Interest in Potential Reinforcers in the Second Year of Life Predicts Outcome of Behavioral Intervention in Toddlers with ASD

Background:  There is substantial variation in language and intellectual outcomes in children with ASD. A recent study suggests that the number of stimuli that children find reinforcing may constitute a predictor of response to behavioral treatment in preschool-aged children with ASD (Klintwall & Eikeseth, 2012). However, very little is known if a similar effect can be found in children presenting for a differential diagnosis of ASD in the second year of life.  Advancing our understanding of factors that affect amenability to treatment at the very early stages of ASD emergence would improve our ability to individualize treatments very early on and to enhance outcome in the affected children. Moreover, in contrast with our original work, which was based on informant report, in this study we intended to design and test a coding system based on children’s responses to play probes in the context of semi-structured interaction with an examiner.

Objectives:  To address this gap we investigated child interest in objects and activities during the Autism Diagnostic Observation Schedule-Toddler Module (ADOS-T; Lord et al, 2000) assessment as a predictor of subsequent development measured by adaptive behavior and verbal and non-verbal developmental quotient (DQ) at ages 2-4 years. 

Methods:  Reinforcing function of stimuli can be approximated from exhibited affect and attention towards the object or activity. Based on our previous work (Klintwall & Eikeseth, 2012), we developed a novel scoring system for capturing toddlers’ interest in potential reinforcers (e.g. edibles, toys): the Assessment of Volition and Object Curiosity (AVOC). The scoring system yields a mean interest score. AVOC was used to score video recordings of 70 toddlers with ASD who underwent assessment between 14 and 32 months (Time 1). Developmental (Mullen Scales of Early Learning; Mullen, 1995) and adaptive skills (Vineland Adaptive Behaviors Scales-II; Sparrow, et al., 2005) were measured concurrently and then again when the children were 21 to 49 months old (Time 2).  Between measurements, the children received a variety of community-based behavioral and developmental interventions. 

Results:  The AVOC scoring system exhibited acceptable levels of test-retest (0.80) and inter-rater reliability (0.90).  AVOC score, ADOS-T total score, Mullen verbal and nonverbal DQ, and Vineland adaptive full scale score (Adaptive Behavior Composite; VABC) was entered into linear regression models at Time 1 to predict outcome at Time 2 defined as VABC, verbal DQ and non-verbal DQ.  The AVOC score was the only significant predictor of Time 2 adaptive behavior (p=.001, R²=.394) as well as non-verbal DQ (p=.024, R²=.277), outperforming Time 1 verbal DQ (p=.216 and p=.226, respectively), non-verbal DQ (p=.652 and p=.856) and ADOS-T scores (p=.305 and p=.285).  AVOC score at Time 1 also contributed, along with non-verbal DQ, to variance in verbal DQ at Time 2 (p=.009, R²=.318). 

Conclusions:  

Interest in potential reinforcers was a powerful predictor of subsequent development for toddlers with ASD. Future studies should aim to evaluate and develop techniques for identifying and expanding the repertoire of stimuli functioning as reinforcers for children who score low on the AVOC, thus possibly improving their outcomes.