Demographic Effects on Longitudinal Semantic Processing, Working Memory, and Cognitive Speed

• Published in: The journals of gerontology. Series B, Psychological sciences and social sciences Vol. 75; no. 9; pp. 1850 - 1862
• Main Authors: Vonk, Jet M J, Higby, Eve, Nikolaev, Alexandre, Cahana-Amitay, Dalia, Spiro, Avron, Albert, Martin L, Obler, Loraine K
• Format: Journal Article
• Language: English
• Published: US Oxford University Press 16-10-2020
• Subjects: Aged; Aging - physiology; Aging - psychology; Cognition; Cognitive Aging - physiology; Cognitive Aging - psychology; Demography; Educational Status; Executive Function; Factor Analysis, Statistical; Female; Humans; Male; Memory, Short-Term; Mental Processes - physiology; Middle Aged; Semantics; Speech Production Measurement - psychology; The Journal of Gerontology: Psychological Sciences; Aging; Cognition; Factor analysis; Trajectory; Language
• ISSN: 1079-5014; 1758-5368
• DOI: 10.1093/geronb/gbaa080

Abstract
Objectives
To better understand and compare effects of aging and education across domains of language and cognition, we investigated whether (a) these domains show different associations with age and education, (b) these domains show similar patterns of age-related change over time, and (c) education moderates the rate of decline in these domains.
Method
We analyzed data from 306 older adults aged 55–85 at baseline of whom 116 returned for follow-up 4–8 years later. An exploratory factor analysis identified domains of language and cognition across a range of tasks. A confirmatory factor analysis analyzed cross-sectional associations of age and education with these domains. Subsequently, mixed linear models analyzed longitudinal change as a function of age and moderation by education.
Results
We identified 2 language domains, that is, semantic control and semantic memory efficiency, and 2 cognitive domains, that is, working memory and cognitive speed. Older age negatively affected all domains except semantic memory efficiency, and higher education positively affected all domains except cognitive speed at baseline. In language domains, a steeper age-related decline was observed after age 73–74 compared to younger ages, while cognition declined linearly with age. Greater educational attainment did not protect the rate of decline over time in any domain.
Discussion
Separate domains show varying effects of age and education at baseline, language versus cognitive domains show dissimilar patterns of age-related change over time, and education does not moderate the rate of decline in these domains. These findings broaden our understanding of age effects on cognitive and language abilities by placing observed age differences in context.