One Short Night of Sleep Can Skew Concussion Test Scores, Study Warns

When a young athlete takes a computerized concussion test the morning after a rough night of sleep, the score on the screen may say as much about their pillow as about their brain. That is the central message of a new analysis by Aderman and colleagues, published in the Journal of Clinical Sleep Medicine, which examined thousands of baseline and post-injury neurocognitive test records from United States service academy cadets and midshipmen drawn from the CARE Consortium. The study, and a companion commentary by sleep medicine specialists Amir Sharafkhaneh, J. Kent Werner Jr., and Ahmed BaHammam, argues that prior-night sleep duration is a powerful, measurable, and almost universally ignored variable in concussion assessment. Even modest differences in how long a person slept the night before testing produced detectable changes in verbal memory, visual memory, visual motor speed, and reaction time scores on the Immediate Post-Concussion Assessment and Cognitive Testing battery, known as ImPACT, as well as on selected outcomes of the Automated Neuropsychological Assessment Metrics, or ANAM. The effects were most pronounced at early clinical time points, precisely when the highest-stakes decisions about return to play, return to duty, and return to the classroom are being made.

To understand why this matters, it helps to consider what computerized neurocognitive testing actually measures. These batteries are designed to probe the cognitive domains most vulnerable to concussion: processing speed, attention, memory, and executive function. Traumatic brain injury, which affects millions of people each year across sports, military service, falls, and traffic accidents, produces persistent deficits in exactly these domains. The large PROTECT-TBI cohort of 15,764 participants demonstrated a dose- and severity-dependent association between lifetime brain injury and impaired attention and executive function in later life, with affected individuals showing roughly half the processing speed and working memory of controls. Patients describe the experience in vivid terms. One person with brain injury put it this way: it is like the brain has a hole, and things tend to fall out. In clinical practice, these deficits surface as failures to follow instructions, forgotten test results, disorganized daily living, and impaired decision making behind the wheel when sleepy.

The relationship between sleep and brain injury is not a one-way street but a bidirectional loop. Up to half of patients with traumatic brain injury develop one or more sleep disorders, including insomnia, hypersomnia, restless legs syndrome, and circadian rhythm disorders. In turn, those sleep disorders are associated with an exacerbation of brain injury symptoms, creating a feedback cycle that can entrench cognitive impairment. Mechanistic work offers a plausible biological explanation for this vicious circle. After traumatic brain injury, the brain’s glymphatic system, the network of perivascular channels that flushes metabolic waste from neural tissue during sleep, becomes disrupted. Researchers have documented mislocalization of aquaporin-4, the key water channel protein that facilitates this clearance, along with reduced perivascular outflow. Fragmented sleep may therefore compound axonal injury by denying the damaged brain its nightly window for waste removal, turning every bad night into a potential biochemical setback.

The new findings sharpen this picture with hard numbers from a uniquely well-characterized population. Aderman and colleagues analyzed large baseline and post-injury cohorts of service academy cadets and midshipmen, showing that shorter sleep duration was associated with poorer performance on core ImPACT metrics across multiple assessment points, including the baseline test taken before any injury and the initial post-injury evaluation. This aligns with a growing body of evidence that sleep disruption directly impairs attention, memory, and processing speed, the very domains that concussion tools are built to detect. A complementary analysis of more than 25,000 ImPACT records found that hours of sleep barely altered baseline composite scores yet meaningfully shifted post-injury verbal memory, visual memory, impulse control, and symptom severity. The pattern is striking: the uninjured brain appears to compensate for a single short night, drawing on cognitive reserves to hold performance steady, whereas the injured brain has thinner reserves and amplifies the cognitive cost of the same sleep loss.

Pre-existing sleep problems may set the stage for worse outcomes even before the injury occurs. In earlier work on athletic populations, athletes with pre-injury sleep difficulties showed slower reaction time and higher symptom burden for up to two weeks after concussion compared with athletes who slept well before their injury. This suggests that sleep history is not merely a confounder to be controlled away but a genuine modifier of recovery trajectory. It also raises the possibility that some of the variability clinicians see in concussion recovery, where two patients with seemingly similar injuries follow very different courses, could be partly explained by differences in sleep that no one thought to measure. The Million Veteran Program has added a subjective dimension to this evidence, showing in a post-deployment cohort that reduced sleep duration and greater sleep disturbance are associated with poorer subjective cognitive symptoms on a validated cognition scale, reinforcing the objective findings from cadet testing.

