Reinforcer magnitude and resurgence

Punishment-induced resurgence

The phenomenon of extinction-induced resurgence is well established, but there is comparatively little experimental evidence for punishment-induced resurgence. Punishment-induced resurgence can by tested by contingent shocks following the alternative response. The purpose of Experiment 1 was to test whether low-intensity shocks, that do not decrease rate of reinforcement, result in resurgence. Four rats served as subjects. Rats were exposed to three sequential conditions: (a) variable-interval (VI) 30-s food delivery for a lever press (target response); (b) VI 30-s food delivery for a nose poke (alternative response) and extinction of the lever press; (c) VI 30-s reinforcement for a nose poke with superimposed VI 60-s shock delivery. In the final condition, shocks increased gradually from 0.1 to 0.5 mA. Experiment 2 evaluated whether an abrupt introduction of a high-intensity shock would result in resurgence. Three rats served as subjects and were exposed to three sequential conditions: (a) VI 30-s food delivery for a lever press; (b) VI 30-s food delivery for a nose poke and extinction of the lever press; (c) continued VI 30-s reinforcement for a nose poke with superimposed VI 60-s 0.6 mA shock delivery. Resurgence was observed in all subjects, including in situations in which rate of responding, but not rate of reinforcement, decreased. The present study provides additional evidence for punishment-induced resurgence, but future studies are warranted to determine the extent to which punishment can produce resurgence with or without decreases in rates of reinforcement.

Resurgence of ethanol seeking following voluntary abstinence produced by nondrug differential reinforcement of other behavior

Resurgence refers to the relapse of a target behavior following the worsening of a source of alternative reinforcement that was made available during response elimination. Most laboratory analyses of resurgence have used a combination of extinction and alternative reinforcement to reduce target behavior. In contingency-management treatments for alcohol use disorder, however, alcohol use is not placed on extinction. Instead, participants voluntarily abstain from alcohol use to access nondrug alternative reinforcers. Inasmuch, additional laboratory research on resurgence following voluntary abstinence is warranted. The present experiment evaluated resurgence of rats' ethanol seeking following voluntary abstinence produced by differential reinforcement of other behavior (DRO). Lever pressing produced ethanol reinforcers during baseline phases. During DRO phases, lever pressing continued to produce ethanol and food reinforcers were delivered according to resetting DRO schedules. Ethanol and food reinforcers were suspended during resurgence test phases to evaluate resurgence following voluntary abstinence. Lever pressing was elevated during baseline phases and occurred at near-zero rates during DRO phases. During the resurgence test phases, lever pressing increased, despite that it no longer produced ethanol. The procedure introduced here may help researchers better understand the variables that affect voluntary abstinence from ethanol seeking and resurgence following voluntary abstinence.

A quantitative analysis of resurgence following downshifts in alternative-reinforcer magnitude

Resurgence is the increase in a previously reinforced and then extinguished target response due to changes in reinforcement conditions for an alternative response, including reductions in the rate or magnitude of reinforcement for the alternative response. Research with nonhumans suggests that reductions in both alternative-reinforcer rate and magnitude produce resurgence, but the present study was the first to examine effects of downshifts in alternative-reinforcer magnitude on humans' resurgence. Moreover, it was the first to evaluate whether the quantitative framework, resurgence as choice in context (RaC² ), could account for those effects. Consistent with predictions of RaC² , resurgence of a target button press occurred with reductions in point gain for an alternative response, with greater reductions producing higher levels of resurgence. However, the model consistently underpredicted and then overpredicted resurgence during tests with low-magnitude reinforcement and extinction. Systematic deviations in model predictions of alternative responding were also evident and consistent with previous fits of RaC² to nonhuman data. Overall, our findings suggest that RaC² could be a useful quantitative theoretical framework for understanding processes contributing to resurgence in humans, but further theoretical development is needed to account for the apparent divergent effects of extinction versus downshifts in reinforcer magnitude.

A Comprehensive Systematic Review of Procedures and Analyses Used in Basic and Preclinical Studies of Resurgence, 1970-2020

Resurgence is the return of a previously reinforced response as conditions worsen for an alternative response, such as the introduction of extinction, reductions in reinforcement, or punishment. As a procedure, resurgence has been used to model behavioral treatments and understand behavioral processes contributing both to relapse of problem behavior and flexibility during problem-solving. Identifying existing procedural and analytic methods arranged in basic/preclinical research could be used by basic and preclinical researchers to develop novel approaches to study resurgence, whereas translational and clinical researchers could identify potential approaches to combating relapse during behavioral interventions. Despite the study of resurgence for over half a century, there have been no systematic reviews of the basic/preclinical research on resurgence. To characterize the procedural and analytic methods used in basic/preclinical research on resurgence, we performed a systematic review consistent with PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses). We identified 120 articles consisting of 200 experiments that presented novel empirical research, examined operant behavior, and included standard elements of a resurgence procedure. We reported prevalence and trends in over 60 categories, including participant characteristics (e.g., species, sample size, disability), designs (e.g., single subject, group), procedural characteristics (e.g., responses, reinforcer types, control conditions), criteria defining resurgence (e.g., single test, multiple tests, relative to control), and analytic strategies (e.g., inferential statistics, quantitative analysis, visual inspection). We make some recommendations for future basic, preclinical, and clinical research based on our findings of this expanding literature.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40614-022-00361-y.

