A biomarker discovery framework for childhood anxiety

Area of Science:

• Neuroscience
• Developmental Psychology
• Biomarkers

Background:

• Anxiety is a common youth psychopathology impacting functioning and adult mental health.
• Current diagnoses occur after symptoms emerge, missing early intervention opportunities.
• Electroencephalogram (EEG) offers a window into brain activity for early detection.

Purpose of the Study:

• To identify nonlinear EEG features distinguishing children with anxiety disorders from healthy controls.
• To explore EEG-based detection of externalizing disorders in youth.
• To determine optimal ages for EEG data collection for disorder identification.

Main Methods:

• A novel supervised tensor factorization method was used to extract latent factors from nonlinear EEG measures.
• Longitudinal EEG data from infancy to 7 years were analyzed.
• Distinguishing children with anxiety and externalizing disorders from controls using cross-validation.

Main Results:

• EEG recordings at age 7 years, combined with earlier data (ages 5 & 7, or 3, 5, & 7), were necessary to distinguish anxiety disorders.
• Externalizing disorders were detectable using EEG data from ages 3 and 5 years.
• Sex and birthweight were significant covariates, while infant EEG data did not improve classification.

Conclusions:

• Latent factors from childhood EEG recordings show potential as biomarkers for anxiety and externalizing disorders.
• Specific age ranges and combinations of EEG recordings are crucial for accurate disorder identification.
• This approach may enable earlier intervention for youth mental health conditions.