Atypical Classical Conditioning in Children with Autism Spectrum Disorder

Background:  

Previous research examining classical conditioning in individuals with autism spectrum disorder (ASD) has indicated a pattern of both impaired (Gaigg & Bowler, 2007; South et al., 2013) and intact learning (Bernier et al., 2005; Sears et al., 1994; South et al., 2011).  One explanation that may account for these findings is that individuals with ASD may acquire a conditioned response (CR) at a slower rate compared to individuals with typical development.

Objectives:  

The primary objective of this study was to examine classical conditioning and the rate of acquisition (i.e., number of trials needed to demonstrate a conditioned response) in children with ASD compared to age and IQ matched children with typical development. Additionally, we examined whether differences in the rate of extinction were similar across both diagnostic groups.

Methods:  

Nineteen children with ASD and 16 age and IQ matched children with typical development (age range: 8 to 14 years) participated in a classical conditioning paradigm (data collection ongoing).  This task consisted of three phases (habituation, acquisition, and extinction).  During the habituation phase participants were shown two different colored squares.  During the acquisition phase one color (i.e., the conditioned stimulus; CS) was paired with a loud aversive sound (i.e., the unconditioned stimulus; UCS), with 100% reinforcement.  Finally, during the extinction phase participants were shown four presentations of the CS without presentation of the loud noise. Participants’ skin conductance responses (SCR’s) were recorded throughout the task.

Results:  

Our findings showed a significant increase in SCR to the CS compared to the neutral stimulus in both individuals with ASD and individuals with typical development, F(1,32) = 13.94, p = .001.  However, despite similar overall group conditioning, a significant diagnosis (ASD vs. typical) by trial order (i.e., CS-UCS trial pairings) interaction was found, F(1, 33) = 4.64, p = .04.  This finding showed that individuals with ASD needed a greater number of trials in order to learn the association between the CS and UCS. Analysis of CS trials during the extinction phase also revealed a significant diagnosis by pairing (CS vs. Neutral Stimulus) interaction, F(1, 33) = 5.13, p = .03, showing that individuals with ASD demonstrated a smaller CR compared to individuals with typical development after reinforcement was no longer provided (i.e., faster extinction).

Conclusions:  

The present findings suggest that although individuals with ASD demonstrated a conditioned response, more trials were needed in order to demonstrate a reliable conditioned response that was similar to children with typical development.  Analysis of the extinction phase also revealed that individuals with ASD were less likely to respond to the CS once reinforcement was no longer provided.  These findings suggest not only a slower rate of associative learning in ASD, but also a greater likelihood of losing this learning once the reinforcement is no longer provided. Many theories of ASD suggest impairments in complex higher-order cognitive processes, however the present study demonstrates that the impairment in ASD may be a much more basic associative learning impairment which may underlie many cognitive impairments observed in this disorder.