Identifying depression-related topics in smartphone-collected free-response speech recordings using an automatic speech recognition system and a deep learning topic model

Large-scale digital phenotyping: Identifying depression and anxiety indicators in a general UK population with over 10,000 participants

BACKGROUND

Digital phenotyping offers a novel and cost-efficient approach for managing depression and anxiety. Previous studies, often limited to small-to-medium or specific populations, may lack generalizability.

METHODS

We conducted a cross-sectional analysis of data from 10,129 participants recruited from a UK-based general population between June 2020 and August 2022. Participants shared wearable (Fitbit) data and self-reported questionnaires on depression, anxiety, and mood via a study app. We examined correlations between mental health scores and wearable-derived features, demographics, health variables, and mood assessments. Unsupervised clustering was used to identify behavioural patterns associated with depression and anxiety. Furthermore, we employed XGBoost machine learning models to predict depression and anxiety severity and compared the performance using different subsets of features.

RESULTS

We observed significant associations between the severity of depression and anxiety with several factors, including mood, age, gender, BMI, sleep patterns, physical activity, and heart rate. Clustering analysis revealed that participants simultaneously exhibiting lower physical activity levels and higher heart rates reported more severe symptoms. Prediction models incorporating all types of variables achieved the best performance (R2 = 0.41, MAE = 3.42 for depression; R2 = 0.31, MAE = 3.50 for anxiety) compared to those using subsets of variables. Several wearable-derived features were observed to have non-linear relationships with depression and anxiety in the prediction models.

LIMITATIONS

Data collection during the COVID-19 pandemic may introduce biases.

CONCLUSION

This study identified several indicators for depression and anxiety and highlighted the potential of digital phenotyping and machine learning technologies for rapid screening of mental disorders in general populations.

Characterising illness stages and recovery trajectories of eating disorders in young people via remote measurement technology (STORY): a multi-centre prospective cohort study protocol

BACKGROUND

Eating disorders (EDs) are serious, often chronic, conditions associated with pronounced morbidity, mortality, and dysfunction increasingly affecting young people worldwide. Illness progression, stages and recovery trajectories of EDs are still poorly characterised. The STORY study dynamically and longitudinally assesses young people with different EDs (restricting; bingeing/bulimic presentations) and illness durations (earlier; later stages) compared to healthy controls. Remote measurement technology (RMT) with active and passive sensing is used to advance understanding of the heterogeneity of earlier and more progressed clinical presentations and predictors of recovery or relapse.

METHODS

STORY follows 720 young people aged 16-25 with EDs and 120 healthy controls for 12 months. Online self-report questionnaires regularly assess ED symptoms, psychiatric comorbidities, quality of life, and socioeconomic environment. Additional ongoing monitoring using multi-parametric RMT via smartphones and wearable smart rings ('Ōura ring') unobtrusively measures individuals' daily behaviour and physiology (e.g., Bluetooth connections, sleep, autonomic arousal). A subgroup of participants completes additional in-person cognitive and neuroimaging assessments at study-baseline and after 12 months.

DISCUSSION

By leveraging these large-scale longitudinal data from participants across ED diagnoses and illness durations, the STORY study seeks to elucidate potential biopsychosocial predictors of outcome, their interplay with developmental and socioemotional changes, and barriers and facilitators of recovery. STORY holds the promise of providing actionable findings that can be translated into clinical practice by informing the development of both early intervention and personalised treatment that is tailored to illness stage and individual circumstances, ultimately disrupting the long-term burden of EDs on individuals and their families.