Babatope Eyitomilayo Yemisi, Ramírez-Acosta Alejandro Álvaro, Avila-Funes José Alberto, García-Vázquez Mireya
Instituto Politécnico Nacional, Centro de Investigación y Desarrollo de Tecnología Digital, Tijuana 22435, Mexico.
MIRAL R&D&I Multimedia, San Diego, CA 92154, USA.
J Clin Med. 2024 Nov 22;13(23):7068. doi: 10.3390/jcm13237068.
The growing incidence of cognitive impairment among older adults has a significant impact on individuals, family members, caregivers, and society. Current conventional cognitive assessment tools are faced with some limitations. Recent evidence suggests that automating cognitive assessment holds promise, potentially resulting in earlier diagnosis, timely intervention, improved patient outcomes, and higher chances of response to treatment. Despite the advantages of automated assessment and technological advancements, automated cognitive assessment has yet to gain widespread use, especially in low and lower middle-income countries. This review highlights the potential of automated cognitive assessment tools and presents an overview of existing tools. This review includes 87 studies carried out with non-neuroimaging data alongside their performance metrics. The identified articles automated the cognitive assessment process and were grouped into five categories either based on the tools' design or the data analysis approach. These categories include game-based, digital versions of conventional tools, original computerized tests and batteries, virtual reality/wearable sensors/smart home technologies, and artificial intelligence-based (AI-based) tools. These categories are further explained, and evaluation of their strengths and limitations is discussed to strengthen their adoption in clinical practice. The comparative metrics of both conventional and automated approaches of assessment suggest that the automated approach is a strong alternative to the conventional approach. Additionally, the results of the review show that the use of automated assessment tools is more prominent in countries ranked as high-income and upper middle-income countries. This trend merits further social and economic studies to understand the impact of this global reality.
老年人认知障碍发病率的不断上升对个人、家庭成员、护理人员和社会都产生了重大影响。当前传统的认知评估工具存在一些局限性。最近的证据表明，认知评估自动化具有前景，可能会带来更早的诊断、及时的干预、改善患者预后以及更高的治疗反应几率。尽管自动化评估具有优势且技术不断进步，但自动化认知评估尚未得到广泛应用，尤其是在低收入和中低收入国家。本综述强调了自动化认知评估工具的潜力，并概述了现有工具。 本综述纳入了87项使用非神经影像数据并带有其性能指标的研究。所纳入的文章将认知评估过程自动化，并根据工具设计或数据分析方法分为五类。这些类别包括基于游戏的、传统工具的数字版本、原始计算机化测试及组合、虚拟现实/可穿戴传感器/智能家居技术以及基于人工智能的工具。对这些类别进行了进一步解释，并讨论了对其优缺点的评估，以加强它们在临床实践中的应用。传统评估方法和自动化评估方法的比较指标表明，自动化方法是传统方法的有力替代方案。此外，综述结果表明，自动化评估工具在高收入和中高收入国家的使用更为突出。这一趋势值得进一步开展社会和经济研究，以了解这一全球现状的影响。
