Advancing Physical Activity Measurement Using Pattern Recognition Techniques

使用模式识别技术推进身体活动测量

• 批准号: 7814900
• 负责人: Patty S. Freedson
• 金额: $ 49.44万
• 依托单位: UNIVERSITY OF MASSACHUSETTS AMHERST
• 依托单位国家: 美国
• 财政年份: 2009
• 资助国家: 美国
• 起止时间: 2009-09-30 至 2011-07-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/7814900
• 关键词: Address; Advisory Committees; Advocate; Affect; American; Area; Behavior; Behavioral; Biological Markers; Biological Neural Networks; Blood Pressure; Cardiovascular Diseases; Cardiovascular system; Cholesterol; Clinical; Data; Detection; Development; Dose; Energy Metabolism; Ensure; Environment; Environmental Exposure; Exercise; Feedback; Funding; Genetic; Guidelines; Health; Heart; Hematological Disease; Indirect Calorimetry; Individual; Insulin; Intervention; Journals; Knowledge; Laboratories; Life; Light; Literature; Lung; Machine Learning; Measurement; Measures; Mediating; Metabolic; Metabolic syndrome; Methodology; Methods; Moderate Exercise; Modification; Monitor; National Heart, Lung, and Blood Institute; Non-Insulin-Dependent Diabetes Mellitus; Obesity; Onset of illness; Outcome; Paper; Participant; Pathway Analysis; Pattern; Pattern Recognition; Persons; Physical activity; Physiology; Population Surveillance; Positioning Attribute; Prevention; Process; Public Health; Publishing; Reporting; Research; Research Personnel; Research Proposals; Risk Factors; Risk Reduction; Role; Structure; Techniques; Technology; Time; Training; Training Programs; Triglycerides; United States Dept. of Health and Human Services; base; blood lipid; cardiovascular disorder risk; computerized data processing; design; fitness; improved; innovation; innovative technologies; mortality; novel; programs; response; sedentary; tool

DESCRIPTION (provided by applicant): This application addresses the broad Challenge Area (01) Behavior, Behavioral Change, and Prevention and specific Challenge Area 01-HL-101: Develop innovative technologies and measurements to assess and provide real-time feedback on behavioral and environmental exposures for disease onset and progression for heart, lung, and blood disease. In October, 2008 the US Department of Health and Human Services issued the first-ever federally mandated Physical Activity Guidelines for Americans. The Guidelines reflect the view of the Physical Activity Guidelines Advisory Committee (PAGAC) and are based on an extensive review of the scientific literature on physical activity (PA) and health. In their report, the PAGAC points out the limited knowledge of the dose-response relationship between PA and health, and identifies poor measures of PA exposure as a major contributing factor to this gap in knowledge. Our application directly addresses this issue by applying innovative technologies to measure PA dose in a free- living environment. We will use these technologies to examine if habitual PA performed outside of purposeful exercise influences biomarkers of cardiovascular health. Although insufficient PA clearly correlates with an increased risk for cardiovascular disease (CVD), research evidence is equivocal regarding the effects of training on CVD risk factors (e.g. insulin action, triglycerides, blood pressure, and cholesterol). Research suggests increases in sedentary behavior may negate the benefits of training however this idea has not been explored experimentally. Our application will consider habitual free-living PA as a possible mechanism mediating the relationship between training and risk factors for cardiovascular disease. In order to elucidate the relationship between PA and biomarkers of cardiovascular disease risk, it is critical that valid, objective measures are used to quantify PA. We propose to use novel analytic techniques known as artificial neural networks (ANN) to process accelerometer-based measurements of PA. The first part of this project (Aim 1) will examine the ANN's sensitivity to change in PA dose by applying the ANN technique to distinguish three distinct patterns of habitual PA - Sedentary, Moderately Active, and Very Active. These three conditions represent common activity patterns that impact health. Accurately assessing changes to habitual PA levels that are relevant to public health will advance the field by further establishing a technique for application in population surveillance research and detection of changes in PA consequent to an intervention. The second part of this project (Aim 2) will apply the ANN methodology to examine the effect of free-living activity and inactivity levels, performed outside of training, on insulin action, blood pressure, triglycerides, cholesterol, and cardiorespiratory fitness following a 12-week exercise training trial in previously sedentary individuals with an elevated risk for CVD. Results from this study have the potential to impact how clinical exercise trials are conducted (e.g. require objective monitoring of PA outside of an exercise training trial) and how exercise is prescribed (e.g. reducing sedentary time AND maintaining sufficient PA). The Physical Activity Guidelines Advisory Committee advocates improved measures of physical activity exposure in order to elucidate the relationship between physical activity dose and health. To address this challenge we will apply and validate innovative accelerometer-based technologies for measuring physical activity to assess its sensitivity to detecting changes in dose of physical activity and to monitor activity outside of a training program designed to improve cardiorespiratory fitness and biomarkers of cardiovascular disease risk. Through improved measures of physical activity this project will promote a better understanding of how the dose of physical activity affects selected health outcomes.

