Key Takeaways
- In a randomized laboratory experiment involving 88 healthy adults ages 18 to 35, people who exercised before a night of restricted sleep performed better on a next-morning memory test than those who had the same restricted sleep without exercise.
- The exercise session involved 30 minutes of moderate-intensity cycling at 55% to 60% of each participant’s measured VO₂max.
- Participants in the sleep-restriction groups were allowed 4.5 hours in bed, compared with 8.5 hours in the normal-sleep groups.
- The researchers found a significant interaction between exercise and sleep condition for next-morning declarative-memory performance.
- Slow-wave sleep was statistically involved in the relationship between experimental condition and memory, but the analysis does not prove that slow-wave sleep caused the memory benefit.
- The findings do not show that exercise can replace adequate sleep or counteract the many other effects of chronic sleep loss.
Introduction
A 30-minute moderate-intensity evening workout improved next-morning memory performance after restricted sleep in a small randomized laboratory study of healthy young adults.
A moderate evening workout may reduce some of the memory impairment associated with a single short night of sleep, according to a randomized laboratory experiment in healthy young adults.
Researchers found that participants who completed 30 minutes of moderate-intensity cycling before being restricted to 4.5 hours in bed performed better on a face-name memory test the following morning than participants who experienced the same sleep restriction without exercising (Frimpong et al., 2026).
The study was published online August 26, 2026, in SLEEP.
Unlike an observational study that simply compares people who exercise more or sleep more, this experiment randomly assigned participants to combinations of exercise and sleep opportunity. The researchers also recorded sleep with overnight polysomnography, allowing them to examine different sleep stages.
The result is intriguing, but narrow.
It does not mean that working out can undo a bad night’s sleep. The researchers tested one exercise session, one experimentally restricted night, and one type of declarative memory in a small group of healthy young adults.
What the Study Examined
Researchers recruited 92 adults ages 18 to 35 from Concordia University, the surrounding Montreal community, and social media between November 2022 and November 2024.
Four participants were excluded from the final analysis, leaving 88 participants.
The study abstract and participant table report a mean age of 24.0 years, with a standard deviation of 3.8 years. Fifty-two percent of participants were female.
Participants were screened to be generally healthy and good sleepers.
The researchers excluded people with current or previous sleep, neurological, psychological, or other relevant medical conditions. Shift workers, people taking medications known to affect cognition or sleep, and people with a body mass index of 30 kg/m² or higher were also excluded.
Participants underwent a baseline night of polysomnography to help rule out sleep disorders.
They were then randomly assigned to one of four conditions:
- 8.5-hour sleep opportunity without exercise
- 4.5-hour sleep opportunity without exercise
- exercise followed by an 8.5-hour sleep opportunity
- exercise followed by a 4.5-hour sleep opportunity
The final groups contained between 21 and 23 participants each.
What Participants Actually Did
At approximately 7 p.m., participants assigned to exercise completed a supervised cycling session.
After a five-minute warm-up, they cycled continuously for 30 minutes at 55% to 60% of their individually measured VO₂max, followed by a five-minute cool-down and stretching period.
Heart rate was monitored during the exercise.
Participants assigned to the non-exercise conditions instead remained seated quietly for approximately 40 minutes.
At 10 p.m., participants learned 80 face-name pairs.
They then completed an immediate retrieval test.
For the normal-sleep conditions, the sleep opportunity lasted from 11 p.m. until 7:30 a.m., or 8.5 hours.
Participants assigned to sleep restriction remained awake in the laboratory until 3 a.m. and then had a sleep opportunity from 3 a.m. to 7:30 a.m., or 4.5 hours.
Sleep was objectively recorded with polysomnography.
At approximately 8:30 the next morning, participants completed the delayed memory test.
What Researchers Found
The main result involved delayed memory performance.
Researchers assessed how well participants distinguished previously learned face-name combinations using a memory measure called the discriminability index, or d′.
There was a statistically significant interaction between exercise and sleep condition for delayed memory performance.
The reported p value was .033, with a partial eta-squared of .056.
Most importantly, the group that exercised before restricted sleep performed significantly better on next-morning delayed memory than the group that experienced restricted sleep without exercise.
That comparison had a p value of .028.
The exercise-plus-restricted-sleep group was not statistically different from either of the groups that received an 8.5-hour sleep opportunity.
That does not prove that exercising made 4.5 hours of sleep equivalent to a normal night.
Failure to detect a statistically significant difference between relatively small groups is not the same as demonstrating that their outcomes are truly equivalent.
Immediate memory testing also did not show the same significant exercise-by-sleep interaction.
Taken together, the results indicate that exercise influenced the pattern of next-morning memory performance under sleep restriction, but they do not show that exercise completely reversed the consequences of lost sleep.
What Slow-Wave Sleep May Have to Do With It
The researchers also examined slow-wave sleep, often called deep sleep or stage N3 sleep.
Slow-wave sleep has been implicated in the consolidation of declarative memories, including newly learned facts and associations.
The researchers performed a statistical mediation analysis and found that the proportion of sleep spent in slow-wave sleep statistically mediated part of the relationship between experimental condition and delayed memory performance.
This finding offers a possible clue about how exercise, sleep, and memory may interact.
But statistical mediation should not be confused with proof of a biological mechanism.
The researchers did not independently manipulate slow-wave sleep. Therefore, the study cannot establish that exercise improved memory specifically because it altered slow-wave sleep.
There is another important detail.
The sleep-restricted groups spent a larger percentage of their shorter sleep period in slow-wave sleep, but that did not mean they obtained more deep sleep overall.
The restricted-sleep group averaged about 87.8 minutes of slow-wave sleep, while the exercise-plus-restricted-sleep group averaged approximately 90.8 minutes.
