identification performance, and the related neural responses, as a function of fixation location on faces. Participants were asked to complete face identification tasks with restriction of the initial fixation location onto the eyes or the nose of faces. The upper-focused group performed better in the eye-fixation condition than in the nose-fixation condition, whereas the lower-focused group showed no significant difference between eye-fixation and nose-fixation conditions. Event-related potentials (ERPs) revealed that the upper-focused group elicited a larger N170 in the eye-fixation condition than in the nose-fixation condition, whereas the lower-focused group elicited similar N170 amplitude between eye-fixation and nose-fixation conditions. Moreover, the P1 was larger in the nose-fixation condition than in the eye-fixation condition for both eye-movement groups, and the P1 difference between fixation conditions was bigger for the lower-focused group than the upper-focused group. These findings suggest that the N170 reflects face identification performance. Since the upper-focused group has the preference to scan eyes directly while the lower-focused group appears to perform a global looking strategy, I wondered whether the upper-focused group and the lower-focused group have advantages of using high spatial frequencies (HSF) information and low spatial frequencies (LSF) information of faces, respectively. In Study 4, results showed that both eye-movement groups identified HSF faces better than LSF faces, and the initial fixation location on HSF faces modulated the P1 and the N170 for the lower-focused group and the upper-focused group, respectively. The results together showed that individual differences in eye movements are associated with different patterns of behavioral performance and neural mechanisms underpinning face perception. These studies suggest that the approach of individual differences in eye movements is a powerful method to exploit face perception and other cognitive processes.