Response effort and resurgence

This study provides an initial translational examination of response effort and resurgence. Eleven typically developing adults and five adolescents with autism served as participants across two experiments. Participants received points for touching moving stimuli on a computer screen. The resurgence evaluation consisted of three phases: establishment wherein R1 was reinforced, elimination wherein R1 was placed on extinction while R2 was reinforced, and extinction wherein R1 and R2 no longer resulted in reinforcement. Rate of R1 during extinction was compared across three conditions: intermediate, easy, and difficult. Disparity in effort was created by manipulations of the size and speed of objects that moved about on a computer screen. In Experiment 2, control stimuli were added to the experimental arrangement. Across the two experiments, the magnitude of resurgence was greater when R1 was easy. In Experiment 2, both R1 and control responding were greater in the extinction phase than in the elimination phase in all conditions with all participants. The present study supports the hypothesis that response effort affects resurgence and that less effortful responses are likely to recur with greater magnitude under conditions that produce resurgence than are their more effortful counterparts.

The effects of large, small, and thinning magnitudes of alternative reinforcement on resurgence

Resurgence refers to an increase of a previously reinforced target behavior following the worsening of conditions for a more recently reinforced alternative behavior. There is evidence to suggest that alternative reinforcers of greater magnitude are more effective at reducing target responding but may also result in more resurgence when removed. Similar effects have been observed with high rates of alternative reinforcement. However, in clinical settings, reinforcement rate thinning is used to reduce the likelihood of resurgence associated with higher rates of alternative reinforcement. Given the clinical importance of alternative reinforcer magnitude, it is necessary to evaluate how reinforcer magnitude thinning may impact resurgence as well. Following Phase 1 in which target responding was reinforced, rats earned either large (six pellets), small (one pellet), or thinned (reduced from six pellets to one across sessions) magnitude reinforcement for alternative responding during target-response extinction in Phase 2. Then, alternative responding was placed on extinction for all groups in Phase 3. Target responding was comparably elevated at the end of Phase 2 for groups Small and Thin compared to group Large. In Phase 3, resurgence was evident only in group Large but target responding remained relatively elevated in groups Small and Thin. These results provide additional evidence of the important interplay between conditions of alternative reinforcement and the persistence and resurgence of target responding.

Resurgence during transitions from variable- to fixed-interval schedules

The effects of local periods of extinction on resurgence following transitions from variable-interval (VI) to fixed-interval (FI) schedules were studied using four pigeons exposed to a within-session resurgence procedure. Each session was divided into a Training (T) Alternative-Reinforcement (AR), and Resurgence Test (RT) phase. During the T phase, key pecking was reinforced under a VI 60-s schedule on one key. In the AR phase, responses reinforced in the T phase were extinguished, while responses to a different key were reinforced under a VI 90-s schedule. Next, responding to the same key that produced reinforcers in the AR phase was reinforced according to four different RT conditions: RT phase I (FI 90 s), RT phase II (FI 180 s), RT phase III (FI 45 s), or RT phase IV (extinction). The frequency of resurgence generally was an inverse function of the rate of reinforcement in the RT phase. Resurgence occurred less often when reinforcers were delivered under the FI 45-s schedule and more often under leaner schedules in the RT phase, peaking under extinction. The results show that resurgence may occur during local periods of extinction, with larger and more consistent effects occurring when the rate of reinforcement in the RT condition is leaner than it was during the preceding AR phase.

Empirically Deriving Omission and Commission Errors for Relapse Tests: A Demonstration of Reverse Translation

Most studies examine treatment relapse by programming contextual changes with perfect treatment integrity or with omission errors in the absence of a context change (i.e., all alternative responses placed on extinction). Recently, Mitteer et al. (2018) examined caregiver behavior in response to a confederate playing the role of a child with destructive behavior, providing the opportunity for researchers to empirically derive reinforcement schedules and test caregiver error patterns within future relapse tests with children. The present study represents a pilot demonstration of methods for reverse translating findings from caregivers to relapse preparations with children. We used a human-operant arrangement with three children with autism spectrum disorder in which they (a) emitted a target response (i.e., pad touch) for a preferred item in a home-like context, (b) emitted an alternative response (e.g., card touch) for the item in a clinic context while the target response was extinguished, and (c) experienced a relapse test in which the experimenter programmed the same low-rate omission and commission errors that caregivers made in the prior study within the home-like context. During the relapse test, target responding approximated or exceeded baseline ranges for all cases, and alternative behavior extinguished for two of the three cases. We discuss how researchers might incorporate similar translation processes in future relapse research.

