Neurons in Area LIP differentially encode saccadic targets and visual distractors

Purpose. We attempted to determine how effectively neurons in Area LIP separate signals related to saccadic targets from signals related to visual distractors. To accomplish this, distractor utility was manipulated by making either fixation offset or distractor offset a movement initiation cue. Methods. Single neurons in Area LIP were studied in two head-fixed, awake-behaving rhesus monkeys using standard physiological techniques. First, the response field of each unit was measured at a 2° resolution using a delayed saccade task. Next, unit responses were recorded while animals performed a selection task. Subjects initially fixated a central yellow LED for 200-800 msec. Two eccentric yellow LEDs were then co-illuminated for 200-800 msec, one above and one below the horizontal meridian. The fixation LED then changed color to either red or green. Red indicated that the monkey would be rewarded for aligning gaze with the upper LED (±6°) after the offset of the fixation LED 200-800 msec later; similarly, green specified the lower LED as the saccadic goal. For each unit, one of the two potential targets was fixed outside the response field of the unit while the location of the other potential target was varied randomly, at a 2° resolution, throughout the visual field. Selection-related response fields were then constructed from the data for each unit by plotting firing rate as a function of the horizontal and vertical amplitude of the movement and as a function of the horizontal and vertical position of the visual distractor. A selectivity ratio was calculated as the peak firing rate for targets divided by the peak firing rate for distracrors. Unit responses were also recorded while subjects performed a distributed selection task. This task was identical to the selection task except that the offset of the visual distractor provided the cue to initiate a saccade at short latency. The effects of distractor utility on unit responses to targets and distractors were assessed by comparing selectivity ratios for each unit across the two tasks.Results. Most units were well-tuned spatially (<5° half-width). On the selection task, most units were selective for saccadic targets over irrelevant visual distractors (selectivity ratio >1) throughout the specification and pre-movement intervals. On average, LIP units responded about twice as well for saccadic targets as they did for irrelevant visual distractors after the specification cue. On the distributed selection task, distractor utility significantly modulated target/distractor selectivity for some units. In this condition, some units increased their selectivity for saccadic targets in the presence of a utile visual distractor, whereas others showed decreased selectivity. Most units were equally selective on both tasks.