Visual Landmarks Influence Eye-Hand Coordination in Memory-Guided Reaches
In the authors' words
The human visual system optimally weighs egocentric cues (e.g., object location relative to the eye) and allocentric cues (e.g., object location relative to other landmarks) to aim actions. The neural mechanisms of this process have been studied in the primate gaze system, but it is not known how these results extend to eye-hand coordination and the reach system. Here, reach and gaze (eye-in-space) data were collected from two head-unrestrained Macaca mulatta (M1, M2) trained to reach toward remembered visual targets on a touch screen. During the delay, animals either saw nothing (no-landmark), a pattern of 'dots' around the target (stable landmark), or a similar landmark that reappeared at a shifted position (landmark shift). In most conditions, increased delay times (up to 1200 ms) reduced reaction times, with minimal effects on gaze and reach errors. In the temporal domain, landmarks reduced gaze reaction time, increased the time to final hand touch, and reduced gaze-hand coupling. In the spatial domain, the stable landmark tended to improve reach accuracy and precision but reduced landmark-centered gaze-hand coupling. Landmark shifts increased errors in pooled datasets, due to weighted (and modestly correlated) shifts in both reach and gaze in the same direction. The 'allocentric weights' for reach were also modulated by initial gaze position. Overall, these data suggest that visual landmarks influence the timing, accuracy, and coordination of gaze / hand movements, likely through visual feedback during reach, and these effectors share a similar mechanism for optimal integration of egocentric and allocentric visual cues.
Appeared: Wednesday, November 18. bioRxiv. Preprint, not yet peer-reviewed.