Collaborative Research: SCH: Assessment of Cognitive Decline using Multimodal Neuroimaging with Embedded Artificial Intelligence

摘要：阿尔茨海默病（AD）和阿尔茨海默病相关性痴呆（ADRD）在世界各地的老年人中非常普遍。认知能力下降的不利影响，从轻度认知障碍（MCI）到AD和ADRD，不仅给老年人、他们的照顾者和社会带来了高昂的经济成本，而且还带来了身体、精神和情感负担。MCI是AD的一个公认的危险因素。然而，传统的诊断程序和生物标志物在识别MCI中观察到的认知下降的脑机制改变方面的效用有限。虽然关于MCI、AD和ADRD的神经影像学相关性的文献相当多，但传统的脑成像方法昂贵、限制性强，并且通常单独进行。此外，使用多模态非侵入性神经成像方法的研究，可以利用自然的设置，以检测大脑为基础的签名MCI一直有限。开发提取这些特征的工具可以识别新的生物标志物，这些生物标志物可以指导开发精确的，个性化的评估和治疗与年龄相关的认知衰退和痴呆症。 在这个项目中，我们将开发一个工具链，用于使用多模态神经成像和机器学习（ML）方法评估MCI。我们提出了三个具体目标：（1）在移动的软件中开发一个与多模态功能近红外光谱和脑电图同步的对MCI敏感的综合认知测试组合（fNIRS-EEG）的神经成像系统，其可以同时提供电生理学、血液动力学和行为测量;（2）从fNIRS-EEG数据中提取、选择和验证时间、空间、频谱和复杂度域中的大量模态内和跨模态生物标志物以及行为生物标志物;（3）提出一种基于fNIRS-EEG和行为特征的多模态机器学习方法来检测MCI。 开发一个移动的应用程序，将fNIRS和EEG结合在一个平台上，可以在更便宜和限制性的测试环境中使用，以确定患有MCI的老年人的脑功能改变，这是非常具有创新性的。该项目的研究结果可以导致早期检测和监测有AD风险的老年人认知能力下降的转变。

Summary: Alzheimer’s Disease (AD) and Alzheimer’s Disease-Related Dementias (ADRD) are highly prevalent among older individuals throughout the world. The adverse impact of cognitive decline, ranging from mild cognitive impairment (MCI) to AD and ADRD, presents not only a prohibitive financial cost but also physical, mental, and emotional burden to older adults, their caregivers, and society. MCI is a well-established risk factor for AD. However, traditional diagnostic procedures and biomarkers have limited utility in identifying alterations in brain mechanisms that underlie the cognitive decline observed in MCI. While the literature concerning neuroimaging correlates of MCI, AD and ADRD is considerable, traditional brain imaging methods are expensive, restrictive, and typically conducted separately. Moreover, research using multimodal noninvasive neuroimaging methods that can be utilized in naturalistic settings to detect brain-based signatures of MCI has been limited. Developing tools to extract such signatures can lead to the identification of novel biomarkers that can guide the development of precise, and individualized assessment and treatment of age-related cognitive decline and dementia. In this project, we will develop a toolchain for the assessment of MCI using multimodal neuroimaging and machine learning (ML) methods. We propose three specific aims: (1) to develop a comprehensive cognitive testing battery sensitive to MCI in a mobile software synchronized with multimodal functional near infrared spectroscopy and electroencephalography (fNIRS-EEG) based neuroimaging system that can concurrently provide electrophysiological, hemodynamic and behavioral measures; (2) to extract, select, and validate the multitude of within and across modality biomarkers from fNIRS-EEG data in temporal, spatial, spectral, and complexity domains together with the behavioral ones; (3) to develop a comprehensive multimodal ML approach to detect MCI based on fNIRS-EEG and behavioral features. Developing a mobile application that combines fNIRS and EEG on one platform that could be used in less expensive and restrictive testing environments to determine functional brain alterations in older adults with MCI is very innovative. The findings of this project can lead to a transformation in early detection and monitoring of cognitive decline in older adults at risk of developing AD.