Background

The effects of aging on circadian patterns of behavior are insufficiently described. To address this, we characterized age-specific features of rest-activity rhythms (RAR) in community dwelling older adults both overall, and in relation, to sociodemographic characteristics.

Methods

We examined cross-sectional associations between RAR and age, sex, race, education, multimorbidity burden, financial, work, martial, health, and smoking status using assessments of older adults with wrist-worn free-living actigraphy data (N=820, Age=76.4 yrs, 58.2% women) participating in the Study of Muscle, Mobility and Aging (SOMMA). RAR parameters were determined by mapping an extension to the traditional cosine curve to activity data. Functional principal component analysis determined variables accounting for variance.

Results

Age was associated with several metrics of dampened RAR; women had stronger and more robust RAR vs. men (all P < 0.05). Total activity (56%) and time of activity (20%) accounted for most the RAR variance. Compared to the latest decile of acrophase, those in the earliest decile had higher average amplitude (P <0.001). Compared to the latest decile of acrophase, those in the earliest and midrange categories had more total activity (P=0.02). Being in a married-like relationship and a more stable financial situation were associated with stronger rhythms; higher education was associated with less rhythm strength (all P < 0.05).

Conclusions

Older age was associated with dampened circadian behavior; behaviors were sexually dimorphic. Some sociodemographic characteristics were associated with circadian behavior. We identified a behavioral phenotype characterized by early time-of-day of peak activity, high rhythmic amplitude, and more total activity.

BACKGROUND: Aging is associated with declines in circadian functions. The effects of aging on circadian patterns of behavior are insufficiently described. We characterized age-specific features of rest-activity rhythms (RAR) in community dwelling older adults, both overall, and in relation, to sociodemographic characteristics. METHODS: We analyzed baseline assessments of older adults with wrist-worn free-living wrist-worn actigraphy data (N=820, Age=76.4 yrs, 58.2% women) participating in the Study of Muscle, Mobility and Aging (SOMMA). We applied an extension to the traditional cosine curve to map RAR to activity data, calculating the parameters: rhythmic strength (amplitude); robustness (pseudo-F statistic); and timing of peak activity (acrophase). We also used function principal component analysis to determine 4 components describing underlying patterns of activity accounting for RAR variance. Linear models were used to examine associations between RAR and sociodemographic variables. RESULTS: Age was associated with several metrics of dampened RAR; women had stronger and more robust RAR metrics vs. men (all P < 0.05). Total activity (56%) and time of activity (20%) accounted for most the RAR variance. Compared to the latest decile of acrophase, those in the earliest decile had higher average amplitude (P <0.001). Compared to the latest decile of acrophase, those is the earliest and midrange categories had more total activity (P=0.02). RAR was associated with some sociodemographic variables. CONCLUSIONS: Older age was associated with dampened circadian behavior; and behaviors were sexually dimorphic. We identified a behavioral phenotype characterized by early time-of-day of peak activity, high rhythmic amplitude, and more total activity.