UVM KID Study: Identifying Multimodal Features and Optimizing Wearable Instrumentation to Detect Child Anxiety

Area of Science:

• Pediatric Psychology
• Biomedical Engineering
• Digital Health

Background:

• Internalizing disorders (anxiety, depression) affect 1 in 5 children, starting in preschool.
• Untreated, these disorders lead to severe long-term health issues, including substance abuse and suicide risk.
• Current diagnostic methods are time-consuming, require expert access, and rely on caregiver observation.

Purpose of the Study:

• To develop an assessment battery for a digital phenotype of internalizing disorders in young children.
• To conduct an early feasibility study using multi-modal wearable sensor data.
• To explore the potential of wearable sensors for rapid, point-of-care diagnostics.

Main Methods:

• Utilized multi-modal wearable sensors measuring movement and surface biopotentials (chest, trapezius).
• Collected data during a short stress-induction task.
• Analyzed sacral movement responses and R-R interval variability.

Main Results:

• Sacral movement and R-R interval data showed potential for facilitating child diagnosis.
• Redundancy was observed between chest and trapezius biopotential sensors, suggesting optimization possibilities.
• Feasibility demonstrated with data from two participants, one with a clinical diagnosis.

Conclusions:

• Multi-modal wearable sensors show promise for developing digital phenotypes of childhood internalizing disorders.
• Specific physiological signals (movement, R-R interval) may aid in early identification.
• Further optimization of sensor placement, signals, and features is needed for clinical deployment.