<p>Comparison of AUUC Scores.</p>

This study analyzes the causal effect of sleep duration on mental health among young adults using a meta-learner-based causal inference framework. Specifically, we applied a T-Learner model with Random Forest and XGBoost as the base learner to data from 1,405 individuals aged 19–34, drawn from the 2022–2023 Korea National Health and Nutrition Examination Survey. The result indicates that adequate sleep increases the probability of maintaining normal mental health. Subgroup analysis comparing individuals with adequate sleep and normal mental health to those with insufficient sleep and poor mental health also reveals a statistically significant causal effect of sleep on mental health improvement. In addition, AST (SGOT) levels and blood creatinine concentration are identified as key confounding factors. Findings suggest that sufficient sleep could enhance mental health among young adults, and policy implications for youth mental health are derived from the perspective of sleep duration. By providing empirically identified causal evidence based on nationally representative data, this study contributes to the growing literature on sleep and mental health and highlights sleep duration as a modifiable target for evidence-based mental health interventions in young adults.本研究采用基于元学习器（meta-learner）的因果推断框架，分析了青年群体睡眠时间对心理健康的因果效应。具体而言，本研究以2022-2023年韩国国民健康与营养检查调查（Korea National Health and Nutrition Examination Survey）中1405名年龄介于19至34岁的个体数据为研究样本，采用以随机森林（Random Forest）和XGBoost为基学习器的T-Learner模型开展分析。研究结果显示，充足睡眠可提升维持正常心理健康状态的概率。针对充足睡眠且心理健康正常者与睡眠不足且心理健康欠佳者的亚组对比分析，同样证实睡眠对心理健康改善存在具有统计学显著性的因果效应。此外，天门冬氨酸氨基转移酶（AST，SGOT）水平与血肌酐浓度被识别为关键混杂因素。本研究结果表明，充足睡眠可改善青年群体的心理健康状况，并从睡眠时间维度为青年心理健康相关政策制定提供了政策启示。本研究基于具有全国代表性的数据集提供了经实证识别的因果证据，为日益丰富的睡眠与心理健康相关研究文献作出了贡献，并强调睡眠时间可作为青年群体循证心理健康干预的可调控靶点。