Association for Behavior Analysis International

The relapse of problem behavior following apparently successful treatment is an enduring problem for the field of behavior analysis. Treatment relapse can occur when, for example, treatment is withdrawn to test for maintenance or when there are lapses in treatment integrity. Behavior momentum theory describes how behavior persists despite changes in the environment. Behavior momentum studies have shown that rate and resistance to change are independent aspects of behavior. Problem behavior is more resistant to change, and thus more likely to relapse in environments that are associated with high rates of reinforcement than in environments associated with low rates, even if an alternative behavior has been reinforced at the same time. Reinstatement, resurgence, and renewal models have been used to demonstrate treatment relapse. An effective method of evaluating the resistance of a behavior to disruption in each of these three models is to use multiple schedules of reinforcement in which differing rates of reinforcement occur in the context of alternating discriminative stimuli. Relapse is greater in the richer of the two multiple schedules. Translational researchers are now developing treatments based on these procedures to successfully reduce the magnitude of the relapse of problem behavior.

The magnitude of treatment relapse has been shown to be greater when high rates of reinforcement are used to reduce problem behavior in both resurgence and reinstatement models. The effects of high rate reinforcement on the relapse of problem behavior within an ABA renewal model was evaluated. The participant was an 18 year-old male with severe intellectual disabilities who presented problem behavior maintained by attention. In the first context (A) problem behavior was reinforced by2 therapists on a multiple schedule of reinforcement. Therapist 1 reinforced problem behavior on a VT 30s schedule and Therapist 2 on a VT 120s schedule until stable responding was achieved. The participant was then moved to a second context (B) and all behaviors were placed on extinction by both therapists in consecutive extinction sessions. The participant was then transferred back to the first context (A) and all behaviors were again placed on extinction by both therapists in consecutive extinction sessions. Results demonstrated that the renewal of problem behavior was greater following high-rate reinforcement, that is, with Therapist 1 (VT 30s), than with Therapist 2 (VT 120s). This result suggests that low rate reinforcement may be more effective in preventing treatment relapse of problem behavior than high rate reinforcement.

Consistent with behavioral momentum theory, arranging alternative sources of reinforcement within a stimulus context decreases target response rates but enhances resistance to disruption. We explored a method developed previously to circumvent enhancing the persistence of the target response while still reducing its rate. Keypecking in pigeons was maintained in 3 mutually exclusive stimulus contexts defined by keylight color. One stimulus context maintained a target response and another stimulus context maintained an alternative response. In a third context, a target response was reinforced concurrently with an alternative response, modeling a differential-reinforcement-of-alternative-behaviour (DRA) schedule. The overall reinforcement rate in the concurrent context was equal to the sum of the separate stimulus contexts. Combining the separate stimulus contexts during extinction enhanced disruption of target responding relative to target responding in the DRA context. In comparison, combining the separate alternative context with the target response from the overall richer concurrent context produced smaller disruptions in target responding. These findings suggest that alternative stimulus contexts function as disrupters of target behavior, and resistance to disruption to the addition of separate stimulus contexts is a function of the overall baseline reinforcement rate in a stimulus context, consistent with the assertions of behavioral momentum theory.