JMIR Publications

ABSTRACT

Despite the global proliferation of mental health screening programs, a substantial proportion of fatal psychiatric outcomes—particularly suicides—occurs in individuals categorized as "low risk" during standard clinical assessments. Retrospective psychological autopsies reveal that many decedents exhibited latent or highly concealed indicators of psychological pain, successfully masking their distress until a terminal crisis. This Viewpoint argues that the field's reliance on explicit symptom reporting via instruments such as the Patient Health Questionnaire-9(PHQ-9)creates a systemic diagnostic blind spot for individuals with "masked" or high-functioning depression, who actively conceal their distress due to stigma, perfectionism, or maladaptive defense mechanisms. To address this, we propose a novel conceptual model: the "Digital Leakage Effect" (DLE). Grounded in cognitive load theory and the neurobiology of allostatic load, the DLE posits that the immense executive functioning required for daytime psychosocial masking inevitably leads to prefrontal exhaustion by evening. Consequently, suppressed psychological distress subconsciously "leaks" through involuntary nocturnal physiological and human-computer interaction (HCI) behaviors. We outline a multimodal digital phenotyping framework mapping these passive metrics directly toDiagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5)criteria: (1) chronobiological disruption (sleep fragmentation, circadian misalignment); (2) psychomotor anomalies (diminished keystroke velocity, increased scrolling entropy); and (3) semantic-temporal divergence (discordance between large language model (LLM)-derived text sentiment and posting timestamps). Finally, we argue that the ethical deployment of these technologies—anchored in federated learning, dynamic informed consent, and culturally calibrated algorithms—is not merely preferable but mandatory.In clinical practice, integrating the DLE framework will bridge the assessment gap between traditional visits, augment initial clinical triage, and enable the deployment of Just-in-Time Adaptive Interventions (JITAIs) during nocturnal vulnerability windows. By shifting the diagnostic focus from what patients consciously report during the day to what they subconsciously reveal through their digital footprint at night, clinicians can finally capture the elusive phenotype of concealed depression.