Validity and Utility of Consumer-based Wearable Fitness Trackers to Monitor Free-Living Physical Activity Energy Expenditure and Sleep in Children 5-12 Years Old

准确测量儿童（5- 12岁）的自由生活体力活动（PA）、能量消耗（EE）和睡眠 复杂，没有一种方法是没有局限性的。PA、EE和睡眠的验证研究表明， 组合HR和加速度测量数据（例如，步骤、计数、原始信号）提供PA、EE 睡觉不幸的是，在常规监测时间范围内同时收集HR和加速度测量（例如， 7天），因为历史上测量HR依赖于不舒服的胸带遥测。 可穿戴技术的进步已经通过纳入HR的非侵入性评估消除了这个问题 经由广泛可用的消费者可穿戴设备中的光电体积描记术（例如，FitBits、Garmin），其中还包括 加速度测量法研究表明，来自消费者可穿戴设备的人力资源估计与通过 心电图或胸带遥测。然而，消费者可穿戴设备的验证研究几乎完全集中在 专有活动输出（例如，步骤），并且大多在健康成人、老年人或临床 群体（例如，有神经肌肉或步态异常的人）。消费者可穿戴设备有望收集 5至12岁儿童的PA、EE和睡眠数据，但尚未进行研究以确定其有效性， 在这一人群中的实用性/可行性。拟议项目的目标是进行一系列研究，包括 实验室和现场协议，以评估消费者可穿戴设备的可靠性，有效性和实用性/可行性 用于测量儿童在自由生活条件下的PA、EE和睡眠。我们将评估设备的不同功能 (e.g., PA，HR）在实验室和现实世界条件下。此外，我们将评估消费者的实用性/可行性 可穿戴设备，实现多日佩戴合规性。我们将实现以下目标。目标1。开发和验证开放式- 使用来自消费者的活动和HR数据估计PAEE和体力活动时间的源方程 与消费者可穿戴设备专有处理算法的PAEE输出相比， PAEE的标准测量（即，间接量热法）。目标2.开发并验证开源方程， 总睡眠时间，睡眠效率，以及消费者可穿戴设备使用其活动和HR数据与 来自消费者可穿戴设备的专有处理算法的睡眠输出和睡眠的标准测量（即， PSG和体动计）。目标3。根据创建的算法评估PAEE和睡眠方程估计值的有效性 在目标1和目标2中，在自由生活的条件下。这个项目意义重大，因为它将是 是第一批确定消费者可穿戴设备对PAEE和儿童睡眠监测的有效性的公司之一。这个项目 它是创新的，因为它将使用先进的统计建模技术，包括机器学习， 儿童可穿戴消费品的有效性和实用性/可行性。我们的愿景是利用收集到的生物特征数据 在消费者可穿戴设备中进行测试，以评估PAEE和儿童睡眠。这将使从业人员和 研究人员希望更准确地测量儿童24小时的运动行为。

Accurate measurement of free-living physical activity (PA), energy expenditure (EE), and sleep of children (5-12yrs) is complex, with no single method free of limitations. Validation studies of PA, EE, and sleep have demonstrated that combining HR and accelerometry data (e.g., steps, counts, raw signal) provides the most accurate estimate of PA, EE, and sleep. Unfortunately, the simultaneous collection of HR and accelerometry over routine monitoring timeframes (e.g., 7 days) has been limited because historically measuring HR has relied on uncomfortable chest strap telemetry. Advancements in wearable technology have eliminated this issue by incorporating the noninvasive assessment of HR via photoplethysmography in widely-available consumer wearable devices (e.g., FitBits, Garmin) that also include accelerometry. Studies have shown that HR estimates from consumer wearables are comparable to those collected via ECG or chest strap telemetry. However, validation studies of consumer wearables have focused almost exclusively on proprietary activity output (e.g., steps) and have mostly been conducted on healthy adults, older adults, or clinical populations (e.g., people with neuromuscular or gait abnormalities). Consumer wearables hold promise for collecting PA, EE, and sleep data with children 5 to 12yrs, yet no studies have been conducted to establish their validity and utility/feasibility in this population. The objectives of the proposed project are to conduct a series of studies that include both laboratory and field-based protocols to evaluate the reliability, validity and utility/feasibility of consumer wearables for measuring children’s PA, EE, and sleep in free-living conditions. We will evaluate the different features of the devices (e.g., PA, HR) in the lab and in real-world conditions. In addition, we will evaluate the utility/feasibility of consumer wearables for multi-day wear compliance. We will accomplish the following aims. Aim 1. Develop and validate open- source equations to estimate PAEE and time spent physically active using the activity and HR data from consumer wearables compared to the PAEE output from the consumer wearables’ proprietary processing algorithms and a criterion measure of PAEE (i.e., indirect calorimetry). Aim 2. Develop and validate open-source equations to estimate total sleep time, sleep efficiency, and timing from consumer wearables using their activity and HR data compared to the sleep output from the consumer wearables’ proprietary processing algorithms and a criterion measure of sleep (i.e., PSG and actigraphy). Aim 3. Evaluate the validity of the PAEE and sleep equation estimates from the algorithms created in Aim 1 & 2 against a criterion (i.e., Actiheart) under free-living conditions. This project is significant because it will be among the first to establish the validity of consumer wearables for PAEE and sleep monitoring of children. This project is innovative as it will use advanced statistical modeling techniques, including machine learning, to systematically test the validity and utility/feasibility of consumer wearables for children. Our vision is to leverage the biometric data collected across consumer wearables to produce estimates of PAEE and sleep in children. This will allow practitioners and researchers alike to more accurately measure 24-hour movement behaviors in children.