Unveiling Depression: Monitoring Daily Activity Changes for Anomalies

According to the world health organisation (WHO), 1 in 26 people globally suffer from depression, many of which go undiagnosed. The daily actions of individuals may reflect the presence of depressive symptoms, yet these signs could be disregarded due to negative social perceptions or insufficient awareness, resulting in untreated symptoms. Efforts to enhance the identification of depressive symptoms propose utilizing data from mobile devices and wearables rather than relying solely on clinical assessments and self-report questionnaires like the patient health questionnaire (PHQ). However, current methods often concentrate on singular aspects, which may restrict their efficacy. This study introduces an innovative method for identifying depressive symptoms by analyzing various data collected from mobile devices, including information on location, sleep patterns, phone usage, and call durations. A discrete mobile application was developed to collect and analyze this data from participants, guaranteeing privacy and confidentiality. Additionally, both participant-generated data and a benchmark dataset were analyzed using multiple anomaly detection algorithms, allowing for robust insights into depressive symptom identification. Results decisively demonstrated a significant association: fewer visited locations correlated with heightened depression severity among both app users and dataset participants. This underscores the potential of multifaceted data analysis in mental health research and intervention strategies.