Multi-scale modeling of sleep behaviors in social networks

社交网络中睡眠行为的多尺度建模

• 批准号: 8453066
• 负责人: Charles A Czeisler
• 金额: $ 49.43万
• 依托单位: BRIGHAM AND WOMEN'S HOSPITAL
• 依托单位国家: 美国
• 财政年份: 2013
• 资助国家: 美国
• 起止时间: 2013-07-18 至 2018-03-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/8453066
• 关键词: Adolescent; Affect; Behavior; Behavioral; Biological; Biological Process; Caffeine; Car Phone; Cardiovascular Diseases; Circadian Rhythms; Cognitive; Cues; Data; Development; Devices; Diabetes Mellitus; Drug usage; Electronics; Fatigue; Habits; Health; Hormones; Human; Individual; Individual Differences; Inpatients; Internet; Learning; Length; Life; Light; Measures; Melatonin; Metabolic; Methods; Modeling; Moods; Nature; Network-based; Obesity; Outcome; Outpatients; Participant; Pattern; Pattern Recognition; Performance; Pharmaceutical Preparations; Physiological; Physiology; Population; Population Heterogeneity; Property; Recovery; Risk-Taking; Schedule; Simulate; Skin Temperature; Sleep; Sleep Deprivation; Social Interaction; Social Network; Statistical Models; Stimulus; Stress; Students; System; Techniques; Teenagers; Testing; Time; Weight Gain; Wit; Work; awake; base; cognitive function; cohort; computerized data processing; immune function; improved; innovation; insight; mathematical model; multi-scale modeling; network models; public health relevance; research study; response; sensor; sleep onset; social; tool; undergraduate student; young adult

DESCRIPTION (provided by applicant): Sleep is critical to a wide range of biological functions. Inadequate sleep results in impaired cognitive performance and mood, and adverse health outcomes including obesity, diabetes, and cardiovascular disease. Recent evidence suggests that sleep behaviors can spread between individuals connected by a social network and that these behaviors can even influence drug use in teenagers. While models exist separately for quantifying connectivity within social networks and for modeling sleep, there are currently no combined models for predicting and studying the emergent dynamics of sleep behaviors within social networks. We therefore propose to develop multi-scale physiologically-based models of the effects of social interactions on sleep behaviors. We have assembled a trans-disciplinary team of individuals who have: (i) developed mathematical methods for quantifying social network interactions; (ii) developed a physiologically based model of sleep and circadian physiology, including the effects of wake-promoting stimuli and drugs; (iii) studied healthy and pathological sleep behaviors under inpatient and outpatient conditions, including in undergraduate students; (iv) developed techniques for collecting multiple physiological and behavioral variables; and (v) studied pattern recognition and signal processing techniques for analyzing multimodal data. We will develop statistical and mathematical models from experimental data collected from 8 groups of closely-connected MIT undergraduates using mobile phones and wearable sensors to measure sleep patterns and duration, light exposure, subjective measures of sleepiness and mood, and social interactions including texting, calls, internet use, and spatial proximity to other participants. We will determine how social interactions, sleep duration and timing, light exposure, sleepiness and mood interact. These social interaction effects will then be added to our physiological sleep and circadian model, which will also be extended from the individual to the population level, while the physiological model results will inform the social network model work. Once developed, the mathematical model will be used to explore how emergent dynamics depend on network properties. Specifically, we will simulate the student network, including the observed rates and effects of social interactions. We will then test the effects of modifying the network properties, including the strengths of interactions and the degree of population heterogeneity (model parameter variability). We anticipate that the mathematical model developed in this project will provide a new means of predicting the dynamics of sleep behaviors within social networks. Due to its multi-scale nature, the model will relate observations at the network level to interactions between individuals. This will allow us to simulate candidate strategies for intervening in populations wit unhealthy sleep behaviors. Given the alarming increase in insufficient sleep in the U.S., and the rapidly escalating use of social media, establishing models that can be used to improve sleep behaviors could potentially improve multiple health outcomes.

描述(由申请者提供)：睡眠对一系列生物功能至关重要。睡眠不足会导致认知能力和情绪受损，并导致肥胖、糖尿病和心血管疾病等不良健康后果。最近的证据表明，睡眠行为可以在通过社交网络连接的个人之间传播，这些行为甚至会影响青少年的药物使用。虽然有单独的模型来量化社交网络中的连通性和对睡眠进行建模，但目前还没有组合的模型来预测和研究社交网络中睡眠行为的紧急动态。 因此，我们建议开发多尺度的基于生理的社会互动对睡眠行为影响的模型。我们已经组建了一个跨学科的个人团队，他们已经：(I)开发了量化社会网络互动的数学方法；(Ii)开发了一个基于生理的睡眠和昼夜生理学模型，包括促醒刺激和药物的影响；(Iii)研究了住院和门诊条件下的健康和病理性睡眠行为，包括本科生；(Iv)开发了收集多个生理和行为变量的技术；以及(V)研究了模式识别和信号处理技术，用于分析多模式数据。 我们将从8组密切联系的麻省理工学院本科生身上收集实验数据，使用手机和可穿戴传感器来测量睡眠模式和持续时间、光线暴露、嗜睡和情绪的主观测量，以及社交互动，包括短信、通话、互联网使用和与其他参与者的空间接近程度，我们将建立统计和数学模型。我们将确定社会互动、睡眠持续时间和时间、光线暴露、困倦和情绪如何相互作用。这些社会互动效应将被添加到我们的生理睡眠和昼夜节律模型中，该模型也将从个体层面扩展到群体层面，而生理模型的结果将通知社会网络模型的工作。 一旦开发完成，该数学模型将用于探索紧急动态如何依赖于网络特性。具体地说，我们将模拟学生网络，包括观察到的社交互动的比率和影响。然后，我们将测试修改网络属性的效果，包括交互的强度和种群异质性的程度(模型参数可变性)。 我们预计，该项目中开发的数学模型将提供一种新的方法来预测社交网络中睡眠行为的动态。由于其多尺度的性质，该模型将把网络层面的观测与个体之间的交互联系起来。这将使我们能够模拟干预有不健康睡眠行为的人群的候选策略。鉴于美国睡眠不足的增加令人担忧，社交媒体的使用迅速升级，建立可用于改善睡眠行为的模型可能会改善多种健康结果。