KMAN Publication Inc. (KMANPUB) Journal of Adolescent and Youth Psychological Studies (JAYPS) 2981-2526 7 Serial Number 48 2026 08 10 Deep Neural Network Prediction of Depression Risk Among Young Adults Based on Psychological, Sleep, Lifestyle, and Social Variables Deep Neural Network Prediction of Depression Risk Among Young Adults Based on Psychological, Sleep, Lifestyle, and Social Variables 1 17 EN Journal Article 2026 04 22 <table> <tbody> <tr> <td> <p><strong>Objective:</strong> This study aimed to develop and evaluate a deep neural network model for predicting elevated depression risk among young adults in Canada using psychological, sleep-related, lifestyle, and social variables.</p> <p><strong>Methods and Materials:</strong> This cross-sectional predictive study included 1,042 Canadian young adults aged 18–35 years. Depression risk was assessed using the Patient Health Questionnaire-9, with scores ≥10 indicating elevated risk. Predictors included perceived stress, anxiety symptoms, self-esteem, loneliness, resilience, sleep quality, sleep duration, sleep-onset latency, physical activity, sedentary behavior, screen time, perceived social support, financial strain, and meaningful social interactions. Data were divided into training (n = 729), validation (n = 156), and independent test (n = 157) sets using stratified sampling. A feedforward deep neural network with three hidden layers was trained using the Adam optimizer and binary cross-entropy loss. Performance was evaluated using accuracy, sensitivity, specificity, precision, F1 score, ROC-AUC, PR-AUC, and Brier score. Logistic regression, random forest, and gradient boosting were used as benchmark models, and SHAP values were applied to interpret predictor importance.</p> <p><strong>Findings:</strong> The deep neural network achieved an accuracy of 0.87, sensitivity of 0.84, specificity of 0.89, precision of 0.82, F1 score of 0.83, ROC-AUC of 0.93, PR-AUC of 0.89, and Brier score of 0.101 in the independent test set. It outperformed logistic regression, random forest, and gradient boosting. The strongest predictors were generalized anxiety symptoms, perceived stress, low social support, low self-esteem, loneliness, poor sleep quality, financial strain, low resilience, shorter sleep duration, and greater sedentary time. Nonlinear effects and cross-domain interactions were also identified.</p> </td> </tr> </tbody> </table> <p><strong>Conclusion:</strong> Depression risk among young adults can be predicted with high accuracy by integrating psychological, sleep, lifestyle, and social variables within an interpretable deep neural network framework.</p> https://journals.kmanpub.com/index.php/jayps/article/download/6215/11802