Abstract
Recent research has identified two common eye movement patterns during face perception: holistic and analytic. People using holistic patterns fixate mainly around the face center, whilst those using analytic patterns look at individual facial features in addition to the face center and exhibit more frequent gaze transitions. Analytic patterns yield better face recognition performance than holistic patterns. However, it remained unclear whether these two gaze patterns are associated with different cognitive and neural mechanisms in face processing. The present study investigated the neural correlates of these two patterns through functional magnetic resonance imaging (fMRI). Participants performed face recognition memory tasks with eye tracking and brain imaging. In addition, they performed three cognitive tests, including planning, verbal and spatial working memory tests. Eye movements were analyzed with a hidden Markov model based approach (Chuk, Chan, & Hsiao, 2014). We trained one model per participant and clustered the individual models into two patterns according to their similarities. For each individual model, the likelihood of being identified as each pattern was computed. The holistic/analytic likelihoods were correlated with the blood-oxygen-level-dependent fMRI response at the whole brain level. Consistent with previous studies, our results showed that analytic patterns were associated with better recognition performance. More importantly, a more holistic pattern was associated with more activation in perceptual regions including the inferior occipital gyrus and lateral fusiform gyrus. In contrast, a more analytic pattern exhibited more activation in areas important for top-down control of visual attention, including the frontal eye field and intraparietal sulcus. In addition, participants who adopted a more analytic pattern had better spatial working memory. These results suggest that analytic patterns are associated with better spatial working memory and recognition performance, and this advantage may be a consequence of more engagement of top-down control of visual attention.
Meeting abstract presented at VSS 2017