Behavioral definition of minimal reaction time in monkeys

Two monkeys (Macaca mulatta) were trained to press a key after onset of a tone and to release it after a 1-sec fixed foreperiod terminated by a light. The effects of imposing temporal contingencies on key release reaction times were determined by reinforcing only those releases whose latencies from the light fell within a "payoff band", two time limits 50 msec apart located at some delay following the light. Over several days this delay was first gradually decreased, shortening the interval between light and payoff band, and then gradually increased again. For each delay, the median reaction time and a measure of variability were obtained from the latency distribution. For both animals, median latency could be decreased to 180 msec with the variability remaining small. Moving the payoff band still closer to the light resulted in further decrease in median latency but an abrupt increase in variability. This is in agreement with a model for simple reaction time derived from human research which suggests that this increased variability results from the inclusion of high-variability foreperiod time estimations in the latency distribution. These results indicate that interpretation of monkey response latencies as "minimal reaction times" requires examination of temporal reinforcement contingencies and variability of latencies.