Association for Behavior Analysis International

This symposium features a series of studies aimed at extending the scope of functional analysis methodology by (a) evaluating the interaction between behavior function and psychotropic medication, (b) validating a set of patterns of responding characteristic of functional analyses of behavior maintained by automatic reinforcement, and (c) evaluating the criterion validity of two indirect approaches to functional assessment. The first presentation compiles a series of studies on the interaction between behavior function and psychotropic medication based on the analysis of 40 published datasets of side-by-side functional analysis conducted in the absence and the presence of specific medication manipulations. The second study presents a set of patterns of functional analysis data often found during the assessment of problem behavior maintained by automatic reinforcement. These patterns will be validated by way of a thorough re-analysis of the literature and a replication under consistent assessment conditions. The patterns identified suggest that specific treatments may be highly effective for specific individuals. We will present several treatment studies to illustrate the extent to which these patterns predict the effects of specific treatment components. Finally, the last study will present a set of concurrent indirect and functional analyses. Specifically, we will evaluate the validity of the Motivation Assessment Scale and the Questions About Behavior Function against experimental functional analyses. Each study will end with a set of clinical recommendations supported by our findings.

This study will examine the criterion validity of the Questions About Behavior Function (QABF) and the Motivation Assessment Scale (MAS) using experimental functional analysis as a reference for comparison. A total of 15 cases of problem behavior presented in individuals with developmental or intellectual disabilities have participated in the study while and a smaller number are currently being assessed. Caregivers of all participants completed the QABF and the MAS. Whenever possible, two caregivers of an individual participant were requested to independently complete both indirect assessments. All participants underwent a full functional analysis according to the methods by Iwata et al. (1982/1994). Level of correspondence between the indirect and the experimental assessments (criterion validity) was computed as the number of cases that showed the same outcome in the indirect assessment and the functional analysis divided by the total number of cases and converting this ratio into a percentage. Potential outcomes included behavior maintained by social positive reinforcement, social negative reinforcement, tangible reinforcement, automatic reinforcement, multiply controlled behavior, and undifferentiated assessment. We also evaluated the level of agreement across caregivers, the concurrent validity of the two indirect assessments, and the incremental validity resulting from the implementation of two, as opposed to one, indirect assessments. We will discuss a set of clinical recommendations supported by our findings.

Several patterns of responding during a functional analysis (FA) are consistent with an automatic reinforcement outcome: (a) low rates during only the attention condition, (b) low rates during only the demand condition, (c) low rates during only the play condition, (d) distinctively higher rates during the alone condition, and (e) similar levels of behavior across all conditions. These patterns are consistent with the action of specific behavioural processes. Namely, the patterns presented above could be explained by way of positive punishment (contingent statements, response blocking), social negative punishment (timeout from activities/attention), extinction (response blocking), noncontingent reinforcement (attention and/or leisure/task items), and differential reinforcement of alternative behavior (via response competition). Therefore, patterns of responding in a full FA of behaviour maintained by automatic reinforcement could be useful in selecting an optimal treatment strategy. The purposes of the present study were to: (a) validate the proposed patterns on the basis of a re-analysis of the literature, (b) replicate the proposed patterns under consistent assessment conditions, and (c) determine, by way of a component analysis, if these patterns help to establish the likely effects of a range of treatments probed during the FA. Our results show that the proposed patterns are present in the literature and can be replicated under consistent assessment conditions. Moreover, the component analysis revealed that discrete behavioural processes drive the reductive effect of specific FA conditions that is characteristic of some of the proposed patterns.

Treatment for problem behaviors typically includes behavioral, pharmacological, and combined interventions. Given that combined interventions are relatively common, analyzing the interactions between behavior function and psychotropic drugs may have some clinical value for practitioners. We reanalyzed all published cases of side-by-side functional analyses conducted in the absence and presence of a target medication manipulation. We identified 40 cases. Medication used included antipsychotics, antidepressants, stimulants, anxiolytics, and opioid antagonists. We conducted three studies in order to determine if: (a) medication had an overall effect on behavior, (b) medication changed the function of the behavior, and (c) medication changed the level of differentiation of the functional analysis. Study 1 assessed the overall effects of psychotropic drugs on problem behavior. Medication had reductive effects in 34 of the 40 datasets reviewed. The magnitude of the medication effect was associated with the baseline level of responding across conditions of the functional analysis (rate-dependency). Study 2 examined drug-induced changes in behavior function. For 34 out of the 40 datasets reviewed medication did not change behavior function. Only two cases showed a potential addition of a new behavior function. Finally, Study 3 evaluated drug-induced changes in the level of differentiation of the condition of the functional analysis showing behavioral maintenance. The analyses showed that the differentiation of the target condition were primarily rate-dependent. No clear effects of medication on function differentiation were established. We will discuss clinical recommendations based on our findings.