In a Chicago-based observational study of 324 adults aged 55 and older (64% women, 37% Black, 14% Hispanic/Latino), four distinct sleep dimensions measured by 7-day wrist actigraphy were independently associated with cognitive performance. Greater sleep fragmentation and less sleep regularity were linked to worse fluid cognition and Trail Making Test B scores (all p<0.01). Longer sleep duration and earlier sleep timing were associated with better attention and processing speed. Notably, no sleep measure predicted cognitive change over an average 1.9-year follow-up in the 133-person subset.
This preprint, not yet peer-reviewed, adds meaningful nuance to an already substantial literature by simultaneously testing four sleep domains rather than isolating duration alone — a common limitation in earlier work. The racial and ethnic diversity of the sample is a genuine strength, since most sleep-cognition studies have been conducted in predominantly white cohorts. The finding that sleep regularity — essentially keeping a consistent bedtime — independently predicts cognition aligns with emerging circadian research suggesting rhythm stability matters beyond hours slept. Practically, adults can target regularity and fragmentation through consistent sleep schedules and treating conditions like sleep apnea, steps that require no prescription. The critical caveat: the cross-sectional associations dominate, and the small longitudinal subsample (n=133, under two years) failed to detect any predictive signal, meaning causality remains unestablished. Larger, longer trials are needed before firm clinical guidance can follow.