OPEN ACCESS SPOKEN LANGUAGE DATA FOR UNDERSTANDING DEMENTIA

摘要 我们建议扩展多媒体交互的共享DementiaBank数据库，用于研究 老年认知交流障碍和阿尔茨海默氏症及相关痴呆症。我们继续 这项工作将依赖于在数据收集、转录、分析、共享和网络交付方面取得的进展， 整个TalkBank项目的背景。我们将部署和扩展TalkBank不断增长的资产， AphasiaBank和DementiaBank回答有关诊断、分类和 痴呆症中语言的纵向追踪。这项工作的总体目标是支持 认知语言结果的措施和治疗，将改善评估和有效性的早期 诊断和识别那些有痴呆症风险的人。为了实现这一目标，我们必须巩固数据分析， 训练方法（即，协作评论、自动语音识别和机器学习 基准）。我们还将注重扩大传播方法，以确保痴呆症银行 工具和资源正在满足研究人员和临床医生的需求。

Abstract We propose the expansion of the shared DementiaBank database of multimedia interactions for the study of cognitive-communication impairments in aging and Alzheimer’s disease and related dementias. Our continued work will rely on advances in data collection, transcription, analysis, sharing, and web delivery achieved in the context of the overall TalkBank Project. We will deploy and extend the growing assets of TalkBank, AphasiaBank, and DementiaBank to answer critical questions regarding the diagnosis, classification, and longitudinal tracking of language in dementia. The overall goal of this work is to support the development of cognitive-linguistic outcome measures and treatments that will improve the evaluation and effectiveness early diagnosis and identifying those at-risk for dementia. To reach that goal, we must solidify data analysis and training methods (i.e., collaborative commentary, automatic speech recognition, and machine learning benchmarks). We will also focus on the expansion of dissemination methods to ensure the DementiaBank tools and resources are meeting the needs of researchers and clinicians.