18052
A Novel fMRI Paradigm for Testing Learning in Adolescents with ASD
Deficits in learning are central to Autism Spectrum Disorders (ASD). One type of learning deficit found in affected individuals is the inability to generalize (or transfer) what they have learned during training to new similar situations. Generalization problems have a profound impact on the academic, social, and adaptive functioning of persons with ASD, and have not been well studied.
Objectives:
We sought to advance our understanding of the cognitive and neural basis of generalization deficits in adolescents with ASD using a transitive inference (TI) paradigm which has been well characterized in typically developing individuals. TI involves learning a series of ordered stimulus pairs (AB, BC, CD, DE, EF where A>B>C>D>E>F), and then transferring or generalizing this learning about order to novel pairs (AC, AD, AE, BD, BE, CE).
Methods:
Participants were 25 medication free adolescents with ASD aged 12-18 evaluated using gold standard ASD diagnostic measures, and 25 age, gender, and IQ matched adolescents with typical development (TYP). We implemented a rapid event-related functional magnetic resonance imaging (fMRI) study of a new TI paradigm for adolescents, with a 5-stimulus pair hierarchy, a game format, frequent feedback, and prizes. Whole brain voxel-wise analyses were conducted using SPM8. We also interrogated regions of interest in the striatum, prefrontal cortex (PFC), and medial temporal lobes, and conducted functional connectivity analyses using these regions as seeds. Based on our prior study of TI (Solomon, Frank, Smith, Ly, & Carter, 2011), we hypothesized that: (1) ASD group performance would be associated with the use of a rote memory strategy involving the hippocampus and visual cortical brain regions; (2) TYP group performance would be associated with recruitment and functional connectivity of the striatum, prefrontal cortex (PFC), and the parietal cortex; and that (3) task performance would be related to math and reading achievement test scores.
Results:
Both groups learned the task to comparable rates by the end of training. At test, the ASD group outperformed the TYP group on the most difficult previously trained CD pair, suggesting they showed superior rote memory. Whole brain and ROI analyses revealed the ASD group exhibited greater medial temporal lobe activation during training that was related to later inference pair performance. At test, there was less activation in prefrontal regions in the ASD group on inference versus premise pairs. However, there was greater evidence of functional connectivity with visually related brain regions in the ASD versus the TYP group that was associated with task performance suggesting these brain regions were used in a compensatory fashion. Indices of cognitive control deficits were related to deficits in reading comprehension and math problem solving.
Conclusions:
ASD rely on a rote learning-based strategy as opposed to a more flexible one that can incorporate rapid updating of reward contingencies, and integrate this information in the service of generalization. This interpretation is supported by the systems-level computational modeling work of Frank et al. (2004, 2005, 2006), and also is consistent with the underconnectivity theory of ASD proposed by Just, Cherkassy, Keller & Minshew, 2004.