Klinger and Dawson (2005) and Klinger, Klinger, and Pohlig (2007) hypothesized that the social and language processing impairments in ASD may be caused by a dysfunction in implicit learning. Implicit learning is learning without the conscious effort to learn, and without the conscious knowledge of what was learned (Reber, 1993), and is thought to underlie the development of social intuition (Lieberman, 2000). Additionally, recent studies have pointed to ASD as a neurological disorder, with anatomical, functional, and connectivity differences associated with some of the behavioral differences in individuals with ASD (Courchesne et al., 2001; Just et al., 2004; Sears et al., 1999). It is theorized that the implicit learning impairments seen in persons with ASD may be due to a more general underlying neural dysfunction evidenced by diminished activation in the basal ganglia, specifically the caudate nucleus, and diminished communication between areas of the brain in persons with ASD, specifically the caudate and medial temporal cortex.
Objectives:
This study examined the relationship between implicit learning deficits in individuals with ASD and parents of persons with ASD and associated differences in brain activation.
Methods:
Twelve high-functioning children and adolescents with ASD and 17 age and verbal ability matched typically developing controls, 10 parents of participants with ASD, and 9 parents of typically developing participants completd this study. The data were collected using a Siemens 3T Allegra fMRI scanner at the Baylor College of Medicine (Houston, TX). While in the scanner, participants viewed twenty-three grammatical letter strings constructed using a Markovian grammar chain (such as “VXVV”). After this initial exposure, participants were told that there were complex rules that defined how the letter strings were formed, and that all the previously seen letter strings followed those rules. During a following test phase, participants saw more letter strings, and were asked to indicate if the strings followed the same rules as those in the exposure phase.
Results:
The grammar effect was calculated as the difference between the proportion of correct responses to grammatical test items and nongrammatical test items. A greater percentage of correct responses to grammatical items during test indicates an automatic learning of the relationships between the letters (the rules), and is taken as evidence of implicit learning. Behaviorally, participants with ASD showed a significantly smaller grammar learning effect (M = +2%) than typically developing participants (M = +16%), t(27) = 2.51, p = .02. Additionally, parents of participants with ASD showed a smaller grammar effect (M = +4%) than parents of typically developing participants (M = +14%), t(17) = 2.93, p < .01. Initial functional analyses reveal less activation in the areas of the anterior cingulate and caudate for participants with ASD compared to typically developing participants in contrasts of grammatical versus nongrammatical stimuli, t(27) = 2.77, p < .005. Similar differences in these areas are also found in the parent groups.
Conclusions:
Diminished activation in the caudate nucleus and cingulate cortex may underlie differences in implicit learning, which are found to be associated with ASD symptomatology.