A 20-Minute Workout Matched a 90-Minute Nap for Memory After a Sleepless Night

A sleepless night measurably weakens episodic memory, the kind that lets someone recall a specific photo, conversation, or event days later. What has been less clear is what, short of sleep itself, can blunt that damage. A study published August 24 in the Proceedings of the National Academy of Sciences tested two candidates head to head in the same experiment for the first time: a short bout of exercise and a nap.

Researchers at McGill University kept 54 healthy adults, ages 18 to 35, awake for 30 consecutive hours under lab supervision, then split them into three groups. One group cycled at moderate to vigorous intensity for 20 minutes. A second group got a 90-minute nap opportunity, sleeping with a median efficiency of about 96 percent. A third group sat on a stationary bike without pedaling, as a control. All three then viewed a set of images and were tested three days later on how many they could recognize.

The gap between each intervention and the control group was large by the study's own statistical measure, with an effect size (Cohen's d) of about 1.0 for exercise and 0.91 for the nap. Madhura Lotlikar, the study's first author and a doctoral candidate in McGill's Department of Neurology and Neurosurgery, framed the practical case for exercise as one of access rather than superiority.

“A nap isn't always possible in the middle of a shift. Exercise is accessible, inexpensive and easy to implement. That makes it a promising tool to help people stay cognitively sharp when sleep is limited.”

Two paths to the same score

Brain recordings taken during the memory task showed the two groups were not doing the same thing internally, even though they landed in the same place. In the exercise group, memory performance was best predicted by activity occurring after each image appeared, specifically P300 amplitude and beta-wave desynchronization, together explaining 64 percent of the variation in results. Marc Roig, the study's senior author and a professor in McGill's School of Physical and Occupational Therapy, said the pattern suggests exercise helps the brain use whatever resources sleep loss left it with more efficiently, rather than restoring those resources directly.

“Sleep loss affects nearly every aspect of how we think and function, but many people can't simply stop what they're doing and get more sleep.”

In the nap group, the predictors ran the other direction. Brain-state and fatigue markers measured before each image appeared, including theta wave activity and slow-wave patterns tied to accumulated sleep pressure, explained 81 percent of the variation in memory scores. That fits the standard account of what naps do: lower the brain's sleep pressure directly, rather than help it work around a deficit. Roig's collaborators on the paper span McGill's neurology and physical therapy departments along with researchers at Montreal's Jewish Rehabilitation Hospital, the University of Glasgow, and the Universite de Montreal.

What the study doesn't show

The finding does not mean a workout is interchangeable with sleep, and the authors do not claim that. The 54 participants were healthy, young, and tested after a single lab-induced night of total sleep loss, not the chronic, partial sleep restriction more common outside a lab. The study did not separate two related but distinct memory processes, recollection and familiarity, and it did not account for participants' education, occupation, or socioeconomic status, any of which can independently affect memory performance. The exercise and nap sessions also ran different lengths by design, 20 minutes against 90, so this is not a test of equal time investment.

The stakes are ordinary rather than dramatic. Nationally, 30.5 percent of American adults reported sleeping less than seven hours a day in 2024, according to a National Center for Health Statistics data brief, and shift workers, medical residents, and new parents routinely go without a full night's rest. For that population, the study offers a specific, narrow, and unreplicated data point: a 20-minute bout of exercise can substitute for a nap on one particular measure, next-day memory for images seen once, when a nap is not an option.