Adolescent Mood Disorders & Digital Phenotyping

Real-Time Mood State Detection Via Smartphone Behavioral Phenotyping

This research investigates computational algorithms that extract mood states from digital behavioral markers including typing patterns, screen interaction velocity, and app usage temporal dynamics. The investigation advances affective computing by establishing validated digital biomarkers for continuous, non-intrusive mood monitoring in ecological contexts.

Temporal Dynamics of Social Media Engagement and Depressive Episode Trajectories

Research examines longitudinal associations between granular social media interaction patterns and prospective trajectories of depressive symptom severity in adolescent cohorts. This work elucidates bidirectional temporal relationships between digital behavioral signatures and affective state fluctuations.

Circadian Disruption Markers and Mood Disorder Risk Stratification

This investigation uses passive sensor data from smartphones and wearables to quantify sleep-wake cycle irregularities and their mechanistic links to adolescent anxiety and depression onset. The research identifies early circadian phenotypes that stratify risk for developing clinical mood disorders.

Natural Language Processing Sentiment Analysis in Encrypted Digital Communications

Research develops privacy-preserving NLP methodologies to extract affective content from adolescent digital communications while maintaining end-to-end encryption and ethical data governance standards. This work bridges the gap between computational phenotyping capability and ethical research implementation in sensitive populations.

Multimodal Sensor Integration for Heterogeneous Depression Phenotype Classification

This research integrates data from accelerometers, heart rate variability sensors, GPS geolocation, and keystroke dynamics to identify distinct biological and behavioral subtypes of adolescent depression. The work advances precision psychiatry by revealing depression''s neurobiological heterogeneity through digital phenotypic clustering.

Predictive Machine Learning Models for Suicidal Ideation Trajectory Detection

Investigation employs longitudinal machine learning approaches trained on multimodal digital phenotypes to prospectively identify adolescents at elevated risk for suicidal ideation escalation. This research produces clinically actionable predictive instruments that operationalize real-time digital monitoring for suicide prevention.

Affective Computing Architectures for Wearable Emotion Recognition Validation

Research develops and validates deep learning architectures that infer momentary emotional states from wearable biosignals including electrodermal activity, photoplethysmography, and inertial measurement unit data. This work establishes psychometric properties and clinical validity of wearable-derived affect inference in naturalistic adolescent environments.

Algorithmic Bias Assessment in Mood Disorder Digital Phenotype Development

This investigation examines how demographic factors, digital literacy, and technology access patterns introduce systematic bias into digital phenotyping algorithms for adolescent mood disorders. The research produces fairness-aware methodologies that ensure digital biomarkers maintain predictive validity across sociodemographically diverse populations.

Passive Smartphone Geolocation Patterns and Social Withdrawal Symptomatology Linkages

Research analyzes passive GPS-derived spatial mobility patterns, location entropy, and place-visitation frequency as digital phenotypes of social anhedonia and withdrawal in adolescent mood disorders. The work establishes geospatial behavioral metrics as objective indicators of clinical symptom severity.

Ecological Momentary Assessment Integration With Passive Digital Phenotyping

This investigation combines ecological momentary assessment of mood self-reports with contemporaneous passive digital phenotypes to validate and calibrate computational mood inference models. The research produces hybrid methodologies that leverage both subjective experience and objective digital behavior for enhanced affective state characterization.