Behavior and interaction imaging at 9 months of age predict autism/intellectual disability in high-risk infants with West syndrome

Area of Science:

• Developmental neuroscience
• Computational psychiatry
• Behavioral pediatrics

Background:

• Current psychiatric assessments face limitations in early detection of developmental disorders.
• West syndrome (WS) is a severe epilepsy associated with significant neurodevelopmental challenges, including autism spectrum disorder (ASD) and intellectual disability (ID).
• Early identification of ASD/ID in high-risk infants is crucial for timely intervention and improved outcomes.

Purpose of the Study:

• To investigate the predictive power of automated behavioral and interactional synchrony analysis in infancy for later ASD/ID diagnosis in West syndrome.
• To identify specific behavioral and interactional features that differentiate infants with West syndrome who develop ASD/ID from those who do not.

Main Methods:

• Recorded 32 infants with West syndrome (WS) and 19 typically developing (TD) controls during mother-infant interactions at 9 months of age.
• Computed infant hand movements (HM), speech turn-taking patterns (vocalization, pause, silences, overlap), and motherese.
• Utilized machine learning algorithms to classify infants based on behavioral and synchrony features, predicting ASD/ID outcomes at 4 years.

Main Results:

• Machine learning models achieved 76.47% accuracy in classifying WS versus TD infants and 81.25% accuracy in distinguishing WS infants who developed ASD/ID (WS+) from those who did not (WS-).
• The top predictive features included infant vocalizations, hand movements, and vocal synchrony patterns between mother and infant.
• Behavioral and interactional imaging successfully predicted ASD/ID in high-risk infants with WS.

Conclusions:

• Automated analysis of infant behavior and mother-infant interactional synchrony at 9 months can predict the later development of ASD/ID in infants with West syndrome.
• Multimodal social signals and interactional synchrony are valuable biomarkers for early risk identification in neurodevelopmental disorders.
• These findings support the use of advanced behavioral analysis tools to overcome diagnostic limitations in pediatric psychiatry and guide early interventions.