JOURNAL OF APPLIED BEHAVIOR ANALYSIS
2015, 48, 390–401
NUMBER
2 (SUMMER)
BASIC AND TRANSLATIONAL EVALUATION OF RENEWAL OF OPERANT RESPONDING MICHAEL E. KELLEY, CLARE J. LIDDON, AURELIA RIBEIRO, AND ABIGAIL E. GREIF THE SCOTT CENTER FOR AUTISM TREATMENT AND FLORIDA INSTITUTE OF TECHNOLOGY
AND
CHRISTOPHER A. PODLESNIK THE SCOTT CENTER FOR AUTISM TREATMENT AND FLORIDA INSTITUTE OF TECHNOLOGY AND THE UNIVERSITY OF AUCKLAND
Treatment relapse, defined as the reemergence of problem behavior after treatment, is a serious difficulty faced by clinicians. Failures of treatment integrity (i.e., failure to implement interventions as intended) are often invoked to explain the reemergence of problem behavior. Basic studies suggest that the prevailing stimulus context might also contribute. We conducted 2 experiments in which reinforcement for a target response was followed by 2 phases of extinction with different or identical stimulus contexts relative to baseline (ABA renewal). In Experiment 1, pigeons served as subjects using procedures typical of those used in basic behavioral research. Experiment 2 was designed as a translational replication of Experiment 1, and children who had been diagnosed with autism served as participants. Returning to the previously reinforced stimulus context in both species produced a clear and immediate increase of extinguished responding. These findings are consistent with previous studies that have suggested that both reinforcement contingencies and stimulus context influence the reemergence of extinguished behavior. Key words: renewal, treatment relapse, behavioral momentum theory, pigeon, translational research
Voluminous research results support the use of applied behavior-analytic methods for solving problems of social significance (e.g., Beavers, Iwata, & Lerman, 2013). Although treatments may be initially effective, researchers and practitioners continue to struggle to understand fully the necessary and sufficient conditions for maintaining treatment efficacy over time, across changes in stimuli (e.g., teachers, locations), and We thank the members of the Experimental Analysis of Behaviour Research Group for their help in conducting these experiments and Mike Owens for looking after the pigeons. Portions of the present data set were presented at the 40th annual meeting of the Association for Behavior Analysis International. Experiment 1 was carried out under Approval AEC/2011/RT909 granted by the Animal Ethics Committee of the University of Auckland. Experiment 2 was carried out under Approval IRB 13-081 granted by the Florida Institute of Technology. Send correspondence to Michael E. Kelley (e-mail:
[email protected]) or Christopher A. Podlesnik (e-mail:
[email protected]). doi: 10.1002/jaba.209
in the presence of disruptive events (e.g., extinction, distraction; Luczynski, Hanley, & Rodriguez, 2014). These threats to long-term efficacy have received attention in the literature, are often discussed as a matter of practice, and continue to warrant the attention of researchers and clinicians. One strategy for improving treatment outcomes is to conduct systematic research on the basic processes that contribute to treatment failure. In practice, interventions may fail for a variety of reasons, including lapses in treatment integrity (St. Peter Pipkin, Vollmer, & Sloman, 2010), changes in motivational variables (Murphy, McSweeney, Smith, & McComas, 2003), and changes in the function of behavior (Lerman, Iwata, Smith, Zarcone, & Vollmer, 1994), among others. Variables of treatment failure are often labeled with descriptive terms (e.g., “lapse in treatment integrity”) that may be neither linked to identifiable conceptual systems nor
390
TRANSLATIONAL ANALYSIS OF RENEWAL suggestive of a strategy for remediation (however, see St. Peter Pipkin et al., 2010, for an in-depth discussion and suggestions for potential ways to mitigate lapses in treatment integrity). Lapses in treatment integrity compromise experiments’ internal and external validity (Gresham, Gansle, & Noell, 1993) and may reduce treatment efficacy (Stephenson & Hanley, 2010; Wilder, Atwell, & Wine, 2006) and response acquisition (Carroll, Kodak, & Fisher, 2013), particularly if these failures are considerable. In addition, different types of lapses may produce different behavioral effects. Development of a conceptual framework into which treatment failures may be functionally categorized would allow one to test specific hypotheses, to identify important variables that might control behavior, and to shed light on which operations govern behavior under certain experimental conditions. Thus, linking specific operations that produce specific treatment failures to broader conceptual systems might guide research for basic processes that affect intervention efficacy. One conceptual framework relevant to understanding basic behavioral processes that underlie treatment relapse is resurgence (see