Characterizing Episodes of Lucidity in Dementia Using Observational and Applied Computational Linguistics Approaches

使用观察和应用计算语言学方法表征痴呆症的清醒发作

• 批准号: 10677987
• 负责人: Andrea L Gilmore-Bykovskyi
• 金额: $ 72.75万
• 依托单位: UNIVERSITY OF WISCONSIN-MADISON
• 依托单位国家: 美国
• 财政年份: 2020
• 资助国家: 美国
• 起止时间: 2020-09-30 至 2025-05-31
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10677987
• 关键词: Address; Affect; Alzheimer' s Disease; Alzheimer' s disease related dementia; Attitude; Caring; Clinical; Code; Cognitive; Communication; Computational Linguistics; Data; Data Collection; Data Sources; Dementia; Development; Disease Progression; Enrollment; Epidemiology; Ethics; Event; Eye; Family Caregiver; Feasibility Studies; Focus Groups; Foundations; Frequencies; Future; Goals; Group Interviews; Individual; Inpatients; Investigation; Language; Legal; Linguistics; Measurement; Measures; Methodology; Methods; Mission; Nonverbal Communication; Output; Persons; Phase; Phenotype; Population; Privacy; Procedures; Reporting; Research; Research Priority; Secure; Source; Speech; Standardization; Structure; Target Populations; United States; Validation; Variant; Work; advanced disease; base; behavioral response; cognitive function; communication behavior; complex data; data infrastructure; data repository; data sharing; end of life; end of life care; gaze; hospice environment; interest; prospective; prototype; recruit; repository; sharing platform; usability

PROJECT SUMMARY/ABSTRACT Individuals with Alzheimer’s disease and related dementias (ADRD) demonstrate heterogeneous cognitive, functional, and behavioral responses to disease progression, yet underlying sources of these variations remain poorly understood. Fluctuations in communication abilities in particular have important implications for care and attitudes toward individuals with ADRD. Of particular interest is the occurrence of episodes of lucidity (EL) in advanced disease stages, which are characterized by the manifestation of spontaneous, meaningful, and relevant communication abilities that were previously believed to be irretrievable. Reports suggest EL are transient and most likely to occur near end of life, yet empiric evidence documenting these events is extremely limited. Efforts to better understand, and ultimately define EL, are hindered by underdeveloped methodological approaches for capturing and characterizing these events. While underutilized in ADRD research, audiovisual observation is well-suited to addressing these gaps, as these data provide an objective data source and enable a robust assessment of verbal and nonverbal communication, the primary means through which EL are observed. Our long-term objective is to clarify conceptual, operational and epidemiologic understandings of EL in ADRD. Our short-term objective is to establish the necessary foundational data and infrastructure to accelerate systematic investigation of EL. To advance these goals, we will develop feasible and acceptable procedures to enable capture of longitudinal audiovisual data of targeted populations and timeframes to maximize opportunities for detecting EL; these rich data sources will then be analyzed via computational linguistic and sequential analysis (timed-event) methods, to assemble a robust, fully characterized understanding of the linguistic, non- linguistic, and non-verbal communication features of EL along with the specific temporal qualities of these events. In the R21 Phase, we will demonstrate the feasibility of collecting and sharing longitudinal audiovisual data among PLWD near end of life through a mixed-methods feasibility study to determine key ethical and practical considerations (R21 Aim 1); and evaluate usability, reliability and capture rate for coding of verbal/nonverbal indicators of EL from collected audiovisual data using computational linguistic and timed-event methods (R21 Aim 2). The R33 Phase will harness procedures optimized in the R21 phase to expand data collection and develop more robust estimates of EL frequency. Specifically, we will expand longitudinal audiovisual observations of persons with ADRD near the end of life to obtain initial estimates of EL frequency, describe event attributes, and inform the development of definitions for EL (R33 Aim1); and establish a repository of longitudinal audiovisual observational data and a secure data sharing platform (R33 Aim 2). Impact: Findings from this research will set the stage for extension and validation of urgently needed objective measurement for EL in ADRD and provide a critically needed foundation for future systematic investigation of lucidity in ADRD. The objectives of this proposal are directly responsive to national ADRD research priorities and the NIA mission.

项目摘要/摘要 阿尔茨海默病和相关痴呆(ADRD)患者表现出不同的认知能力， 对疾病进展的功能和行为反应，但这些变化的潜在来源仍然存在 人们对此知之甚少。尤其是沟通能力的波动对护理和 对ADRD患者的态度。特别令人感兴趣的是清醒发作(EL)在 晚期疾病，其特征是表现为自发的、有意义的和 以前被认为是无法挽回的相关沟通能力。有报道称，El正在 这些事件是短暂的，最有可能发生在生命的末期，但记录这些事件的经验证据是极其丰富的 有限的。更好地理解并最终定义EL的努力受到方法学不发达的阻碍 捕捉和描述这些事件的方法。虽然在ADRD研究中未得到充分利用，但视听 观察法非常适合于弥补这些差距，因为这些数据提供了客观数据来源，并使 对语言和非语言交流的强有力的评估，这是El通过的主要手段 观察到的。我们的长期目标是澄清对流行性感冒的概念、操作和流行病学的理解 在ADRD。我们的短期目标是建立必要的基础数据和基础设施，以加快 EL.系统调查。为了推进这些目标，我们将制定可行和可接受的程序，以 能够捕获目标人群和时间范围的纵向视听数据，以最大限度地增加机会 用于检测EL；然后将通过计算语言和顺序分析这些丰富的数据源 分析(计时事件)方法，以集合对语言、非 EL的语言和非语言交际特征以及这些事件的特定时间性质。 在R21阶段，我们将论证收集和共享纵向视听数据的可行性 通过一项混合方法的可行性研究，在即将结束的PLWD中确定关键的伦理和实用 考虑事项(R21目标1)；评估语言/非语言编码的可用性、可靠性和捕获率 使用计算语言和定时事件方法(R21)从收集的视听数据中获得EL的指示器 目标2)。R33阶段将利用在R21阶段优化的程序，以扩大数据收集和 开发更可靠的EL频率估计。具体来说，我们将扩大纵向视听 对ADRD患者临终时的观察，以获得EL频率的初步估计，描述事件 属性，并为制定EL(R33 Aim1)的定义提供信息；建立一个纵向信息库 视听观测数据和安全数据共享平台(R33目标2)。影响：由此得出的结论 研究将为扩展和验证EL迫切需要的客观测量奠定基础 并为今后系统研究ADRD的透明度提供了迫切需要的基础。这个 该提案的目标直接响应国家ADRD的研究优先事项和NIA的使命。