The influence of `switching' on the psychometric function in the free-operant psychophysical procedure

Fifteen rats were trained under the `free-operant psychophysical procedure', using a sucrose reinforcer. The training sessions consisted of 50-s trials in which reinforcers were available on a variable-interval 30-s schedule; in the first 25 s of each trial, reinforcers were only available for responses on lever A, whereas in the last 25 s they were available only for responses on lever B. Data were collected in probe trials (4 per session) in which no reinforcers were delivered, during the last 10 sessions of each phase of the experiment. In phase 1 (70 sessions), repetitive switching between the levers was prevented by withdrawal of lever A after the first response on lever B in each trial. In phase 2 (40 sessions), this constraint on switching was removed. In phase 3 (40 sessions), it was reinstated. In all 3 phases, the response rates on lever A declined and the response rates on lever B increased as a function of time from the trial onset. The response rate on lever B, expressed as a percentage of the overall response rate, conformed to a two-parameter logistic function. Removal of the constraint on switching did not alter the indifference point (the time corresponding to 50% responding on lever B), but did reduce the slope of the function, this being reflected in an increase in the Weber fraction. The changes were reversed when the constraint on switching was reinstated. The results show that constraining switching altered the slope of the psychometric function; thus caution is needed in interpreting psychometric functions obtained with the free-operant psychophysical procedure in terms of theoretical models of timing processes.