Autism is a neurodevelopmental disorder characterized by alterations in the intellectual, social, communication, and behavioral capabilities of an individual, and is rarely detected in children before 24 months of age. Early diagnosis and intervention may be more effective at a younger age. Functional connectivity magnetic resonance imaging (fcMRI) of 6-month old infants may be able to identify brain connection patterns related to at least one of the characteristics of autism, which normally appear at 24 months of age, by using a mathematical model to analyze the neuroimaging data.

Clinical studies published up to December 2018 that used fcMRI to detect autism in infants were reviewed. The literature databases searched included PubMed, Web of Science, the Trip Database, DynaMed, the Cochrane Library, the International Clinical Trials Registry Platform, and ClinicalTrials.gov. Early assessments of fcMRI analysis were identified through the Early Awareness and Alert System of the Agencia de Evaluación de Tecnologías Sanitarias.

Only one prospective study of 59 infants at 6-months of age was retrieved. A fcMRI analysis was performed to identify 2,635 pairs of functional connections from 230 brain regions. The infants were subsequently assessed for autism at 24 months of age using gold standard tests. The functional connections correlated with at least one of the behaviors related to autism evaluated at 24 months of age. Eleven infants (19%) were diagnosed with autism at 24 months. Compared with the gold standard test results, the predictive model achieved the following: sensitivity 0.82 (95% confidence interval [CI]: 0.52 - 0.95); specificity 1.00 (95% CI: 0.93–1.00); positive predictive value 1.00 (95% CI: 0.70–1.00); negative predictive value 0.96 (95% CI: 0.87–0.99); and negative likelihood ratio 0.18 (95% CI: 0.05–0.64). Adverse effects were not reported in the study.

The fcMRI analysis could help in early detection of autism and the development of preventive interventions. However, the evidence is sparse and more well-designed studies are needed.