描述（由申请人提供）：该申请涉及广泛的挑战领域（01）行为、行为改变和预防，以及特定的挑战领域01- hl -101：开发创新技术和测量方法，以评估和提供心脏、肺和血液疾病发病和进展的行为和环境暴露的实时反馈。2008年10月，美国卫生与公众服务部发布了第一份联邦授权的美国人体育活动指南。该指南反映了身体活动指南咨询委员会（PAGAC）的观点，并以对身体活动与健康的科学文献的广泛审查为基础。在他们的报告中，PAGAC指出，人们对PA与健康之间的剂量-反应关系的认识有限，并确定对PA暴露的不良测量是造成这种知识差距的主要因素。我们的应用程序直接解决了这一问题，通过应用创新的技术来测量PA剂量在一个自由的生活环境。我们将使用这些技术来检查在有目的的运动之外进行的习惯性PA是否会影响心血管健康的生物标志物。尽管PA不足明显与心血管疾病（CVD）风险增加相关，但关于训练对CVD危险因素（如胰岛素作用、甘油三酯、血压和胆固醇）的影响，研究证据尚不明确。研究表明，久坐行为的增加可能会抵消训练的好处，但这一观点尚未得到实验的证实。我们的应用将考虑习惯性自由生活的PA作为一种可能的机制，调解训练与心血管疾病危险因素之间的关系。为了阐明PA与心血管疾病风险的生物标志物之间的关系，使用有效、客观的测量方法来量化PA是至关重要的。我们建议使用称为人工神经网络（ANN）的新型分析技术来处理基于加速度计的PA测量。该项目的第一部分（目标1）将通过应用人工神经网络技术来区分三种不同的习惯性PA模式——久坐、适度运动和非常运动，来检查人工神经网络对PA剂量变化的敏感性。这三种情况代表了影响健康的常见活动模式。准确评估与公共卫生相关的习惯性PA水平的变化，将进一步建立一种技术，用于人口监测研究和检测干预后PA的变化，从而推动这一领域的发展。该项目的第二部分（目标2）将应用人工神经网络方法，对先前久坐且心血管疾病风险升高的个体进行为期12周的运动训练试验后，在训练之外进行的自由活动和不活动水平对胰岛素作用、血压、甘油三酯、胆固醇和心肺健康的影响。这项研究的结果有可能影响临床运动试验的进行方式（例如，在运动训练试验之外需要客观监测PA）以及如何规定运动（例如，减少久坐时间并保持足够的PA）。体力活动指南咨询委员会提倡改进体力活动暴露的措施，以阐明体力活动剂量与健康之间的关系。为了应对这一挑战，我们将应用并验证基于加速计的创新技术来测量身体活动，以评估其检测身体活动剂量变化的敏感性，并监测旨在改善心肺健康和心血管疾病风险生物标志物的训练计划之外的活动。通过改进身体活动的措施，该项目将促进更好地了解身体活动的剂量如何影响选定的健康结果。