Cargando…

Developmental changes of cortical white–gray contrast as predictors of autism diagnosis and severity

Recent studies suggest that both cortical gray and white-matter microstructural characteristics are distinct for subjects with autism. There is a lack of evidence regarding how these characteristics change in a developmental context. We analysed a longitudinal/cross-sectional dataset of 402 magnetic...

Descripción completa

Detalles Bibliográficos
Autores principales: Bezgin, Gleb, Lewis, John D., Evans, Alan C.
Formato: Online Artículo Texto
Lenguaje:English
Publicado: Nature Publishing Group UK 2018
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6240045/
https://www.ncbi.nlm.nih.gov/pubmed/30446637
http://dx.doi.org/10.1038/s41398-018-0296-2
Descripción
Sumario:Recent studies suggest that both cortical gray and white-matter microstructural characteristics are distinct for subjects with autism. There is a lack of evidence regarding how these characteristics change in a developmental context. We analysed a longitudinal/cross-sectional dataset of 402 magnetic resonance imaging (MRI) scans (171 subjects with autism and 231 with typical development) from the Autism Brain Imaging Data Exchange, cohorts I–II (ABIDE-I-II). In the longitudinal sample, we computed the rate of change in the white–gray contrast, a measure which has been related to age and cognitive performance, at the boundary of the cerebral cortex. Then, we devised an analogous metric for the cross-sectional sample of the ABIDE dataset to measure age-related differences in cortical contrast. Further, we developed a probabilistic model to predict the diagnostic group in the longitudinal sample of the cortical contrast change data, using results obtained from the cross-sectional sample. In both subsets, we observed a similar overall pattern of greater decrease within the autistic population in intensity contrast for most cortical regions (81%), with occasional increases, mostly in primary sensory regions. This pattern correlated well with raw and calibrated behavioural scores. The prediction results show 76% accuracy for the whole-cortex diagnostic prediction and 86% accuracy in prediction using the motor system alone. Our results support a contrast change analysis strategy that appears sensitive in predicting diagnostic outcome and symptom severity in autism spectrum disorder, and is readily extensible to other MRI-based studies of neurodevelopmental cohorts.