During a memory-guided delayed saccade task a monkey's attention, as measured by an improvement in contrast threshold at the attended location, lies at the goal of the saccade, unless a distractor appears elsewhere in the visual field. The distractor attracts attention for several hundred ms, after which it returns to the saccade goal. Neurons in the lateral intraparietal area (LIP) respond to the onset of the saccade target, and maintain activity tonically during the memory delay period. They also respond phasically to a distractor flashed in the receptive field while a monkey plans a saccade elsewhere. We have previously shown that the time at which attention returns to the saccade goal from the distractor location can be predicted, on a monkey-by-monkey basis, by the crossing time at which the LIP population activity evoked by the distractor falls below the activity at the saccade plan. Here we show that this occurs at a similar time in the individual neurons that make up the population, despite the fact that the tonic level of delay activity varied greatly between cells. Neurons with higher delay period activity had less rapid decay of their phasic distractor response than neurons with lower delay period activity. Thus the population average of crossing time, and presumably its psychological sequela, are determined by some general idiosyncratic characteristic relating the level of delay activity with the decay rate of the distractor response, and not by differing distributions of individual cells with differing response characteristics which average out in the population.

This work was supported by the NEI, HFSP & MacDonnell Foundation.