PARIETAL CODING OF OBJECT-BASED SACCADES: TEMPORAL ASPECTS

P. N. Sabes^1,2, B. Breznen¹, and R. A. Andersen¹

¹Biology Division, 216-76 Caltech, Pasadena, CA 91125, U.S.A.,
²Sloan Center for Theoretical Neurobiology, Salk Institute, La Jolla, CA 92037, U.S.A.

A macaque monkey was trained to saccade to a cued location on an abstract object: a red dot briefly cued a part of the object while the monkey fixated its center, and when the object reappeared in a new orientation and/or position, the monkey made a saccade to the previously cued portion of the object. Two versions of the experiment were randomly interleaved across trials: either the object was moved visibly in front of the animal or it was extinguished for a brief interval before reappearing at the new location. We recorded neurons in the posterior parietal cortex of the monkey. The temporal course of the neural coding was explored. A population vector analysis was used, across trials and cells, to infer the evolution of area's movement plan over the course of a trial. In the visible rotation task, the population tracked the previously cued location of the object as it moved, as if the area were performing a pursuit of the remembered location. In the non-visible rotation task, the population vector shifted rapidly from the cue direction to the movement direction after the reappearance of the object. This transition appeared to be a discrete change in plan, with a marked decoherence of the population tuning during the transition. We also investigated the temporal relationship between the population vector transition and the monkey's behavior. Trials were sorted into groups by reaction time, and the analysis was repeated for each group. We found that the completion of the transition was temporally locked to the saccade onset time rather than the reappearance of the object. Furthermore, there was an inverse correlation between the length of the population vector code and the reaction time, suggesting that increased coherence in posterior parietal tuning can facilitate faster movement initiation. This work was supported by the NEI and the Sloan Foundation.