Not every result in the new analysis points in the same direction, and the discrepancies are scientifically instructive. The sleep effect was smaller on the ANAM composites than on the ImPACT composites, and the commentary authors are candid that the reason is genuinely unclear. The two test batteries may simply differ in how sensitive they are to a short night of sleep. But alternative explanations are equally plausible: practice effects from repeated testing, ceiling effects that compress scores on ANAM, and the sheer sample size needed to detect a small statistical signal. The authors caution that the ANAM result should not be read as evidence that sleep is irrelevant to that battery. ImPACT memory and reaction time measures look particularly susceptible to short sleep, and a smaller effect on a different instrument is a difference in degree, not a license to ignore sleep when interpreting ANAM results.

The most clinically actionable insight from the study is that sleep is a modifiable and easily assessed contributor to test variability, yet it is typically neither standardized nor routinely recorded in many clinical settings. The commentary proposes a concrete protocol: sleep duration should be documented before every computerized neurocognitive test, and abnormal post-injury scores should be interpreted cautiously when the prior night’s sleep was less than six hours. A repeat assessment after a normal night of sleep can help separate true cognitive signal from measurement noise, especially before decisions about duty or activity clearance. The stakes run in both directions. Short sleep can mimic or amplify concussion-related deficits, potentially causing unnecessary delays in return-to-activity or return-to-duty decisions, while adequate sleep may temporarily mask subtle deficits and allow an impaired individual to slip through clearance thresholds. Consumer wearables, now capable of providing objective sleep measures, offer a promising route to capturing this variable at scale, and the World Sleep Society has issued recommendations for their use in monitoring sleep.

The commentary also urges active screening for insomnia, hypersomnia, and sleep-disordered breathing throughout concussion recovery, with treatment when indicated, on the grounds that managing sleep disorders may improve both cognitive functioning and the overall recovery trajectory. But the authors are careful to draw boundaries around what the new study can and cannot claim. The partial r-squared values reported were under 0.05, which are statistically real but clinically modest effects. Whether an added hour of sleep actually shortens the time to return-to-activity has not been tested. Sleep was scored from a single self-report item collected the night before testing, so sleep quality, chronic sleep restriction, and shifts in circadian phase were never captured. The data cannot say whether sleep disturbance without a change in sleep duration would have similar or different effects on test scores. And because concussion can worsen sleep while poor sleep slows recovery, an observational design cannot disentangle those two causal directions.

Generalizability presents further limits. The cadets and midshipmen in the study were young, otherwise healthy adults, whereas traumatic brain injury occurs at all ages and often in people with multiple comorbid conditions. Sleep-disordered breathing was not assessed, and the study did not record medication, caffeine, alcohol, or stimulant use, each of which can alter neurocognitive test performance. Extending these findings to older populations and to patients with chronic sleep pathology will require further data. Still, the bottom line for clinical practice is straightforward and hard to ignore: prior-night sleep should be recorded on every computerized neurocognitive test, abnormal post-injury scores should be reinterpreted when sleep was short, and ongoing sleep complaints should be evaluated and treated rather than dismissed as an expected nuisance of recovery.

What makes this story resonate beyond the clinic is its sheer simplicity. Concussion assessment has become a multibillion-dollar enterprise of sophisticated testing platforms, biomarker research, and neuroimaging, yet one of the strongest levers on test performance may be the number of hours a patient slept the night before walking into the testing room. Sleep is modifiable, biologically relevant, and cheap to measure. Overlooking it, the commentary argues, leaves an avoidable gap in concussion assessment, one that can distort scores, misdirect recovery plans, and delay or hasten return-to-play decisions for the wrong reasons. For athletes, service members, and the clinicians who clear them, the prescription begins the night before the test: document the sleep, question the score, and retest after a full night of rest before making a decision that could shape a young brain’s future.

Subject of Research: The effect of prior-night sleep duration on computerized neurocognitive testing performance at baseline and after concussion

Article Title: Lose sleep, lose score? Interpreting computerized neurocognitive testing after concussion

Article References: Sharafkhaneh, A., Werner, J. K., Jr., & BaHammam, A. (2026). Lose sleep, lose score? Interpreting computerized neurocognitive testing after concussion. Journal of Clinical Sleep Medicine, 22(1), Article 108. https://doi.org/10.1007/s44470-026-00107-6

Image Credits: AI Generated

DOI: 10.1007/s44470-026-00107-6

Keywords: concussion, traumatic brain injury, sleep duration, computerized neurocognitive testing, ImPACT, ANAM, glymphatic system, sleep disorders, return-to-play, CARE Consortium, cognitive function, sports medicine