Pritchard, Hoerger, & Mace, 2014, for a recent review on treatment relapse and behavioral momentum theory). Applied studies of resurgence demonstrated that eliminating a recently reinforced desirable behavior could produce relapse of previously extinguished problem behavior (e.g., Volkert, Lerman, Call, & Trosclair-Lasserre, 2009). These applied studies were inspired by basic studies that assessed extinction processes that revealed the potential for extinguished responding to reemerge simply by eliminating reinforcement for more recently reinforced responses (e.g., Leitenberg, Rawson, & Mulick, 1975). More recent applied studies have begun to integrate specific operations identified to affect resurgence in basic studies to affect treatment outcomes. For example, extinguishing problem behavior while reinforcing desirable behavior for longer durations will reduce
391
resurgence of problem behavior (e.g., Wacker et al., 2011). Therefore, linking the methods of applied researchers and clinicians to broader conceptual systems based on basic science is an effective strategy (see Mace, 1994). Another conceptual framework relevant to understanding the basic behavioral processes that underlie treatment relapse suggests that the reemergence of previously eliminated behavior depends on the prevailing stimulus context (see Bouton, 1993 2002; Bouton, Winterbauer, & Todd, 2012, for reviews). Consistent with prior usage (e.g., Bouton et al., 2012), we use the term stimulus context to refer to any change in the environmental stimulus conditions that have been demonstrated to control behavior. Altering stimulus context after treatment is sufficient to precipitate a reemergence, or relapse, of previously eliminated behavior in both respondent- and operant-conditioning procedures. The advantage of this particular conceptualization of relapse is that it specifies a common set of behavioral processes that potentially underlie several different types of relapse. For example, renewal is characterized by a three-phase arrangement in which responding reinforced during a baseline phase is followed by two phases of extinction (see Bouton, Todd, Vurbic, & Winterbauer, 2011; Nakajima, Tanaka, Urushihara, & Imada, 2000; Podlesnik & Shahan, 2009). The independent variables of interest in renewal are the stimulus contexts associated with Phases 1, 2, and 3, respectively. In ABA renewal, the stimulus contexts are identical in Phases 1 and 3 and different in Phase 2. Conversely, in ABC renewal, the stimulus contexts differ in all three phases. In both ABA and ABC renewal, responding extinguished in Context B returns when transitioned to either Context A or Context C (e.g., Bouton et al., 2011). Therefore, performance of the extinguished response depends on the current stimulus context; transitioning out of the extinction context (B) appears to be sufficient to produce relapse.
392
MICHAEL E. KELLEY et al.
Although a majority of the renewal research comes from studies of respondent conditioning, Bouton et al. (2011) and others demonstrated the generality of renewal effects by extending the procedures and results to operant conditioning (see also Nakajima et al., 2000; Podlesnik & Shahan, 2009; Todd, Winterbauer, & Bouton, 2012; Welker & McAuley, 1978). Bouton et al. trained rats to press levers in a particular set of stimulus contexts with distinct environmental features (e.g., visual, tactile, and olfactory stimuli; Context A). They extinguished responding under a stimulus context with different environmental features (Context B). Finally, returning to Context A or transitioning to a novel context (Context C) while the extinction contingency was maintained renewed responding. These results are important for several reasons. First, these data demonstrate the generality of renewal effects to operant behavior. Second, the data show that responding was not only dependent on the operating contingencies. Reinforced responding occurred reliably in Phase 1 and eventually decreased to low levels during Phases 2 and 3 after lever pressing had contacted the extinction contingency. All of these response patterns are predictable and expected based on reinforcement contingencies. However, the reemergence of responding in Phase 3 of ABA and ABC renewal procedures reveal the influence of the prevailing stimulus context. Third, the concept of renewal has not been assessed or used to understand relapse in applied behavior analysis in any comprehensive way, despite its potential relevance. That is, despite the potential for renewal research to influence treatment programming and training, very little research with operant behavior appears in the literature. Fourth, the processes that underlie renewal might provide a conceptual framework for further understanding the conditions that precipitate relapse (e.g., Bouton, 1993, 2002; Bouton et al., 2012; Winterbauer & Bouton, 2010), similar to the above example with resurgence (Wacker et al., 2011).
