Presentation
Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Poster Slams
Reduced phonological working memory in autism is associated with altered speech-motor engagement
Poster A57 in Poster Session A, Thursday, October 6, 10:15 am - 12:00 pm EDT, Millennium Hall
Amanda M. OBrien1,2, Tyler K. Perrachione3, Helen Tager-Flusberg3, John D. E. Gabrieli2, Zhenghan Qi4; 1Harvard University, 2Massachusetts Institute of Technology, 3Boston University, 4Northeastern University
During language development, children rapidly learn new vocabulary through the complex coordination of the neural systems that underlie speech perception, working memory, and speech production. Phonological working memory - the ability to maintain and manipulate speech information - is believed to play an important role in language development, and is often disrupted in many developmental communication disorders, including autism and reading disabilities (Macizo et al., 2016; Peter et al., 2011). Clinicians and researchers often use nonword repetition as a measure of phonological working memory, which has high sensitivity and specificity for developmental language disorders (Estes et al., 2007). Autistic individuals, for example, frequently demonstrate worse performance on nonword repetition tasks (Gabig, 2008; Williams et al., 2006). Little is known about why phonological working memory is disrupted in autistic children, or why their nonword repetition difficulties appear unique from those found in other language disorders (Nadig & Mulligan, 2017; Williams et al., 2013). The primary aim of this study was to identify the brain bases of phonological working memory difficulties in autism. The second aim was to ascertain the specificity of the functional brain differences observed in autistic children during nonword repetition. Children with autism (N = 23, mean age = 12.2, SD = 3.4), children with reading disabilities (N = 16, mean age = 11.5, SD = 3.1) and typically developing children (N = 28, mean age = 11.9, SD = 3.2) participated in this functional magnetic resonance imaging (fMRI) study. While in the fMRI scanner, participants repeated nonwords distributed across four syllable loads (2- to 5-syllables). We used complementary univariate and multivariate pattern analyses to examine whether the atypical phonological working memory abilities seen in autism were associated with functional differences in each of the three key brain networks involved in nonword repetition: speech perception, working memory, and speech production. We then investigated whether the neural differences seen in autistic children were also found in a group of nonautistic children with a reading disability. Compared to the neurotypical group, the autism group and the reading disability groups performed worse on the nonword repetition task (p < 0.0001 for both comparisons). The univariate analyses (cluster-defining threshold p < 0.01, cluster-level FWE-corrected threshold p < 0.05) revealed that the autism group, compared to the neurotypical group, had reduced responses to increasing syllable length in the supplementary motor area. The multivariate pattern analyses showed that decoding accuracy between 2- and 5-syllables was significantly higher within the speech production network for the NT group than the autistic group (t(47) = 2.5, FDR-corrected p < 0.05). Higher decoding accuracy in autism was associated with less severe autism-related symptoms (R’s > 0.3). The atypical motor findings held when ASD and neurotypical groups were equated for language proficiency. The reading disability group with nonword-repetition difficulties, in contrast, did not show such atypicalities in speech-motor functions. Our findings suggest that brain-based alterations in motor planning underlying nonword repetition may uniquely contribute to phonological working memory difficulty in autism.
Topic Areas: Disorders: Developmental, Signed Language and Gesture