Timing compound stimuli: Relative reinforcer probabilities divide stimulus control in the multiple peak procedure.

Abstract

Previous research suggests that when multiple stimuli signal the location of future reinforcers, the extent of control by each stimulus depends on the relative reinforcer probability associated with that stimulus. In this experiment, we asked whether relative reinforcer probabilities also divide control between stimuli that signal the time to future reinforcers. Six pigeons responded in a multiple peak procedure in which 2 dimensions of a compound stimulus—a color (red or green) and a frequency of on/off alternation (fast or slow)—signaled a fixed-interval 2- or 8-s schedule. Across conditions, we varied the probability of reinforcer deliveries associated with each dimension from .1 to .9. Relative reinforcer probabilities determined the degree of control by each stimulus dimension; as the probability of reinforcer deliveries associated with one dimension increased, a peak in response rates at the time signaled by that dimension became apparent while a peak at the time signaled by the other dimension diminished. However, these effects of relative reinforcers were smaller than in previous research, probably because elapsed time also competed for control over behavior. These findings extend the relation between relative reinforcers and divided stimulus control from spatial to temporal discriminations.

Divided stimulus control depends on differential and nondifferential reinforcement: Testing a quantitative model

“…Like Shahan and Podlesnik, Davison and Elliffe found that as the probability of reinforcer deliveries for correct responses according to one dimension increased, so did the frequency of correct responses according to that dimension, whereas correct responses according to the other dimension decreased. Thus, these experiments demonstrate that divided control between compoundstimulus dimensions depends on relative reinforcer rates (see also Davison, 2018;Gomes-Ng et al, 2019a, 2019bPodlesnik et al, 2012;Shahan & Podlesnik, 2007, 2008. Davison and Elliffe (2010) likened the contingencies in their experiment to those in a "reinforcementfor-errors" procedure, in which some reinforcers are delivered after incorrect responses.…”

“…Although these dimensions may differ in their discriminability-color is discriminable immediately, whereas alternation frequency is only discriminable across time-we presented the sample stimulus for at least 5 s and arranged maximally different alternation frequencies with the goal to enhance the discriminability of the alternation-frequency dimension. We have previously used these two stimulus dimensions in other studies of divided stimulus control (Cowie et al, 2020;Gomes-Ng et al, 2019a, 2019b, and previous research-including Davison and Elliffe (2010)-shows that pigeons can discriminate well between frequencies of keylight-color alternation (e.g., Cowie et al, 2020;Gomes-Ng et al, 2019a, 2019bKrägeloh & Davison, 2003;Roberts, 1997;Roberts et al, 2000;Roberts & Mitchell, 1994).…”