There are many potential examples of clinical treatments that resemble conditions arranged in basic behavioral studies of renewal, with ABA and ABC renewal being two types. For example, a child may engage in problem behavior at school (Phase 1; Stimulus Context A). Next, treatment may be executed at a specialized center for assessing and intervening for problem behavior (Phase 2; Stimulus Context B). Finally, the child is returned to the school (Phase 3; Stimulus Context A), and problem behavior returns. Alternatively, a child might attend a new school after treatment (thus, Phase 3 might be Context C). Interestingly, basic research on renewal specifically predicts that problem behavior will reemerge when the student in this scenario returns to school. The typical applied treatment strategy to combat the reemergence of problem behavior when returning to the original context (A) in Phase 3 is to train teachers, caregivers, and others in the treatment developed in Phase 2 so that treatment will be effective when the student returns to the original context. Although this might be partially effective, basic research on renewal suggests merely transitioning from the treatment context alone is sufficient to produce renewal of problem behavior. Therefore, focusing only on contingencies and not stimulus control processes ignores an entire set of potentially relevant behavioral processes that are relevant to determining long-term treatment effectiveness. This likely reflects a disconnect both conceptually and methodologically between basic research on renewal and the execution of applied treatments (Mace, 1994). The disconnect may be partially explained by the primary reliance on contingency management in practice, whereas basic research suggests that stimulus control processes also play a role in whether treatments will be successful. The general purpose of this study was to evaluate the renewal paradigm in the context of collaborative basic science and translational approaches and to bridge an apparent gap in conceptualizing the causes of treatment relapse
TRANSLATIONAL ANALYSIS OF RENEWAL between basic and applied literatures. Specifically, the purpose of Experiment 1 was to replicate previous renewal research both with pigeons and in the context of purposeful translation to humans. The purpose of Experiment 2 was to conduct a translational evaluation of the renewal paradigm and to lay the groundwork for applied renewal research. Thus, the combination of Experiments 1 and 2 demonstrates the generality of using context renewal as a conceptual framework from which to interpret subsequent renewal findings in applied research and clinical treatment. GENERAL METHOD Experiments 1 and 2 shared common features as a part of the coordinated basic and translational nature of the research (see Table 1 for a summary). Both experiments employed different contexts for Phases A and B. Although the specific stimuli that made up those contexts between Experiments 1 and 2 differed (see below), the execution of stimulus presentation was identical. That is, pigeons in Experiment 1 and children with autism in Experiment 2 were exposed to the stimulus contexts in an ABA design (i.e., ABA renewal), and to the contingencies in an ABB design (reinforcement, extinction, extinction). ABA renewal procedures are defined by this particular experimental arrangement (see Bouton, 2002). The design differs from common applied behavior-analytic designs (e.g., withdrawal, reversal, multiple baseline, and multiple schedule), but has similarities to the reversal and withdrawal designs. The contingency arrangement (ABB) predicts that responding will occur in the A phase, extinguish in the B phase, and remain extinguished in the third phase. The stimulus context arrangement (ABA) predicts that responding will occur in the A phase, extinguish in the B phase, and reemerge in the third phase. Thus, control is demonstrated by violating the predictions generated by the contingency
393
arrangements (i.e., reemergence of responding in the third phase). EXPERIMENT 1 Subjects and Apparatus Six experienced homing pigeons, numbered 111 to 116, were individually housed and maintained at 85% 15 g of their free-feeding body weights with postsession supplementary feeding of mixed grain as necessary. Water and grit were available at all times. The pigeons’ home cages also served as the experimental chambers (see Podlesnik, Bai, & Elliffe, 2012, for a detailed description). Three keys, 85 mm apart center to center, could be transilluminated red or white, and pecks exceeding 0.1 N closed a microswitch. During hopper presentations, all key lights were turned off and the hopper was raised and illuminated for 2 s. All experimental events were arranged and recorded by an IBM PC-compatible computer running MED-PC IV software. The colony room lighting was switched on at 12:00 a.m. and off at 4:00 p.m. daily. Sessions began at 1:00 a.m. daily. No personnel entered the room while sessions were conducted. Procedure Pigeons were exposed to three phases: reinforcement (Context A), extinction (Context B), and extinction (Context A), sequentially. The rate at which key-light color stimuli alternated defined the different stimulus contexts across phases. All sessions lasted approximately 15 min, excluding reinforcement time. We conducted a fixed number of sessions per condition to establish stable response rates in Phase 1 and clear extinction functions reaching near-zero response rates in Phases 2 and 3 (see also Podlesnik & Kelley, 2014). Reinforcement (Context A). Keypecking produced reinforcement according to a fixedinterval (FI) 10-s schedule. The center key alternated between red and white every 0.1 s. This phase was conducted for 15 sessions.
MICHAEL E. KELLEY et al.