By comparison, the normal-sleep group averaged about 127.2 minutes, and the exercise-plus-normal-sleep group averaged approximately 136.9 minutes.
A higher percentage of a short night can therefore coexist with fewer total minutes of deep sleep.
How It Compares With Earlier Research
The new experiment builds on earlier research from some of the same investigators.
In 2023, Frimpong and colleagues randomized 92 healthy young adults to combinations of normal or restricted sleep and either rest or a 15-minute high-intensity interval training session.
Participants again learned 80 face-name pairs before sleep.
That experiment found evidence that evening high-intensity exercise might partially attenuate the effect of sleep restriction on long-term declarative memory, although the exercise-plus-restricted-sleep group was not significantly better than the restricted-sleep-only group (Frimpong et al., 2023).
The newer study introduced tighter experimental controls.
Exercise intensity was individualized using measured VO₂max, heart rate was monitored, and overnight sleep was measured objectively with polysomnography.
The findings also fit with previous research showing that evening exercise does not necessarily interfere with sleep.
A systematic review and meta-analysis of 23 studies found that evening exercise generally did not worsen sleep in healthy participants and was associated with a modest increase in slow-wave sleep. However, vigorous exercise ending within an hour of bedtime could adversely affect some sleep measures (Stutz et al., 2019).
In the new experiment, exercise began at 7 p.m., several hours before participants were allowed to sleep.
What the Results May Mean
The study raises the possibility that a single bout of moderate-intensity exercise could reduce the effect of an occasional short night on this specific type of declarative memory.
That possibility should not be confused with exercise compensating for insufficient sleep generally.
Researchers tested one cognitive outcome involving long-term memory for face-name associations.
They did not show that exercise prevents the effects of inadequate sleep on attention, reaction time, decision-making, mood, metabolic health, cardiovascular function, driving performance, or accident risk.
They also did not study chronic sleep restriction.
Someone who regularly sleeps 4.5 hours a night cannot assume that exercising beforehand will neutralize the consequences.
The experiment provides evidence about one evening exercise session followed by one experimentally restricted night in healthy young adults.
Limitations
The study was relatively small.
Only 88 participants were divided among four groups, leaving between 21 and 23 people in each condition. Small groups produce less precise estimates and increase the importance of replication.
Participants were also young and unusually healthy.
The results cannot automatically be generalized to older adults, people with chronic illnesses, people with obesity, shift workers, or people with insomnia and other sleep disorders.
The study used a between-subjects design, meaning different people experienced the four conditions.
Although randomization reduces systematic differences between groups, a crossover experiment in which the same participants complete several conditions could control more effectively for individual variation.
The groups also differed in chronotype. Researchers adjusted statistically for chronotype, but statistical adjustment cannot guarantee that all residual differences were eliminated.
Another limitation is particularly important for interpreting the memory result.
Participants exercised before learning the face-name pairs.
The experiment therefore cannot completely separate a possible effect of exercise on the initial encoding of information from an effect on subsequent overnight memory consolidation.
The sleep-restricted participants also remained awake substantially longer after learning than participants assigned to normal sleep. That additional waking period could provide more opportunity for interference or forgetting.
Finally, the study examined only one night.
It cannot establish whether the same effect occurs after repeated nights of insufficient sleep or whether repeated exercise would provide a similar benefit.
Funding and Conflicts of Interest
The study was supported by the Canadian Institutes of Health Research, including grant PJT 166167.
The paper also reports additional research support for investigator Thien Thanh Dang-Vu from Canadian research organizations, the Weston Family Foundation, the American Sleep Medicine Foundation, and Concordia University. Emmanuel Frimpong received support through a Canadian Institutes of Health Research fellowship.
The authors reported no financial or nonfinancial disclosures (Frimpong et al., 2026).
Final Thoughts
In this randomized laboratory experiment, healthy young adults who completed 30 minutes of moderate-intensity evening cycling before a night restricted to 4.5 hours in bed performed better on a next-morning declarative-memory test than participants who experienced the same sleep restriction without exercise.
That is a meaningful experimental finding.
It does not show that exercise can erase sleep loss, make four and a half hours of sleep equivalent to a full night, or protect against the many other consequences of inadequate sleep.
The slow-wave-sleep results provide an interesting possible explanation for the memory effect, but they do not establish the biological mechanism.
The most defensible takeaway is narrower: one moderate evening workout improved performance on a specific next-morning memory task after one short night of sleep in healthy young adults.
Whether that benefit extends to chronic sleep loss, older adults, shift workers, people with sleep disorders, or other areas of cognition remains unknown.
References
Frimpong, E., Mograss, M., Dautet-Plourde, J., Milbrath, N., Pointeau, F., Xu, M., Nguyen, N. L. K., Zvionow, T., Aubertin-Leheudre, M., Bherer, L., Pepin, V., Robertson, E. M., Paez, A., & Dang-Vu, T. T. (2026). Slow-wave sleep mediates the compensatory effects of acute moderate-intensity exercise on declarative memory deficits in sleep restriction. SLEEP, zsag235. https://doi.org/10.1093/sleep/zsag235
Frimpong, E., Mograss, M., Zvionow, T., Paez, A., Aubertin-Leheudre, M., Bherer, L., Pepin, V., Robertson, E. M., & Dang-Vu, T. T. (2023). Acute evening high-intensity interval training may attenuate the detrimental effects of sleep restriction on long-term declarative memory. SLEEP, 46(7), zsad119. https://doi.org/10.1093/sleep/zsad119
Stutz, J., Eiholzer, R., & Spengler, C. M. (2019). Effects of evening exercise on sleep in healthy participants: A systematic review and meta-analysis. Sports Medicine, 49(2), 269–287. https://doi.org/10.1007/s40279-018-1015-0
