Percentile schedules to shape touchscreen response locations in cynomolgus macaques.

Variations in the location of emitted responses are often critical in defining functional outcomes. Percentile schedules evaluate responses dynamically and mathematically against a set criterion that must be met for reinforcement to occur. Across five experiments, percentile schedules differentially selected shorter or longer distances between emitted response locations and an invisible, arbitrary target location on a touchscreen, in the absence of any stimulus cues other than the differential reinforcement of each response. Experiment 1 selected responses that minimized distances from the target. Experiment 2 arranged percentile schedule contingencies that reinforced minimizing or maximizing (or did not differentially reinforce) distance. Experiment 3 reinforced horizontal and vertical deviations separately and with differential weightings. Experiment 4 changed whether deviations were to be maximized or minimized halfway through each session. Experiment 5 reinforced minimizing deviations to novel random locations. In summary, percentile schedule contingencies controlled emitted response locations relative to the target location. This method extends the use of percentile schedules to nonhuman primates and a touchscreen apparatus and demonstrates control of operant responding across a continuous two-dimensional plane, with broad relevance for shaping, the study of spatial dimensions of responding, and for evaluating the effects of other variables on spatial learning.