394
Table 1
Subjects Responses Stimuli Context A B Contingencies Reinforcers Design (contexts) Design (contingency)
Experiment 1
Experiment 2
Pigeons Keypecks Key lights Key-light flash rate 0.1 fps 0.5 fps FI 10 s 2-s access to wheat ABA ABB
Children with autism Mastered tasks Colors (tasks, placards, shirts) Context-specific colors Color 1 Color 2 FR 1 Preferred edible items ABA ABB
Extinction (Context B). Keypecking was extinguished, and the center key alternated between red and white every 0.5 s. This phase was conducted for eight sessions, during which response rates decreased below 10% of baseline response rates for all pigeons (see Podlesnik & Shahan, 2009). Extinction (Context A). Keypecking remained on extinction, and the key light again alternated between red and white every 0.1 s for 4 sessions. EXPERIMENT 2 Participants and Setting Drew, a 4-year-old boy, and John, a 9-year-old boy, participated in this study. Both participants had been diagnosed with autism. John had a 6year history of early intensive behavioral intervention at the time of the experiment. Both participants scored as Level 2 (moderate) learners on the Verbal Behavior Milestones Assessment and Placement Program (Sundberg, 2008). John and Drew participated because they were the first two individuals whose families responded to recruitment flyers. Experimenters conducted sessions in individual treatment rooms in an early behavioral intervention clinic. The treatment rooms contained one table, two chairs, contextual stimuli (intended to serve as a context for the condition), and task materials. Only the
experimenter and the participant were present in the room during sessions. A second experimenter collected data through a oneway observation panel for interobserver agreement evaluations. As in Experiment 1, contextual stimuli were arranged in each phase. For Context A, the stimuli included a yellow T-shirt worn by the experimenter, a yellow poster board (0.9 m by 1.2 m) attached to the wall in front of the participant, and yellow task materials. For Context B, the stimuli included a green T-shirt worn by the experimenter, a green poster board (0.9 m by 1.2 m) attached to the wall in front of the participant, and green task materials. Experimenters did not conduct a preassessment discrimination task for the colors. Any effects of the participants’ discrimination of the colors were not evident until implementation of the experiment. Task materials included previously mastered tasks that were selected by the participants’ caregivers. Caregivers were provided with an array of mastered tasks, such as tracing letters and completing math problems, from which to choose. The purpose of choosing previously mastered tasks was twofold. First, we wished to avoid challenges related to acquisition as a confounding variable. Second, participation in this study provided the opportunity to practice mastered skills more frequently than programmed in the clinic.
TRANSLATIONAL ANALYSIS OF RENEWAL Response Measurement and Interobserver Agreement Across all conditions, experimenters counted correct responses on a laptop computer with a specially designed data-collection program. For Drew, a correct response was matching a picture card to the correct sample, defined as placing a picture in the corresponding bin or placing the picture card on top of the sample card. For John, a correct response was tracing letters and numbers on a worksheet, defined as continuously marking the target letter or number with a pencil within 1 mm of the defined line. All tasks were presented in a free-operant arrangement. We assessed interobserver agreement using two independent observers who collected data either simultaneously with the primary observer or from a video recording. Agreement scores for each session consisted of dividing agreements by agreements plus disagreements in each 10-s interval, converting the result to a percentage, adding each quotient together, and dividing by the number of 10-s intervals. Interobserver agreement was calculated for 30% and 33% of sessions for John and Drew, respectively. Mean agreement was 100% for John and 90% (range, 67% to 100%) for Drew. Procedure As stated above, participants were exposed to three phases, including reinforcement (Context A), extinction (Context B), and extinction (Context A) sequentially. Sessions lasted 5 min. Before participation, a multiple-stimulus-without-replacement (DeLeon & Iwata, 1996) preference assessment was conducted to identify edible items (Fritos and marshmallows) to be delivered contingent on task completion. The experimenter initiated the session by saying, “You may [task] as much or as little as you want.” Reinforcement time (i.e., time to consume the edible items) was not removed from the session time. That is, each session lasted a total of 5 min. The participants typically quickly ate the edible items with little to no interruption in the sessions.
395
Reinforcement (Context A). During this condition, the experimenter presented the participant with a mastered task. Contingent on task completion, a highly preferred edible item was delivered on a fixed-ratio (FR) 1 schedule. All reinforcement sessions were conducted in Context A (i.e., yellow stimuli) for both participants. Extinction (Context B). Procedures in this condition were identical to those in the reinforcement condition except no programmed consequences were provided for task completion (i.e., extinction) and sessions were conducted in Context B (i.e., green stimuli). Extinction (Context A). Procedures in this condition were identical to those in the previous extinction condition except all sessions were conducted in Context A (i.e., yellow stimuli).
RESULTS Experiment 1 Results for Experiment 1 for Pigeons 111 through 116 are depicted in Figure 1. For all pigeons, responding during Context A in Phase 1 (reinforcement baseline) was stable (see mean response rate and ranges depicted in each baseline phase: Ms ¼ 56.2, 27.2, 103.3, 47.6, 59.4, and 67.6 responses per minute, respectively, for Pigeons 111 through 116). In Phase 2, when reinforcement for key pecking was discontinued and Context B was introduced, responding decreased and remained at low levels throughout the phase (Ms ¼ 14.9, 4.5, 20.6, 8.9, 6.2, and 11.9 responses per minute, respectively, for Pigeons 111 through 116). Finally, in Phase 3, when pigeons were returned to Context A, responding increased above extinction levels for all six pigeons, despite the continuation of the extinction contingency (Ms ¼ 35.8, 17.1, 26.5, 39.9, 19.1, and 13.0 responses per minute, respectively, for Pigeons 111 through 116). As expected, response rates for all six pigeons decreased to low or zero levels when key pecking contacted the continued extinction contingency.
396
MICHAEL E. KELLEY et al.
Figure 1. Pigeons’ responses per minute of target responses during Phase 1, Phase 2, and Phase 3 of the renewal procedure in Experiment 1.
