A Translational Examination of Alternative-Response Discrimination Training and Resurgence

Individuals with neurodevelopmental disorders such as autism spectrum disorders often engage in severe forms of problem behavior. Reward-based behavioral interventions are highly effective at reducing levels of problem behavior and teaching more appropriate and adaptive alternative behaviors. Despite successful reduction in problem behavior during treatment, problem behaviors are susceptible to reoccurrence or relapse. Resurgence is a type of behavioral relapse that is particularly relevant to the treatment of problem behavior and may occur following the worsening of conditions of a more recently learned alternative behavior. That is, if the rewards that were used to teach the alternative behavior are removed or lessened, problem behavior may increase as a result. Importantly, resurgence of problem behavior poses a major obstacle for these individuals and their families. Recent clinical research has suggested that resurgence of severe destructive behavior may be prevented using a specific signal to indicate that a particular behavior will not be rewarded. While this may be a promising method for preventing resurgence, the generality of this finding is unknown. Laboratory research with animal subjects is a useful way to study resurgence under highly controlled settings and can provide important information for the development of behavioral interventions in clinical settings. The general procedures of behavioral interventions used in the clinic were approximated in Experiments 1 and 2 with rats as subjects to expand on this previous clinical finding. The goal of Experiment 1 was to determine whether resurgence would still be prevented by the signal that indicates reward unavailability when the reward is removed under conditions in which it was previously available. Contrary to previous findings, the signal for reward unavailability did not prevent resurgence; however, the conditions under which resurgence was tested were different between Experiment 1 and the previous clinical research. The goal of Experiment 2 was to determine whether this difference contributed to the discrepant findings. Resurgence was compared under conditions identical to Experiment 1 as well as conditions that more closely resembled those in the clinic. Resurgence was not differentially impacted by the testing procedures and, importantly, was not reduced under conditions similar to those used in the clinic. These results suggest that the conditions under which this signal may mitigate resurgence are limited and suggest avenues for future research to determine the necessary and sufficient conditions for this effect