Autor: |
Hale, Madeline R, Van Hulle, Carol A., Langhough, Rebecca E, Basche, Kristin E, Chin, Nathaniel A., Christian, Bradley T., Betthauser, Tobey J, Hermann, Bruce P, Johnson, Sterling C., Mueller, Kimberly D |
Zdroj: |
Alzheimer's & Dementia: The Journal of the Alzheimer's Association; Dec2023 Supplement 18, Vol. 19, p1-5, 5p |
Abstrakt: |
Background: Changes in connected speech (CS; language used in everyday conversations) have been documented throughout the Alzheimer's disease (AD) spectrum, including the very early stages of disease. Our group identified four factors comprised of 10 measures of CS: fluency, semantic content, lexical diversity, and grammatical complexity. We demonstrated that fluency and semantic content were associated with subclinical cognitive decline (Mueller et al., 2018a). However, the relationship between these connected speech factors and AD biomarkers in preclinical AD is unknown. Method: Cognitively unimpaired (CU) Wisconsin Registry for Alzheimer's Prevention participants with positron emission tomography (PET) amyloid (A) and tau (T) were identified; 26 were A+/T+; an equal number matched 1:1 using nearest neighbor propensity score matching for age, sex, literacy, and race were selected from A‐/T‐ and A+/T‐ subsets (n = 78). A priori power calculation based on similar matched designs revealed >80% power to detect a medium effect size. We analyzed CS samples from a picture description task (Cookie Theft) at the neuropsychological visit closest to the most recent PET scans (median visit number = 6). Language measures were extracted from transcripts and transformed into standardized z‐scores that were used to create the factors (details in Mueller et al. (2018b)). ANOVA models investigated whether factor scores and two exploratory individual measures of CS efficiency (content information units per minute (CIU/Min) and semantic unit idea density (SUID)) differed across PET biomarker profiles (A‐/T‐ = reference group). Post‐hoc pairwise comparisons were then completed, and Cohen's d characterized effect sizes. Result: The mean(sd) years between CS sample and PET scan was 2.3(1.8) for amyloid and 2.1(1.7) for tau (Table 1). The lexical factor was associated with PET biomarker status, such that participants who were A+/T+ had lower lexical diversity than the A‐/T‐ group. Lower SUID was associated with both A+/T‐ and A+/T+ classification (Figure 1). Associations did not change when analyses were repeated without outliers. Effect sizes were highest for SUID (Table 2). Conclusion: Of four previously derived language factors, lexical diversity was associated with PET biomarker status. However, a measure of connected speech efficiency (SUID) was also associated with biomarker status in this preclinical, CU sample. Replication in additional cohorts is needed. [ABSTRACT FROM AUTHOR] |
Databáze: |
Supplemental Index |
Externí odkaz: |
|