Experiment 2 Results for Experiment 2 are depicted in Figure 2. Responses per minute for John and Drew are shown in the left and right panels, respectively. John’s responding in Context A during Phase 1 (reinforcement baseline) was stable (M ¼ 10.2 responses per minute). In Phase 2, when reinforcement for academic work was discontinued and Context B was introduced, responding initially increased above baseline levels and then decreased and remained at zero throughout the phase (M ¼ 5.6). Finally, in Phase 3, when participants were returned to
Context A, responding increased above extinction levels for four sessions, despite the continuation of the extinction contingency. Responding then decreased to zero levels for the remainder of the phase (M ¼ 5.5). Drew’s responding in Context A during Phase 1 (reinforcement baseline) was less stable than John’s and was on an upward trend (M ¼ 5.2 responses per minute). In Phase 2, when reinforcement for academic work was discontinued and Context B was introduced, Drew’s responding decreased to low levels immediately and remained at low or zero levels throughout the phase (M ¼ 0.5). Finally, in Phase 3, when he
TRANSLATIONAL ANALYSIS OF RENEWAL
397
Figure 2. Responses per minute of target responses during Phase 1, Phase 2, and Phase 3 of the renewal procedure in Experiment 2 for John and Drew.
returned to Context A, responding recovered for two sessions, despite the continuation of the extinction contingency, before responding decreased to low or zero levels for the remainder of the phase (M ¼ 1.1). Sessions in Phases 1 through 3 were conducted across 8 days for John and 7 days for Drew. For both John and Drew, Phase 2 began on a different day than Phase 1. For John, Phase 3 began on a different day than Phase 2, and for Drew, Phase 3 was initiated on the same day as the last two sessions of Phase 2. Thus, John’s response recovery (but not Drew’s) in Phase 3 is consistent with both a renewal and a spontaneous recovery interpretation (Lerman, Kelley, Van Camp, & Roane, 1999). However, for John, Sessions 9 and 10 were conducted on 1 day, and Sessions 11 and 12 were conducted on a different day. Similarly, Drew’s Sessions 10 through 13 were conducted on 1 day, Sessions 14 through 17 on another day, and Sessions 18 through 20 on another day. These data show no evidence of spontaneous recovery across both participants, making spontaneous recovery an unlikely explanation for the increases in responding during Phase 3. DISCUSSION The present study assessed ABA renewal with (a) a basic replication using pigeons and (b) a
translational evaluation with children with autism. Results from both experiments generally were consistent with one another and with past research on ABA renewal of operant responding (see Bouton et al., 2011; Nakajima et al., 2000; Podlesnik & Shahan, 2009; Welker & McAuley, 1978). The results of Experiment 1 replicate and extend voluminous basic research on renewal by demonstrating contextual stimulus control by rate of alternating a localized stimulus with pigeons. Much of the previous research was conducted in studies of respondent conditioning with rats (see Bouton, 2004 for a review). Thus, one contribution of this experiment is the “proof of concept” that renewal is robust in both respondent and operant arrangements across different species. Experiment 2 demonstrated that a basic analogue with nonhumans could be translated to humans, producing a renewal effect similar in form and magnitude. If relapse by ABA renewal is general across a range of experimental conditions, the potential for renewal of problem behavior in clinical situations and the need for understanding its determinants should be examined systematically in translational research. These experiments demonstrate the plausibility of arranging basic research questions with translation as a fundamental goal of its outcome (see also Mace et al., 2010). The value of such purposeful
398
MICHAEL E. KELLEY et al.
translational research is to highlight behavioral processes and the likely implications for influencing treatment outcomes. For example, one might collect data on problem behavior in a school or home (Context A) before delivering services in a center-based facility (Context B). Based on the data we collected in Experiments 1 and 2, one might expect that problem behavior will relapse after return to Context A. Translational and applied models of the ABA renewal phenomenon might produce specific predictions about the likelihood and magnitude of relapse and may also produce new technologies for mitigating such effects. Ultimately, the hope is that translational research will set the stage for influencing the design of behavioral treatments to reduce the negative impact of the processes that underlie relapse on treatment outcomes. Programmed translational evaluations such as the current study are particularly important in light of a paucity of research on treatment efficacy in the applied literature that is specifically conducted in the context of treatment relapse. Subsequent to intervention, clinicians seek to ensure that treatment effects are maintained by training multiple individuals to conduct the final intervention. This tactic implicitly protects against resurgence (Lieving & Lattal, 2003; Volkert et al., 2009). That is, clinicians ensure that those who are responsible for conducting the intervention can do so with high levels of treatment integrity so that the intervention continues to function as intended. Protecting against resurgence focuses on the operant contingencies subsequent to treatment. Renewal, on the other hand, provides a framework for understanding the role of stimulus control, rather than operant contingencies, for maintaining treatment results. Like the resurgence effect, the renewal effect is a predictable phenomenon. Translational research on resurgence and renewal provides an opportunity to expand our understanding of behavior and the risks associated with intervention. For example, Lerman and Iwata (1995) provided clinicians
