Can machine learning identify childhood characteristics that predict future development of bipolar disorder a decade...

Area of Science:

• Child and Adolescent Psychiatry
• Computational Psychiatry
• Developmental Psychology

Background:

• Early identification of bipolar disorder (BD) is crucial for timely intervention, yet reliable prediction in children is lacking.
• Machine learning (ML) offers a promising avenue for developing data-driven predictive models for complex psychiatric conditions.
• Existing research lacks statistically validated methods to predict the onset of BD in pediatric populations.

Purpose of the Study:

• To investigate the feasibility of predicting the development of bipolar disorder in children using baseline clinical data.
• To develop and validate a machine learning model for early identification of pediatric bipolar disorder risk.
• To identify key clinical predictors associated with the future development of bipolar disorder in youth.

Main Methods:

• A longitudinal case-control study involving 492 children (ages 6-18 at baseline) followed for 10 years.
• Data collection included sociodemographic information, psychometric scales, structured diagnostic interviews, and cognitive/social functioning assessments.
• The Balanced Random Forest algorithm was employed to predict the outcome of full or subsyndromal bipolar disorder.

Main Results:

• 10% (45 children) developed bipolar disorder at follow-up.
• The ML model achieved 75% sensitivity, 76% specificity, and an Area Under the ROC curve of 0.75 in predicting BD.
• Key predictors included Child Behavioral Checklist (CBCL) scores for externalizing/internalizing behaviors, overall functioning, anxiety/depression, and aggression.

Conclusions:

• This study presents the first quantitative, ML-based model for predicting childhood-onset bipolar disorder.
• The findings suggest that ML algorithms can effectively identify children at high risk for developing BD.
• This predictive model holds significant potential for clinical application in assessing psychopathology and informing early intervention strategies.