Big Surprises: Jackpot Reinforcers in Research and Practice

Unusually large, infrequent reinforcers, described as jackpots, are the subject of considerable discussion among applied animal behaviorists. Such reinforcers offer considerable promise in applied behavior analysis as a means of both potentiating training of new responses and response classes and enhancing previously learned ones. The concept of jackpot reinforcement, however, is rife with not only a lack of definitional and procedural clarity but also a paucity of research, either basic or applied, on such reinforcement. Considerations in undertaking such research include defining the parameters of jackpot reinforcers, identifying suitable dependent variables, and creating experimental designs appropriate for their assessment. The few experimental analyses of jackpot reinforcer effects on either response acquisition or maintenance have produced little evidence of systematic effects, despite the use of several different methods and behavioral measures. Negative results could reflect either the absence of systematic effects of these jackpot reinforcers or unsuitable methods of analysis. Only further research will take the topic of jackpot reinforcers beyond opinion and testimonial and into the realm of the science of behavior.

Resurgence without overall worsening of alternative reinforcement

Three experiments were conducted with pigeons to assess discriminated periods of nonreinforcement as precipitators of resurgence. Each experiment occurred in three phases. In the Training phase, key-pecking was reinforced according to variable-interval schedules that alternated between two response keys (Experiment 1) or were concurrently available on two response keys (Experiments 2a & 2b). In the Alternative-Reinforcement phase, responding to one key was extinguished, while that to the other was reinforced according to tandem schedules. These then were replaced by chained schedules with the same programmed reinforcement rate in the Resurgence-Test phase. Resurgence occurred both when the signaled period of nonreinforcement was a darkened keylight in the terminal link of the chain schedule (Experiment 1) and a darkened keylight (Experiment 2a) or keylight color change (Experiment 2b) in the initial link of the chain schedule. Thus, signaled periods of extinction, without accompanying reductions in reinforcement rate, precipitated resurgence, suggesting that resurgence is not the result of worsening of overall reinforcement conditions, but also occurs when local conditions of reinforcement are worsened.

Repeated resurgence with and without a context change

Resurgence is the reoccurrence of a target response when reinforcement for a more recently reinforced alternative response is eliminated or reduced. The present study arranged two successive three-phase procedures to assess whether resurgence decreases with repeated assessments. Moreover, we arranged a contextual change from the first to second assessment for some groups. Phase 1 reinforced a target response on a touchscreen computer with typically developing adults as participants according to either variable-ratio or variable-interval schedules of reinforcement. Phase 2 extinguished target responding and reinforced alternative responding. Phase 3 tested for resurgence by extinguishing alternative responding. Resurgence reliably occurred in all tests and decreased from the first to second exposure to the procedures but there were no effects of context change. Therefore, repeated exposures to resurgence tests reduced those effects but contextual changes had no effect.

The control response in assessing resurgence: Useful or compromised tool?

Resurgence experiments sometimes include an operandum on which a history of reinforcement has not been experimentally established. The purpose of this control operandum is to rule out a generalized increase in responding when the alternative response is extinguished as being the cause of the resurgent target response. A review of the results of experiments conducted with both nonhumans and humans in which a control operandum was included shows that control- operandum responding is more common in the latter and almost nonexistent in the former. Both the presence and absence of responding on the control operandum, however, are subject to multiple interpretations thereby rendering it a compromised tool. Alternatives to using a control operandum to rule out extinction induction as the basis for resurgence include a preresurgence test control procedure and a differential resurgence procedure.

Hierarchical resurgence

In resurgence, conventionally a target response is trained and then extinguished while some alternative response is reinforced. In the most common procedure, when the latter is extinguished, the former resurges. The present experiments examined resurgence after two responses were trained sequentially and subsequently extinguished. In Experiments 1 and 2, keypecking to one key was trained and then extinguished as keypecking to a different key was trained then later extinguished. In both experiments, regardless of the spatial location of the different keys, the last-trained response resurged before the first-trained one. The results were replicated in Experiment 3 where reinforcement rate of the first-trained response was four times that of the second-trained response. The results in conjunction with earlier experiments suggest that resurgence occurs hierarchically, although whether more or less recently trained target responses resurge first or later may depend on both current and historical variables. The results also raise questions about the interpretation of responding on a control key that sometimes is included in resurgence experiments to isolate resurgence from extinction-induced responding.