with descriptive data for predicting the likelihood of extinction bursts (i.e., 36% if extinction is used alone; 12% if extinction is combined with differential reinforcement). This information is useful for informing consumers about the probability of an extinction burst, which to consumers might suggest that the intervention is not effective. On the contrary, an extinction burst is a predictable process that occurs as a function of the extinction operation. Additional renewal data may provide clinicians with additional means to predict and counter threats to long-term treatment efficacy. Experiment 2’s results are important for furthering translational research on relapse for several reasons. First, the results are consistent with past research that was conducted with nonhumans in the context of respondent conditioning. This is an example of systematic replication (Sidman, 1960/1988). Arguably, systematic replication has the potential to more quickly advance scientific knowledge because several key variables change at once instead of in incremental steps. Second, the results of Experiments 1 and 2 were consistent with one another. This is important because of the systematic, collaborative nature of the research. That is, we first asked a basic research question: Would the renewal phenomenon be apparent with operant behavior using nonhumans? When the renewal phenomenon was apparent under typical experimental conditions, we asked a translational question: Would renewal be apparent with operant behavior with humans and well-controlled responses and reinforcers? Taken together, the data in Experiments 1 and 2 suggest a measure of generality of the processes that underlie renewal, provide a platform to study these processes further, and offer a potential platform for assessing questions relevant to treatment maintenance. The study of variables that influence treatment maintenance is critical for the simple reason that it often eludes behavior analysts in at least two ways. First, behavioral research typically
TRANSLATIONAL ANALYSIS OF RENEWAL does not include data collected beyond the final treatment phase (see Nevin & Wacker, 2013 for a discussion of the implications). For example, in 115 studies published in the Journal of Applied Behavior Analysis in 2012 and 2013, only 15 (13%) reported generalization data. None reported data on long-term efficacy of clinical practices. (We thank Henry S. Roane for providing these descriptive data.) Reliable reporting of long-term treatment outcomes is the first step toward understanding and developing technologies for combating breakdowns in treatment maintenance. Second, parents, teachers, and caregivers often report that treatments do not continue to be effective over time and conditions. In practice, clinicians may invoke “failure to generalize” as a conceptual framework for understanding why treatments may fail over time. Stokes and Baer (1977) categorized techniques that were designed to program for generalization, thus providing a framework for ensuring that treatments generalize beyond the training conditions. Some of the techniques seem particularly relevant to renewal. For example, one may “train loosely” in Phase B for the return to Phase A after treatment by allowing multiple stimulus conditions to be present during Phase B treatment (Thomas, Larsen, & Ayres, 2003) or by including some components of Phase A in Phase B. Alternatively, one may program common stimuli by systematically ensuring that common stimuli presented in Phase B are present when returning to Phase A after treatment (see Brooks & Bouton, 1994; Collins & Brandon, 2002) or across a range of settings in which problem behavior occurred (see Piazza, Hanley, & Fisher, 1996). These approaches provide potential avenues for reducing renewal in clinical situations and, therefore, are worthy of additional study. Our research strategy was influenced by Mace (1994) in which he suggested a threefold, interconnected approach for engaging in translational research. Mace suggested that behavior analysis, like all natural sciences, benefits from
399
specific, programmatic, collaborative research between basic and applied sectors. First, basic questions are asked with a specific application in mind, typically with nonhumans as subjects. Second, those basic findings are replicated with humans with controlled conditions, responses, and reinforcers. Finally, the generality of the findings are tested with humans and behaviors of social significance. There are inherent advantages to this strategy. First, informed basic research questions may be better translated and applied than questions that are not informed by problems of social significance. Second, problems of social significance that are not easily studied under natural conditions may be solved in the laboratory while both protecting humans and adding to the scientific knowledge base. For example, Mace et al. (2010) developed an animal model of a treatment that was ultimately designed to solve a serious problem behavior. The model was then tested with nonhumans in a basic arrangement (see also Podlesnik et al., 2012). After the operating principles had been identified, Mace et al. designed a treatment based on the findings of the animal model. Thus, the model provided guidance for application, rather than a trial-and-error approach with humans who might engage in responses that could be dangerous to self or others. One limitation of this study is the lack of the third component of the three-component research strategy. That is, we did not test the renewal phenomenon with humans and problems of social significance. Had we included all components, as suggested by Mace (1994) we might have identified a client who engaged in problem behavior in one environment, executed a treatment in another environment, and then returned the client to the initial environment. Future researchers could assess the renewal phenomenon with behaviors of social significance to satisfy the three-component research strategy. We intend for the current study to serve as a model for translational research of other types of
MICHAEL E. KELLEY et al.
