21738
Sensitivity to Subtle Changes: A Signal Detection Analysis of Memory for Faces, Objects, and Spoken Words in ASD
Individuals with autism spectrum disorder (ASD) demonstrate hyperspecificity of information processing (McClelland, 2000). More detail in initial sensory/perceptual encoding (Mottron et al 1999) could potentially play a causal role in social and communicative deficits (Eigsti & Fein, 2013). A diagnosis of ASD will broadly impact recall for socially-relevant stimuli, but hyperspecific perceptual encoding might impact recognition of stimuli when presented with novel perceptual features.
Objectives:
The current study was designed to assess memory for stimuli that varied in social relevance and sensory modality. We predicted that individuals with ASD would be less accurate than their peers with typical development (TD) in recalling faces and spoken words, relative to non-social objects. The study also assessed the impact of enhanced sensory encoding in ASD on recall of non-meaningful changes in a memory task. We expected the ASD group to show greater effect of non-meaningful changes, across categories.
Methods:
Adolescents ages 12-18 with high-functioning ASD (n=10, IQs > 85) and TD (n=12; no group differences in gender or FSIQ) completed a stimulus recall task. Subjects were presented with, and asked to remember, three blocks of randomly-ordered stimuli: pictures of household objects, pictures of emotionally-neutral faces, and spoken high-frequency words. Order of block presentation was random. After a 30-minute delay, subjects were presented with the original and novel stimuli, and asked to report whether each item had been presented previously (“Old”) or not (“New”), ignoring (with explanation) any changes. Of the Old stimuli, 50% were unchanged; 50% (“Change”) were presented at 30° rotation (Faces, Objects) or spoken by a new talker (Words).
Results:
Accuracy was calculated using d'. Analyses revealed main effects of category and group, and a category-by-group interaction, p’s<.05. Across groups, accuracy was highest for Words, then Objects, and lowest for Faces (see Table 1). While the ASD group was generally less accurate, group differences were significant only for Faces. Second, a within-group analysis of accuracy for Old versus Change stimuli showed a near-significant (p=.07) difference in Word recall, for the TD group only. Across groups, Faces were most difficult to recall, with low accuracy regardless of Change status; recall for Objects was relatively unaffected by small changes. In contrast, the TD group was somewhat less accurate in recalling Change Words, whereas the ASD group showed no such decrement. Our final sample of 30 will be powered to reveal whether these preliminary results hold.
Conclusions:
As expected, the ASD group showed impaired recall; however, this deficit only appeared for Faces, and not for spoken Words. Faces accuracy was very low, obscuring effects of changes in orientation. Contrary to prediction, results for Objects and Words did not suggest heightened specificity of encoding in the ASD group. Indeed, the data suggested that any such effect was limited to the TD group, and to Words only. This suggests either that sensory information is not encoded with hyperspecificity in ASD, or that recall does not benefit from hyperspecific encoding. Future work will aim to disentangle these accounts.