SCH: INT: Collaborative Research: Passive sensing of social isolation: A digital phenotying approach

社会孤立--既包括客观上的孤独现象，也包括主观上的孤独体验 (感知到的孤立)-是全球的一个主要问题。我们在提议的项目中的目标是利用统计学 利用基于智能手机的连续、不显眼和实时测量的能力的方法 对社会孤立的评估。我们汇集了一支在社会行为方面具有专业知识的临床科学家团队 动力学、处于数字信号处理研究前沿的工程师/计算机科学家和生物统计学家 具备被动传感技术方面的专业知识，以提供执行 研究的目的。使用数字表型方法(即，个体水平的时刻量化 使用来自个人智能手机的数据现场识别人类表型)，我们将开发和测试算法，将 主动(生态即时评估)和被动(移动、位置、对话)指标均可改进 社会隔离的特征和预测。然后我们将对承诺的一个 社会隔离过渡状态的动态网络分析，并将该方法应用于临床 以社会孤立为特征的样本。 相关性(请参阅说明)： 该项目的发现将对全球健康产生深远的影响，因为我们·正在形成对社会孤立作为一种 早期死亡率和其他重大健康问题的主要贡献者强调了对可扩展、全面和 个性化评估和干预方法。开发新方法以改进对社会行为的推断 时间密集的智能手机数据将有利于数字健康研究领域的不断扩大。这一贡献超出了 项目的应用目标，随着我们将开发的方法的进步，可以应用于现有的大型语料库 数据和未来项目。最终，这项工作将有助于提供针对社会孤立的可持续干预措施。 在时间和日常环境中。

Social isolation-including both the objective phenomenon of 'aloneness' and the subjective experience of loneliness (perceived isolation)-is a major problem globally. Our goal in the proposed project is to capitalize on statistical methods for harnessing the power of smartphone-based measurement of continuous, unobtrusive, and real-time assessment of social isolation. We bring together a team of clinical scientists with expertise in social behavior dynamics, engineers/computer scientists at the forefront of research on digital signal processing, and biostatisticians with expert knowledge in passive sensing technology to provide robust methodological rigor needed to execute the study's aims. Using a digital phenotyping approach (i.e., the moment-by-moment quantification of the individual-level human phenotype in situ using data from personal smartphones), we will develop and test algorithms that incorporate both active (ecological momentary assessment) and passive (movement, location, conversation) metrics to improve characterization and prediction of social isolation. We will then conduct a preliminary evaluation of the promise of a dynamic network analysis of social isolation transition states, followed by application of this approach to a clinical sample characterized by social isolation. RELEVANCE (See instructions): Findings from this project will have far-reaching application to global health, as our·emerging understanding of social isolation as a key contributor to early mortality and other significant health problems highlights the need for a scalable, comprehensive, and personalized assessment and intervention approach. Developing new methods for improving inference of social behavior from temporally-dense smartphone data will benefit an expanding area of research in digital health. This contribution extends beyond the applied aims of the project, as the methodological advancements we will develop can be applied to a large corpus of existing data and future projects. Ultimately, this work will inform the delivery of sustainable interventions targeting social isolation in ieal- time and in daily contexts.