Analysis of the superiority of discrete-trials over free-operant procedures in temporal response differentiation

Two experiments examined the effects of various relations between 81 male Wistar rats' leverpress duration and occurrence of intertrial intervals (ITIs). Exp I employed fixed minimum-duration requirements for reinforcement in 4 groups that received ITIs following all presses, no presses, only presses of subcriterion duration, or only those of criterion duration. The major effect was that ITIs following subcriterion press durations or nonreinforced responses increased mean press duration and efficiency. However, this procedure confounded overall frequency of ITIs with their relation to press duration. Exp II used percentile schedules to remove this confounding. Reinforcement always followed the longest 20% of Ss' press durations. Four groups received ITIs following the longest 20%, the shortest 20%, the shortest 80%, or a random quarter of the shortest 80% of their press durations. Again, ITIs following nonreinforced responses greatly facilitated differentiation of press duration. Results support the existence of a mechanism in which ITIs serve as feedback for brief lever releases to prevent Ss from mistaking the resulting 2 shorter durations for a single longer duration press. (7 ref)