Review of assessment methods for dementia

Abstract

AbstractBackgroundMore than 50 million people worldwide live with various types of dementia, which incorporates a progressive decline in patients' cognition. Early detection of dementia in older adults is crucial and helps them apply for intervention programs. Numerous assessment methods have been developed to be used in clinical practice and help clinicians accurately detect dementia and diagnose its types. The recent advancement in the artificial intelligence domain has revolutionized the early detection of dementia and allowed clinicians to use AI‐based assessment technologies for detecting dementia at its very mild decline stage. The AI community aims to develop valid, efficient, practical, reliable, and accurate AI‐based assessment technologies to detect dementia; thus, to help the community, this paper reviews clinical assessment methods and recently developed AI‐based assessment technologies, particularly AI‐based language and speech assessment techniques. Moreover, it collects information about available datasets, python libraries for developing AI assessment methodologies to detect dementia.MethodWe have extracted around 300 peer‐reviewed articles from PubMed and MEDLINE related to AI‐based assessment technologies (i.e., focusing on automated language and speech‐based assessment methods) and cognitive, behavioral, neurophysiological, functional assessment methods.ResultBy reviewing more than 50 various clinical and AI assessment methods, we listed the advantages and limitations of clinical and AI assessment methods and available machine learning datasets, python scripts, and libraries to develop AI assessment methods to detect dementia.ConclusionThe paper could provide valuable information for the AI community to have better progress in developing quick, accurate, multicultural, practical assessment technologies with high specificity and sensitivity. The paper can be beneficial for mental health clinicians to learn about known AI‐based assessment methods to detect dementia.