Perceptual decision making has been studied extensively using a random dot coherent visual motion task, in which trial difficulty is varied by the fraction of the dots moving coherently. In primates, reaction time is longer and accuracy is lower in more difficult trials; for a given difficulty, increasing reaction time increases accuracy (Palmer et al 2005). It remains unclear whether the same is true of rodents, at least in other perceptual tasks (Uchida & Mainen 2003; Rinberg et al 2006). In order to compare perceptual decision making of primates and rats more directly we trained rats to perform the identical visual motion task studied in primates.

We report results for training blocks of >=1000 trials during which the stimulus distribution was fixed and performance was stationary. We define the reaction time as the time between the trial-initiating request and the trial-terminating response. 

We find: (1) Performance ranged from 50% correct at low coherence to a plateau level exceeding 85% correct at high coherence. Thus rats’ accuracy depends on task difficulty, as expected. (2) For trials of a given coherence, rats’ accuracy increased with reaction time. Thus rats show a trade-off between speed and accuracy. (3) The reaction time distribution did not depend on trial difficulty.