400
treatment relapse (e.g., resurgence, reinstatement), among other areas that might fit into this conceptual framework. If this conceptual framework proves to be insufficient to account for a given set of basic or applied findings, the framework nevertheless was useful in highlighting the involvement of additional behavioral processes that might be accounted for by other conceptual or theoretical frameworks (e.g., behavioral momentum theory; see Nevin & Wacker, 2013). Finally, we hope that the present study occasions additional research for extending basic research findings into translational research and translational findings into application for problems of social significance. REFERENCES Beavers, G. A., Iwata, B. A., & Lerman, D. C. (2013). Thirty years of research on the functional analysis of problem behavior. Journal of Applied Behavior Analysis, 46, 1–21. doi: 10.1002/jaba.30 Bouton, M. E. (1993). Context, time, and memory retrieval in the interference paradigms of Pavlovian learning. Psychological Bulletin, 114, 80–99. doi: 10.1037// 0033-2909.114.1.80 Bouton, M. E. (2002). Context, ambiguity, and unlearning: Sources of relapse after behavioral extinction. Biological Psychiatry, 52, 976–986. doi: 10.1016/S0006-3223 (02) 01546-9 Bouton, M. E. (2004). Context and behavioral processes in extinction. Learning and Memory, 11, 485–494. doi: 10.1101/lm.78804 Bouton, M. E., Todd, T. P., Vurbic, D., & Winterbauer, N. E. (2011). Renewal after the extinction of free operant behavior. Learning and Behavior, 39, 57–67. doi: 10.3758/s13420-011-0018-6 Bouton, M. E., Winterbauer, N. E., & Todd, T. P. (2012). Relapse processes after the extinction of instrumental learning: Renewal, resurgence, and reacquisition. Behavioural Processes, 90, 130–141. doi: 10.1016/j. beproc.2012.03.004 Brooks, D. C., & Bouton, M. E. (1994). A retrieval cue for extinction attenuates response recovery (renewal) caused by a return to the conditioning context. Journal of Experimental Psychology, 20, 366–379. doi: 10.1037//0097-7403.20.4.366 Carroll, R. A., Kodak, T. & Fisher, W. W. (2013). An evaluation of programmed treatment-integrity errors during discrete-trial instruction. Journal of Applied Behavior Analysis, 46, 379–394. doi: 10.1002/jaba.49 Collins, B. N., & Brandon, T. H. (2002). Effects of extinction context and retrieval cues on alcohol cue
reactivity among nonalcoholic drinkers. Journal of Consulting and Clinical Psychology, 70, 390–397. doi: 10.1037//0022-006x.70.2.390 DeLeon, I. G., & Iwata, B. A. (1996). Evaluation of a multiple-stimulus presentation format for assessing reinforcer preferences. Journal of Applied Behavior Analysis, 29, 519–533. doi: 10.1901/jaba.1996.29-519 Gresham, F. M., Gansle, K., & Noell, G. H. (1993). Treatment integrity in applied behavior analysis with children. Journal of Applied Behavior Analysis, 26, 257–263. Leitenberg, H., Rawson, R. A., & Mulick, J. A. (1975). Extinction and reinforcement of alternative behavior. Journal of Comparative and Physiological Psychology, 88, 640–652. doi: 10.1037/h0076418 Lerman, D. C., & Iwata, B. A. (1995). Prevalence of the extinction burst and its attenuation during treatment. Journal of Applied Behavior Analysis, 28, 93–94. doi: 10.1901/jaba.1995.28-93 Lerman, D. C., Iwata, B. A., Smith, R. G., Zarcone, J. R., & Vollmer, T. R. (1994). Transfer of behavioral function as a contributing factor in treatment relapse. Journal of Applied Behavior Analysis, 27, 357–370. doi: 10.1901/ jaba.1994.27-357 Lerman, D. C., Kelley, M. E., Camp, C. M. V., & Roane, H. S. (1999). Effects of reinforcement magnitude on spontaneous recovery. Journal of Applied Behavior Analysis, 32, 197–200. doi: 10.1901/jaba.1999.32-197 Lieving, G. A., & Lattal, K. A. (2003). Recency, repeatability, and reinforcer retrenchment: An experimental analysis of resurgence. Journal of the Experimental Analysis of Behavior, 80, 217–233. doi: 10.1901/jeab.2003.80-217 Luczynski, K. C., Hanley, G. P., & Rodriguez, N. M. (2014). An evaluation of the generalization and maintenance of functional communication and selfcontrol skills with preschoolers. Journal of Applied Behavior Analysis, 47, 246–263. doi: 10.1002/jaba.128 Mace, F. C. (1994). Basic research needed for stimulating the development of behavioral technologies. Journal of the Experimental Analysis of Behavior, 61, 529–550. doi: 10.1901/jeab.1994.61-529 Mace, F. C., McComas, J. J., Mauro, B. C., Progar, P. R., Taylor, B., Ervin, R., & Zangrillo, A. N. (2010). Differential reinforcement of alternative behavior increases resistance to extinction: Clinical demonstration, animal modeling, and clinical test of one solution. Journal of the Experimental Analysis of Behavior, 93, 349–367. doi: 10.1901/jeab.2010.93-349 Murphy, E. S., McSweeney, F. K., Smith, R. G., & McComas, J. J. (2003). Dynamic changes in reinforcer effectiveness: Theoretical, methodological, and practical implications for applied research. Journal of Applied Behavior Analysis, 36, 421–438. doi: 10.1901/jaba.2003.36-421 Nakajima, S., Tanaka, S., Urushihara, K., & Imada, H. (2000). Renewal of extinguished lever-press responses upon return to the training context. Learning and
