Radha Nila Meghanathan¹, Stefan Pollmann^1,2; ¹University of Magdeburg, ²Center for Behavioral Brain Sciences, Magdeburg

Gaze paths, like the images that prompt them, are represented in the cortex. Previous research from our lab has shown that gaze paths associated with faces and houses are distinctly represented in the fusiform face area, the parahippocampal place area, the superior parietal lobule and frontal eye fields. Although face and house gaze paths differ spatially, they encode both location and order information. Do neural representations of gazepaths contain both spatial and temporal information? To investigate whether and where location and order information in gaze paths are represented in the cortex, first, we recorded eye movement patterns while participants viewed images of faces and houses. Second, we recorded fMRI data while a different sample of participants followed these gaze paths by tracking dot sequences presented on a gray background in both ordered and shuffled sequence. Using multivariate pattern classification analysis with a linear support vector machine estimator on a brain parcellated according to the Schaefer atlas, we could distinguish activation patterns elicited by face and house-related gaze sequences in ventral occipitotemporal regions, specifically, the fusiform, lingual and occipital gyri, for both ordered and shuffled gaze paths. In contrast, along the borders of the horizontal and descending segments of the left intraparietal sulcus, we were able to classify face from house activation patterns only for ordered gaze paths. Notably, eye movements recorded while participants tracked dot sequences showed better dot tracking for ordered than shuffled sequences. Our findings show that the order of gaze paths is represented in the dorsal stream, while location information is encoded in the ventral stream. Ventral occipitotemporal cortex, which is involved in shape perception, may process predominantly spatial fixation locations but not temporal order. In contrast, dorsal regions, involved in visuomotor control, may use order information to predict upcoming saccade goals.