Linguistic Features from Paragraph Recall are Markers of Cognitive Impairment
Preprint
Abstract
Objective: Cognitive impairment is associated with language changes that may be elicited from verbal responses during neuropsychological assessments that are not captured in traditional scoring. The current study investigated the utility of a linguistic analysis of paragraph recall responses for differentiating participants with and without cognitive impairment. Methods: Digital voice recordings of Logical Memory (LM) were available from 598 participants from the Long Life Family Study with normal cognition and 112 with cognitive impairment. Linguistic polyfeature scores for immediate (PFS-IR) and delayed recall (PFS-DR) were created from a weighted sum of features associated with cognitive impairment. Logistic regression models assessed the predictive value of each PFS and demographics for classifying cognitive impairment. Repeated measures models with Generalized Estimating Equations assessed whether PFSs predict decline on a cognitive screener. Results: Both immediate and delayed PFSs were significantly associated with cognitive status (PFS-IR β = 0.05, p< .001; PFS-DR β = 0.07, p< .001). A classifier with PFS-DR and demographics closely approximated the accuracy of the traditional LM score and demographics (AUC-PR = 0.81 vs 0.84, respectively). A higher PFS-DR was also associated with greater cognitive decline over an average of 5 years of follow-up (β = -0.08, p< .001). Conclusion: Quantification of linguistic features from paragraph recall using a linguistic PFS provides sufficient information for detecting cognitive impairment and predicting incident cognitive decline. The linguistic PFS has the potential to be integrated into automated testing, recording, and scoring pipelines allowing for the implementation of sensitive neuropsychological assessments in broader clinical and research settings.Keywords
This publication has 37 references indexed in Scilit:
- Increased genetic protection against Alzheimer’s disease in centenariansGeroScience, 2025
- Automated remote speech‐based testing of individuals with cognitive decline: Bayesian agreement of transcription accuracyAlzheimer's & Dementia: Diagnosis, Assessment & Disease Monitoring, 2024
- Automated detection of mild cognitive impairment and dementia from voice recordings: A natural language processing approachAlzheimer's & Dementia, 2022
- A finer-grained linguistic profile of Alzheimer's disease and Mild Cognitive ImpairmentJournal of Neurolinguistics, 2022
- Linguistic markers predict onset of Alzheimer's diseaseeClinicalMedicine, 2020
- Episodic memory in normal aging and Alzheimer disease: Insights from imaging and behavioral studiesAgeing Research Reviews, 2015
- Semantic memory and language dysfunction in early Alzheimer's disease: a reviewInternational Journal of Geriatric Psychiatry, 2012
- Clinical‐neuroimaging characteristics of dysexecutive mild cognitive impairmentAnnals of Neurology, 2009
- Prediction of individual genetic risk of complex diseaseCurrent Opinion in Genetics & Development, 2008
- A Validation Study of the Dementia QuestionnaireArchives of Neurology, 1994