Brain Waves Used to Identify Autism Disorders in Infants
EEGs and machine-learning algorithms may be useful for predicting infant's risk for autism
WEDNESDAY, Feb. 23 (HealthDay News) -- A new noninvasive test, using the standard electroencephalogram (EEG) to compute modified multiscale entropy (mMSE), may be a useful predictor of an infant's risk for autism, according to a study published online Feb. 22 in BMC Medicine.
William J. Bosl, Ph.D., from Harvard Medical School in Boston, and colleagues used mMSE to identify infants at high risk for autism spectrum disorders (ASDs). They recorded resting EEG signals from 79 babies repeatedly between the ages of 6 and 24 months to identify very early markers of autism. Forty-six of the infants had an older sibling with an ASD diagnosis, and the other 33 had no family history of ASDs.
The researchers classified 9-month-old infants into control and high risk for autism groups with accuracy of over 80 percent using MSE. Accuracy of classification for male infants at 9 months was close to 100 percent, and remained high (70 to 90 percent) at 12 and 18 months. For girls, the most accurate classification was at 6 months and declined thereafter.
"Infants in families with a history of ASD have quite different EEG complexity patterns from 6 to 24 months of age that may be indicators of a functional endophenotype associated with ASD risk," the authors write.
One of the study authors is named on a provisional patent application that includes parts of the signal analysis methods evaluated in this study.