What Can We Measure with Actigraphy Data? A Descriptive and Component Structure Study from a Population-Based Sample of São Paulo city (EPISONO 4th Edition)

Keywords

accelerometry - epidemiologic methods - circadian rhythm

Introduction: Actigraphy is a well-established tool in sleep and circadian medicine, used to record activity and rest rhythms and to estimate sleep. Despite its growing importance, there is a lack of epidemiological and population-based studies investigating actigraphy parameters, highlighting the need to establish reference and normative values.

Objective: Present descriptive actigraphy data in a population-based sample and identify components that capture the greatest amount of data variation.

Methods: The data set included 402 participants (60% women, 49.5±14.9 years) from a cohort of São Paulo city (EPISONO 4th edition). Data collection took place between August 2018-April 2019 (excluding data collected 15 days after the start of daylight-saving time and 15 days after its end). Sleep-wake classification was conducted using the Cole-Kriple algorithm and revised in accordance with the Consenso Brazileiro de Actigrafia guidelines. Activity and rest parameters were extracted using the COSINOR method and non-parametric analysis. The daily (Sleep Regularity Index) and weekly sleep regularity (social jet lag) were also calculated. A principal component analysis was conducted using 16 actigraphy features and linear regressions were employed (effect of sociodemographic and behavioral variables on each component).

Results: The 16 features were reduced to 5 components explaining 69.6% of the data variation: sleep quantity and opportunity (21.7% variance), sleep and rhythm timing (16.0% variance), sleep quality and continuity (12.8% variance), waking activity (11.3% variance), and sleep regularity and rhythm robustness (7.8% variance). Women had a greater amount and opportunity for sleep (β=0.35) and more activity while awake (β=0.24). Older subjects had earlier sleep and rhythm timing (β=-0.02), less activity during wakefulness (β=-0.02), and a more regular and robust rhythm (β=0.01). Subjects with earlier preferences had earlier sleep and rhythm timing (β=-0.04) and greater activity during wakefulness (β=0.01). Individuals from lower socioeconomic classes had higher activity while awake (β=0.52, 1.04) and lower sleep quality and continuity (β=-0.67). Higher BMI, lower the activity while awake (β=-0.02); higher daytime sleepiness, lower sleep regularity and rhythmicity (β=-0.03).

Conclusion: Our results can be considered a benchmark in the field. Future studies may investigate the relashionship between the components and diseases, such as insomnia and sleep apnea.

Die Autoren geben an, dass kein Interessenkonflikt besteht.

Publikationsverlauf

Artikel online veröffentlicht:
16. September 2025