STUDY OBJECTIVES Machine learning (ML) may provide insights into the underlying sleep stages of accelerometer-assessed sleep duration. We examined associations between ML sleep patterns and behaviour problems among preschool children. METHODS Children from the CHILD Cohort Edmonton site with actigraphy and behaviour data at three-years (n=330) and five-years (n=304) were included. Parent-reported behaviour problems were assessed by the Child Behavior Checklist. The Hidden Markov Model (HMM) classification method was used for ML analysis of the accelerometer sleep period. The average time each participant spent in each HMM-derived sleep state was expressed in hours/day. We analyzed associations between sleep and behaviour problems stratified by children with and without sleep-disordered breathing (SDB). RESULTS Four hidden sleep states were identified at three years and six hidden sleep states at five years using HMM. The first sleep state identified for both ages (HMM-0) had zero counts (no movement). The remaining hidden states were merged together (HMM-mov). Children spent an average of 8.2±1.2 hours/day in HMM-0 and 2.6±0.8 hours/day in HMM-mov at three years. At age five, children spent an average of 8.2±0.9 hours/day in HMM-0 and 1.9±0.7 hours/day in HMM-mov. Among SDB children, each hour in HMM-0 was associated with 0.79-point reduced externalizing behaviour problems (95%CI -1.4, -0.12; p<0.05), and a 1.27-point lower internalizing behaviour problems (95%; -2.02, -0.53; p<0.01). CONCLUSIONS ML-sleep states were not associated with behaviour problems in general-population of children. Children with SDB who had greater sleep duration without movement had lower behavioural problems. The ML-sleep states require validation with polysomnography.

中文翻译：

---

机器学习推导的睡眠参数与 3 岁和 5 岁儿童行为问题的关系；儿童队列研究的结果

研究目标 机器学习 (ML) 可以深入了解加速度计评估的睡眠持续时间的潜在睡眠阶段。我们研究了学龄前儿童 ML 睡眠模式与行为问题之间的关联。方法 来自 CHILD Cohort Edmonton 站点的儿童包括三年 (n=330) 和五年 (n=304) 的动作记录和行为数据。家长报告的行为问题由儿童行为清单评估。隐马尔可夫模型 (HMM) 分类方法用于加速度计睡眠周期的 ML 分析。每个参与者在每个 HMM 衍生的睡眠状态中花费的平均时间以小时/天表示。我们分析了睡眠呼吸障碍 (SDB) 和非睡眠呼吸障碍 (SDB) 儿童分层的睡眠与行为问题之间的关联。结果 使用 HMM 在三年时确定了四种隐藏的睡眠状态，在五年时确定了六种隐藏的睡眠状态。为两个年龄段（HMM-0）确定的第一个睡眠状态的计数为零（没有运动）。剩余的隐藏状态合并在一起（HMM-mov）。三岁时，儿童在 HMM-0 中平均每天花费 8.2±1.2 小时，在 HMM-mov 中花费 2.6±0.8 小时/天。五岁时，儿童在 HMM-0 中平均每天花费 8.2±0.9 小时，在 HMM-mov 中花费 1.9±0.7 小时/天。在 SDB 儿童中，HMM-0 每小时与 0.79 分的外化行为问题（95%CI -1.4，-0.12；p<0.05）和内化行为问题降低 1.27 分（95%；-2.02）相关，-0.53；p<0.01)。结论 ML 睡眠状态与一般儿童的行为问题无关。睡眠时间较长且不运动的 SDB 儿童的行为问题较少。ML 睡眠状态需要使用多导睡眠图进行验证。