No statistically significant longitudinal changes in actigraphy-derived sleep parameters were detected between the subacute phase and later stages after stroke, with sleep metrics remaining outside normal ranges at both time points.
Key Findings
Results
No statistically significant longitudinal changes in any actigraphy-derived sleep parameter were detected between the subacute phase and follow-up after stroke.
Study included 18 individuals with stroke who completed seven-day wrist-actigraphy monitoring at baseline and at a follow-up assessment more than three months after stroke.
The median interval between assessments was 161 days (IQR 102.3–194.0).
Parameters compared using paired t-tests included time in bed, total sleep time, sleep latency, sleep efficiency, WASO, and number of awakenings.
All changes were non-statistically significant and effect sizes were described as small.
Results
Time in bed showed a mean change of -0.12 hours between the two assessment points.
Mean paired difference for time in bed was -0.12 h (95% CI -0.73 to 0.49).
This change was not statistically significant.
Effect size was small.
Results
Total sleep time showed a mean change of -0.05 hours between assessments.
Mean paired difference for total sleep time was -0.05 h (95% CI -0.59 to 0.50).
This change was not statistically significant.
Effect size was small.
Results
Sleep latency showed a mean reduction of 7.93 minutes between the subacute phase and follow-up.
Mean paired difference for sleep latency was -7.93 min (95% CI -19.55 to 3.69).
This change was not statistically significant.
Effect size was small.
Results
Sleep efficiency showed a mean change of -0.36 percentage points between assessments.
Mean paired difference for sleep efficiency was -0.36% points (95% CI -4.19 to 3.48).
This change was not statistically significant.
At the individual level, sleep efficiency increased in 10 participants and decreased in 8, indicating heterogeneous trajectories.
Results
Wake after sleep onset (WASO) showed a mean increase of 3.63 minutes between assessments.
Mean paired difference for WASO was 3.63 min (95% CI -9.22 to 16.48).
This change was not statistically significant.
Effect size was small.
Results
Number of awakenings showed a mean increase of 0.43 between the two time points.
Mean paired difference for number of awakenings was 0.43 (95% CI -0.74 to 1.59).
This change was not statistically significant.
Effect size was small.
Discussion
Sleep metrics remained outside normal ranges at both the subacute baseline and the follow-up assessment.
Despite the passage of a median of 161 days between assessments, sleep parameters did not normalize.
The persistence of abnormal sleep metrics at both time points suggests that sleep disturbances may not self-resolve during the subacute-to-later recovery period.
Wide confidence intervals and a small, heterogeneous sample prevented definitive conclusions about long-term trends.
Results
Individual-level analysis revealed heterogeneous trajectories of sleep efficiency after stroke.
Of the 18 participants, 10 showed increased sleep efficiency and 8 showed decreased sleep efficiency at follow-up.
This heterogeneity was described as an exploratory individual-level finding.
The authors note that group-level null findings may obscure clinically meaningful variation between individuals.
What This Means
This research followed 18 stroke survivors over time to track how their sleep changed during recovery. Using wrist-worn activity monitors (actigraphy) worn for seven days at two separate time points — one in the early recovery phase and one more than three months later (on average about 5 months apart) — the researchers measured several aspects of sleep including how long people slept, how long it took to fall asleep, how often they woke up, and how efficiently they slept. Despite the time that passed, none of these sleep measures changed significantly between the two assessments.
Importantly, sleep quality remained poor at both time points — meaning that sleep problems did not appear to resolve on their own during this period of recovery. When looking at individuals rather than the group as a whole, there was notable variation: some people's sleep improved while others' worsened, suggesting that stroke survivors do not all follow the same path when it comes to sleep recovery. The wide range in outcomes and the small sample size mean that firm conclusions cannot be drawn.
This research suggests that sleep disturbances after stroke may persist well into the recovery period without resolving naturally, which could have implications for rehabilitation. The finding that sleep remained outside normal ranges at both time points highlights a potential need for targeted sleep interventions for stroke survivors. However, because the study was small and findings varied considerably between individuals, larger studies are needed to better understand who is at risk for ongoing sleep problems and what factors might help or hinder sleep recovery.
Tenório R, Biasibetti I, Mota Alvarenga M, Teixeira-Salmela L, Avelino P, Silva A, et al.. (2026). Trajectories of sleep disturbances in the subacute phase after stroke: a longitudinal actigraphy study.. Sleep & breathing = Schlaf & Atmung. https://doi.org/10.1007/s11325-026-03810-z