Another look at the Müller-Lyer illusion: different gaze patterns in vision for action and perception.

Following Goodale and Milner's [Goodale, M. A., & Milner, A. D. (1992). Separate visual pathways for perception and action. Trends in Neurosciences, 15(1), 20-25] proposal to distinguish the dorsal and ventral systems on basis of the functional demands they serve (i.e., action and perception), a vast literature has emerged that scrutinized if the dorsal and ventral systems indeed process information into egocentric and allocentric codes in accordance with their respective functions. However, a corollary of Goodale and Milner's original proposal, that these functional demands also impose different constraints on information detection, has been largely overlooked. In the present study, we measured gaze patterns to investigate how information detection for action and perception differs. In two conditions, participants (N=9) grasped or made a manual estimate of the length of a shaft embedded in a Müller-Lyer configuration. The illusion significantly affected the manual estimates, but not the hand aperture during grasping. In line with these behavioral findings, significant differences in gaze patterns were revealed between the two tasks. Participants spent more time looking at areas that contain egocentric information (i.e., centre of the shaft) when grasping as compared to making a manual length estimate. In addition, participants, made more gaze shifts (i.e., especially between the two areas surrounding the shaft endpoints and including the arrowheads) when making the manual length estimate, enabling the pick up of allocentric information. This difference was more pronounced during task execution as compared to task preparation (i.e., before movement onset). These results support the contention that the functional distinction between the dorsal and ventral systems is not limited to the processing of information, but also encompasses the detection of information.