Beyond sleep: Rest and activity rhythm as a marker of preclinical and mild dementia in older adults with less education.

Background: Although sleep duration and sleep-related breathing disorders were associated with dementia previously, few studies examined the association between circadian rhythm association and cognitive status.

Objective: We aimed to investigate the association of rest and activity rhythm with cognitive performance in older people with cognitive complaints and less education.

Methods: Activity rhythm was evaluated with wrist actigraphy in 109 community-dwelling older people with cognitive complaints without diagnosed dementia. Each participant completed a neuropsychological battery and was classified as having cognitive impairment (MCI), dementia, or normal cognition. We used adjusted multinomial logistic regression and linear regression models to compare sleep and circadian non-parametric measures with cognitive groups and cognitive z-scores, respectively.

Results: The mean age of the 109 participants was 79.3 ± 6.3 years old, 74% were women, 68% were white, and the mean education was 5.6 ± 5.2 years. Daytime activity intensity was associated with better language (β = 0.178; 95% CI = 0.022, 0.334; p = 0.03) and visuospatial performance (β = 0.158; 95%CI = 0.008, 0.308; p = 0.04). Also, less fragmented rhythm was associated with better visuospatial (β = 0.172; 95%CI = 0.025, 0.320; p = 0.02) and global cognitive scores (β = 0.134; 95%CI = 0.005, 0.263; p = 0.04). More interdaily stability was associated with a lower risk of MCI and dementia (RR = 0.54; 95%CI = 0.29-0.99; p = 0.04, and RR = 0.44; 95%CI = 0.21-0.94; p = 0.03, respectively). Moreover, more daytime activity (RR = 0.40; 95%CI = 0.18-0.89; p = 0.02) and less rhythm fragmentation (RR = 0.31; 95%CI = 0.14-0.73; p = 0.007) were associated with lower risk for dementia.

Conclusion: Daytime activity intensity and fragmented rhythm during the day and night may play an important role as markers for cognitive impairment in less educated populations. Future studies with larger samples should confirm these findings.