Target selection for visually guided reaching in macaque

Abstract

We examined target selection for visually guided reaching in monkeys using a visual search task in which an odd-colored target was presented with distractors. The colors of the target and distractors were randomly switched in each trial between red and green, and the number of distractors was varied. Previous studies of saccades and attention have shown that target selection in this task is easier when a greater number of homogenous distractors is present. We found that monkeys made fewer reaches to distractors and that reaches to the target were completed more quickly when a greater number of homogenous distractors was present. When the target was presented in a sparse array of distractors, reaches had longer movement durations and greater trajectory curvature. Reaching errors were directed more often to a distractor adjacent to the target, suggesting a spatially coarse-to-fine progression during target selection. Reaches were also influenced by the properties of trials in the recent past. When the colors of the target and distractors remained the same from trial to trial rather than switching, reaches were completed more quickly and accurately, indicating that color priming across trials facilitates target selection. Moreover, when difficult search trials were randomly intermixed with easier trials without distractors, reach latencies were influenced by the difficulty of previous trials, indicating that motor initiation strategies are gradually adjusted based on accumulated experience. Overall, these results are consistent with reaching results in humans, indicating that the monkey provides a sound model for understanding the neural underpinnings of reach target selection.

Figure 1

Schematic diagrams of a single target trial (A), and color-oddity search trials (B-E). In single target trials, a lone target is presented without distractors, while in search trials an odd-colored target is presented with 3, 5, 7, or 11 distractors.

Figure 2

Effect of the number of distractors on reaching performance. The upper panels show target selection error rate (proportion of reaches that end more the 5° from the target location) for each monkey as a function of the number of distractors in the search array. The middle panels show the total time as a function of the number of distractors for each monkey. Total time is defined as the time from stimulus onset to finger contact with the target location. For comparison, the error rate and total time for single-target trials without distractors are shown at the extreme left of the plots. In the lower panels, total time was decomposed into initiation latency (the time from target onset to finger lift-off) and movement duration (time from lift-off to contact with the touchscreen). For comparison, latency and duration for single-target trials without distractors are shown at the extreme left. Error bars indicate standard error of the mean in this and all subsequent Figures.

Figure 3

Reach endpoints of search trials with 3, 5 and 7 distractors for Monkey H (upper) and 3, 5 and 11 distractors for Monkey J (lower). The endpoints are normalized by a rotation, such that the correct target location is always represented at an angle of 0° and amplitude 1. For instance, in the 3-distractor case, the distractors are located at 90°, 180°, and 270°. Errors were more likely to be directed to one of the distractors near the target than to the distractor furthest from the target location.

Figure 4

Effect of color priming on reaching performance. Target selection error rate (upper), total time (middle), and movement initiation latency and duration (lower) for each monkey are plotted as a function of the number of consecutive same-color trial in the past. Reaching performance improves when the color of the target remains the same from trial to trial. One on the abscissa denotes trials in which the target color differed from its color in the previous trial; two denotes trials in which the target color was the same as in the previous trial, but differed from its color two trials previously, and so on.

Figure 5

Effect of distractors on reach trajectories. A, B show examples of reach trajectories in single target (left) and 3-distractor search (right) for Monkey H. These trajectories are tracked in three dimensions, but, for clarity, only the components in the plane of the touchscreen monitor are shown. Trajectories associated with each target location are depicted by distinct colors: green (45°), blue (135°), red (225°) and black (315°). The scale bar indicates 1cm. C shows a plot of radial direction vs. time during the movement for Monkey J in the single-target task. Radial direction has been normalized such that the target direction corresponds to a direction of 0°. D, E show similar plots of radial direction vs. time in 3-distractor search, for movements with durations in the lower third (short duration: D) and upper third (long duration: E) of the distribution.

Figure 6

Correlation between movement duration and maximum curvature in search. Maximum curvature is plotted against movement duration for 3-distractor search trials in each monkey.

Figure 7

Movement initiation latencies of randomly-intermixed search and single-target trials as a function of the number of consecutive same-type trials. One on the abscissa indicates trials that differed in type from the previous trial. Initiation latencies gradually become shorter with more consecutive single-target trials, and longer with more consecutive search trials.