Evaluating Natural Language Processing for Early Dementia Detection: A Systematic Review

Dementia is a progressive neurodegenerative disorder characterized by cognitive decline that affects memory, communication, and daily functioning. Early detection is crucial. However, traditional diagnostic methods remain limited by their invasiveness, cost, and accessibility. Recent studies have identified language and speech analysis as promising, non-invasive digital biomarkers for the early detection of cognitive decline, particularly in Alzheimer’s disease (AD) and mild cognitive impairment (MCI). Changes in linguistic fluency, complexity, and semantics have been shown to correlate with both clinical assessments and neuroimaging markers of AD. Advances in Natural Language Processing (NLP) and Machine Learning (ML) now allow for the automated extraction and classification of these linguistic features, enabling scalable, efficient, and cost-effective screening for cognitive impairment.