Presentation
Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Lightning Talks
Brain age predicts sentence processing declines in healthy aging beyond chronological age and domain general working memory abilities
There is a Poster PDF for this presentation, but you must be a current member or registered to attend SNL 2023 to view it. Please go to your Account Home page to register.
Poster D29 in Poster Session D, Wednesday, October 25, 4:45 - 6:30 pm CEST, Espace Vieux-Port
William Matchin1; 1University of South Carolina
Sentence processing abilities decline with age and are thought to be explained by declines in general working memory (WM) abilities (van Boxtel & Lawyer, 2020). However, there is strong evidence for the existence of language-specific brain regions adjacent but spatially distinct from domain-general regions (Fedorenko et al., 2011). This raises questions about the extent to which aging declines on sentence processing and general WM abilities diverge, potentially reflecting differential changes to these brain systems. We analyzed data from 154 subjects aged 20-80 who completed a variety of behavioral and neuropsychological assessments, including a simplified self-paced reading task (SPR) with materials adapted from Fedorenko et al. (2007), as part of the Aging Brain Cohort at the University of South Carolina (Newman-Norlund et al., 2021). The SPR task involves presentation of two sentence types, region-by-region, which were identical except for the relative clause: subject-relatives (e.g., the janitor who frustrated the plumber lost the key on the street), and object-relatives (e.g., the janitor who the plumber frustrated lost the key on the street). Following each sentence, a comprehension question was presented. We removed subjects who had average reaction times (RT) in the critical region three or more standard deviations above the mean (Fedorenko et al., 2007), leaving 152 subjects, analyzing RT in a repeated measures ANOVA. We analyzed the NIH Toolbox List Sorting WM task (Tulsky et al., 2013) as a control for more general cognitive abilities, requiring subjects to reorder sequences presented visually and auditory simultaneously, designed to assess both phonological and visuospatial components of WM ability. For 123 subjects, we were able to compute gray matter, white matter, and cerebrospinal fluid components from high-resolution T1 images. We retained top components accounting for 80% of variance in PCA. We estimated brain age, effectively a measure of how old a subject’s brain looks independently of their real chronological age, using the BrainAgeR pipeline (github.com/james-cole/brainageR) v2.1, which was trained on n = 3377 healthy individuals (mean age = 40.6 years, SD = 21.4, age range 18-92 years) from seven publicly-available datasets. When including chronological age, brain age, education, and WM ability as covariates, there were significant interactions between structure and chronological age, F(1,117) = 7.474, p = 0.007, and between structure and brain age, F(1,117) = 4.420, p = 0.038. WM ability was significantly predicted by chronological age, t(3,119) = -2.441, p = 0.016 and education, t(3,119) = 1.998, p = 0.048 but not brain age, f(3,119) = -0.224, p = 0.823. These results suggest that the online speed with which people handle structural complexity in SPR declines with age independently of general WM ability, consistent with the hypothesis that sentence processing relies on separate/additional neurological resources relative to domain-general cognitive abilities (Caplan & Waters, 1999; Fedorenko, 2014; Matchin, 2018). Finally, brain age appears to be important independent predictor of age-related sentence processing declines, sensitive enough to detect effects of structure for self-paced reading but not redundant with chronological age or domain-general processing abilities.
Topic Areas: Syntax and Combinatorial Semantics, Reading