Discourse-based assessment in clinical context for persons with mild cognitive impairment: A scoping review

BackgroundDiscourse assessment may capture subtle communication changes in persons with mild cognitive impairment (MCI) that are not readily detected by standardized test batteries.ObjectiveTo identify and map clinically applicable methods and linguistic measures for assessing discourse changes in persons with MCI.MethodsThis scoping review followed JBI methodology and PRISMA-ScR reporting guidance. Observational studies assessed spoken discourse in persons with MCI, including comparisons with normocognitive controls or persons with Alzheimer's disease (AD). Studies published in English or French were identified through systematic searches of MEDLINE, PsycINFO, Linguistics and Language Behavior Abstracts, and Scopus, updated in January 2025. Studies based exclusively on automated language-processing approaches were excluded. Data from 17 of the 76 studies in the broader review program (22.4%) underwent independent double extraction, with high inter-rater agreement. Data from the remaining studies were extracted by one reviewer, with uncertainties resolved collectively. Findings were summarized descriptively.ResultsSeventeen studies were included. Narrative and descriptive tasks predominated. Across studies, 119 distinct linguistic measures were identified across six domains. Only implausible or irrelevant details showed statistically significant MCI-control differences consistently across independent studies. No measure showed this pattern for MCI-AD comparisons. Most other measures were reported in single studies or yielded inconsistent findings across tasks and samples.ConclusionsHeterogeneity in elicitation procedures, operational definitions, and linguistic measures limits comparability across studies. Methodological harmonization and further evaluation of psychometric properties and clinical feasibility are needed to support evidence-based discourse assessment by speech-language pathologists.