TRANSLATIONAL ANALYSIS OF RENEWAL Motivation, 31, 416–431. doi: 10.1006/lmot.2000. 1064 Nevin, J. A., & Wacker, D. P. (2013). Response strength and persistence. In G. J. Madden, W. V. Dube, T. D. Hackenberg, G. P. Hanley, & K. A. Lattal (Eds.), APA handbook of behavior analysis (Vol. 2, pp. 109–128). Washington, DC: American Psychological Association. doi: 10.1037/13938-005 Piazza, C. C., Hanley, G. P., & Fisher, W. W. (1996). Functional analysis and treatment of cigarette pica. Journal of Applied Behavior Analysis, 29, 437–450. doi: 10.1901/jaba.1996.29-437 Podlesnik, C. A., Bai, J. Y. H., & Elliffe, D. (2012). Resistance to extinction and relapse in combined stimulus contexts. Journal of the Experimental Analysis of Behavior, 98, 169– 189. doi: 10.1901/jeab.2012.98-169 Podlesnik, C. A., & Kelley, M. E. (2014). Resurgence: Response competition, stimulus control, and reinforcer control. Journal of the Experimental Analysis of Behavior, 102, 231–240. doi: 10.1002/jeab.102 Podlesnik, C. A., & Shahan, T. A. (2009). Behavioral momentum and relapse of extinguished operant responding. Learning and Behavior, 37, 357–364. doi: 10.3758/LB.37.4.357 St. Peter Pipkin, C., Vollmer, T. R., & Sloman, K. N. (2010). Effects of treatment integrity failures during differential reinforcement of alternative behavior: A translational model. Journal of Applied Behavior Analysis, 43, 47–70. doi: 10.1901/jaba.2010.43-47 Sidman, M. (1960). Tactics of scientific research: Evaluating experimental data in psychology. New York, NY: Basic Books. (Reprinted by Authors Cooperative, Boston, MA, 1988) Stephenson, K. M., & Hanley, G. P. (2010). Preschoolers’ compliance with simple instructions: A descriptive and experimental evaluation. Journal of Applied Behavior Analysis, 43, 229–247. doi: 10.1901/jaba.2010.43-229 Stokes, T. F., & Baer, D. M. (1977). An implicit technology of generalization. Journal of Applied Behavior Analysis, 10, 349–367. doi: 10.1901/jaba.1977.10-349 Sundberg, M. L. (2008). Verbal behavior milestones assessment and placement program: The VB-MAPP. Concord, CA: AVB Press.
401
Thomas, B. L., Larsen, N., & Ayres, J. J. B. (2003). Role of context similarity in ABA, ABC, and AAB renewal paradigms: Implications for theories of renewal and for treating human phobias. Learning and Motivation, 34, 410–436. doi: 10.1016/S0023-9690(03) 00037-7 Todd, T. P., Winterbauer, N. E., & Bouton, M. E. (2012). Effects of the amount of acquisition and contextual generalization on the renewal of instrumental behavior after extinction. Learning and Behavior, 40, 145–157. doi: 10.3758/s13420-011-0051-5 Volkert, V. M., Lerman, D. C., Call, N. A., & TrosclairLasserre, N. (2009). An evaluation of resurgence during treatment with functional communication training. Journal of Applied Behavior Analysis, 42, 145–160. doi: 10.1901/jaba.2009.42-145 Wacker, D. P., Harding, J. W., Berg, W. K., Lee, J. F., Schieltz, K. M., Padilla, Y. C., … Shahan, T. A. (2011). An evaluation of persistence of treatment effects during long-term treatment of destructive behavior. Journal of the Experimental Analysis of Behavior, 96, 261–282. doi: 10.1901/jeab.2011.96-261 Welker, R. L., & McAuley, K. (1978). Reductions in resistance to extinction and spontaneous recovery as a function of changes in transportational and contextual stimuli. Animal Learning and Behavior, 6, 451–457. doi: 10.3758/BF03209643 Wilder, D. A., Atwell, J., & Wine, B. (2006). The effects of varying levels of treatment integrity on child compliance during treatment with a three-step prompting procedure. Journal of Applied Behavior Analysis, 39, 369–373. doi: 10.1901/jaba.2006.144-05 Winterbauer, N. E., & Bouton, M. E. (2010). Mechanisms of resurgence of an extinguished instrumental behavior. Journal of Experimental Psychology: Animal Behavior Processes, 36, 343–353. doi: 10.1037/ a0017365
Received June 2, 2014 Final acceptance October 9, 2014 Action Editor, John Borrero
