Association for Behavior Analysis International

The present studies assessed variables involved in behavioral persistence under resistance to change and relapse procedures. By using a four-component multiple schedule, Bai, Xue, Podlesnik and Elliffe established key pecking under different reinforcement frequencies (rich, lean and intermediate). Their results suggest a role for stimulus generalization as a determinant of resistance to change. Teixeira and Cançado assessed the effects of response-reinforcer dependency on resistance to change and choice in two experiments. Their results suggest that contingencies which generate high or low response rates affect resistance to change, but have not impacted choice. Madrigal, Craig and Flores reported similar results by isolating response-rate effects and training length on ABA-operant renewal. Rats responded at high or low rates and were exposed to either short or extended training. They found a relation between length of training and renewal, which was affected by differences in training response rates. Finally, Nist and Shahan explored the temporal dynamics of resurgence during alternative reinforcement thinning via within-session progressive-interval in two experiments. In both, resurgence occurred during thinning but not after alternative-response extinction. The present set of experiments extend previous findings and provide novel strategies for the assessment of the variables involved in behavioral persistence and relapse.

Practitioners, teachers, basic and applied researchers, and translational researchers

Behavioral momentum theory suggests that the relation between a discriminative stimulus and reinforcement determines the persistence of responding in the presence of that discriminative stimulus. However, responding also generalizes across similar discriminative stimuli and it remains unclear how generalization may affect behavioral persistence. The present experiment arranged food reinforcement for pigeons’ key-peck responses in a four-component multiple schedule. Components were signalled by different wavelengths projected onto the response key. Two components arranged equal, Intermediate variable interval (VI) 60-s schedules in the presence of 510 nm and 580 nm, while the other two components arranged Rich VI 15-s and Lean VI 240-s schedules in the presence of 530 nm and 560 nm, respectively. Responding in the Intermediate component flanking the Rich Component was more persistent than responding in the Intermediate component flanking the Lean Component. Additionally, generalization tests and preference probes provided converging evidence that stimuli associated with Richer and Leaner reinforcement schedules impacted the value of the stimuli signalling the Intermediate components. These data suggest a role for generalization in establishing the stimulus-reinforcer relation that determines behavioral persistence.

Two experiments assessed the effects of the response-reinforcer dependency on resistance to change and choice. In both, pigeons were first exposed to a two-component multiple schedule and then to a concurrent-chains schedule. Reinforcement rate was equated between components and alternatives. In Experiment 1, the dependency was low in one multiple-schedule component (concomitant variable interval, VI, variable time, VT), and high in the other (VI). Response rates were lower in the low than in the high-dependency component, and resistance to extinction and VT schedules were greater in the former. Also, pigeons preferred the terminal link associated with low rather than high dependency. In Experiment 2, the procedure was replicated but response rates were lower in the high-dependency than in the low-dependency multiple-schedule component and terminal link. Resistance to extinction and VT schedules were greater in the high than in the low-dependency component, but preference was greater for the terminal link associated with low dependency. These results extend previous studies on the relation between resistance to change and choice. In addition, they suggest that contingencies that generate high or low response rates are important determinants of resistance to change as a function of dependency in multiple and concurrent-chains schedules of reinforcement.

The duration of acquisition has been positively associated with ABA renewal. However, in studies that have showed this effect, response rates generally are higher for subjects that experience prolonged training than for subjects that experience shorter training. Thus, it is unclear if this differences can be explained by the duration of acquisition or by any difference on response rates. Therefore, the present experiment was developed with the purpose of further assessing this possibility. During the first condition, two groups of rats were exposed to either short or long training in Context A. Within each group, high and low response rates were established according to a VI30s (100% and 50% response-dependent food, respectively). Rats were then exposed to extinction sessions under Context B. Finally, we tested for renewal by returning the rats to Context A. The second condition occurred in the same manner as the first, with the exception that subjects’ response rates were reversed (i.e. High – Low; Low – High). Greater ABA renewal was observed after long training than short training. However, independently of the length of training, greater renewal was observed after low lever-pressing response rates. These results extend those of previous studies where renewal was greater after long training. Additionally, they provide evidence on the effects of response rates on ABA renewal.