Behavior bests brain as an Autism Spectrum Disorder biomarker

Monday, September 22, 2014 — Poster Session I
12:00 p.m. – 2:00 p.m.	FAES Academic Center	NIMH	NEURO-23

Authors

• M Plitt
• KA Barnes
• GL Wallace
• IW Eisenberg
• B Orionzi
• L Kenworthy
• SJ Gotts
• A Martin

Abstract

Autism Spectrum Disorders (ASD) are diagnosed based on early-manifesting clinical symptoms, including markedly impaired social communication. We ascertained whether resting-state functional connectivity MRI (rs-fMRI) could serve as possible ASD biomarkers and, if so, which brain features were most informative about diagnosis. We investigated the best methods for performing classification of rs-fMRI scans, compared these best methods to classification using simple behavioral measures (Social Responsiveness Scale; SRS), and interrogated the large-scale brain networks which predict an ASD diagnosis. High classification accuracy was achieved through several methods (peak accuracy 76.67%). However, classification via behavioral measures (peak accuracy 95.19%) consistently surpassed rs-fMRI classifiers. The class probability estimates, P(ASD|fMRI data), of the top performing brain-based classifiers significantly correlated with scores on a measure of social functioning, SRS, for all regions sets, as did the most informative features from 2 sets of brain-based features. The most informative connections predominantly originated from regions strongly associated with social functioning. While individuals were classified as having ASD with a high degree of accuracy from rs-fMRI scans alone, behavioral measures are significantly more accurate. Classification methods provided further evidence that ASD functional connectivity is characterized by dysfunction of large-scale functional networks, particularly those involved in social information processing.

back to top