After a Concussion, Under 2 Hours of Screens Tracked Slower

A teenager comes home from a soccer field with a headache and a note from an athletic trainer, and the household reorganizes itself around a rule repeated in emergency departments and school offices for most of the past decade: no screens. Phone away, television off, homework suspended. The rule is easy to say, hard to enforce, and — according to a cohort study published in the British Journal of Sports Medicine on 9 June 2026 — probably wrong at one end.

The study, led by Jingzhen Ginger Yang of Nationwide Children's Hospital with Thomas Pommering, is small and unusually well instrumented. Eighty adolescents aged 11 to 17 with physician-diagnosed concussion were enrolled within 72 hours of injury: mean age 14.3, 65% boys, 71.3% white. Each wore a Narrative Clip camera that captured a still image every 30 seconds through the first seven days, which researchers coded into out-of-school screen time by type — smartphone, television, computer or tablet, gaming. Symptoms were tracked by daily Post-Concussion Symptom Scale surveys, and participants were followed until symptoms resolved or 45 days elapsed. Mean PCSS scores fell from 33.1 on day one to 14.1 on day seven.

The instrumentation is the point. Most prior evidence on screens after concussion rests on adolescents telling a researcher how much they used their phone, which is a question adolescents answer imprecisely in the best conditions and worse after being told the correct answer is 'not much.' A camera on the chest does not care what the correct answer is.

Adjusted for covariates, 141 minutes of screen time a day across the first three days was associated with a 35% faster rate of symptom resolution than 260 minutes. That is the sentence the coverage carried, and it is worth reading precisely: it compares two points on a modeled dose-response curve, not two groups of children who were assigned anything.

The categorical results are easier to hold onto. Youth averaging 120 to 240 minutes a day resolved symptoms 2.27 times as fast as those under 120 minutes, with a confidence interval of 1.20 to 4.29. Measured against the heavy end — more than 240 minutes — the same moderate band resolved 1.58 times as fast, but that interval ran from 0.86 to 2.93 and so includes no difference at all. The lower arm of the U is where the statistical evidence actually sits. The upper arm, the part that would justify telling a teenager to put the phone down at four hours, is suggestive and no more.

Device type behaved differently from total minutes. Smartphone use of 120 to 240 minutes a day was associated with resolution 2.47 times as fast as under 120 minutes (1.25 to 4.86), and television of 60 to 120 minutes 3.10 times as fast as more than 120 (1.29 to 7.48). Computers, tablets and gaming showed no significant association with faster resolution — a pattern the authors do not explain and which, with 80 participants split across four device categories, should be treated as a hypothesis rather than a finding.

The central problem is direction. A concussed 14-year-old with photophobia, a splitting headache and nausea does not choose to spend forty minutes on a phone; she cannot. Low screen time in the first three days is therefore an excellent candidate marker for a severe concussion, which is also, independently, a slow-resolving one. The camera captures behavior with real precision and still cannot separate 'screens helped her recover' from 'recovering let her use screens.' Statistical adjustment narrows that gap; it does not close it, and the authors are direct that randomized trials, standardized definitions and uniform measurement are what the question requires.

The exposure data has one specific hole. School hours were not recorded. Because returning to class is itself a recovery milestone, the participants with the most missing screen data are systematically the ones who got better fastest — a gap whose effect on the estimates the accessible materials cannot sign. The per-device numbers hint at how loosely these categories map onto a day: reported averages of 224.4 minutes of smartphone, 203.5 of television, 112.8 of computer or tablet and 59.7 of gaming sum to roughly 600 minutes against a reported daily total of 358.4 — so the categories cannot be exclusive slices of an afternoon, whether from simultaneous use or from means taken over different groups of users, which the published materials do not say.

What the camera also cannot see is what was on the screen. The authors list content, brightness, resolution, color settings, use of blue-light-blocking glasses and cognitive engagement among the things they did not measure — a relevant list, because two hours of a sitcom and two hours of algebra homework are not the same neurological demand. And 80 adolescents, predominantly white, from a single institution, is not a sample from which a figure carried to three significant figures should be lifted as a target. The BMJ's own materials suggest clinicians might consider recommending roughly 141 minutes a day. That number is a point on a fitted curve wearing a prescription's clothes.

Concussion medicine has moved consistently in one direction over the past decade: away from strict rest and toward early, tolerated, sub-symptom-threshold activity. This paper is that argument arriving in the last domain where the old asceticism held on, and it arrives with objective measurement instead of recall, which is why it is worth more than its sample size alone would suggest.

Its usable content is not the number 141. It is that the floor may matter more than the ceiling — that a child kept under two hours a day appeared to do worse than one at two to four hours — and that some of what the old advice cost may have been isolation rather than photons. The earlier evidence favoring avoidance concerned the first 48 hours and rested on self-report; this study covers the first three days and rests on a camera. That is a real advance in measurement attached to a study too small to settle anything.

There is something almost funny about a decade of dark-room advice being challenged by a chest-mounted camera photographing teenagers every thirty seconds, and the result deserves to be taken seriously — but as a measurement advance, not a guideline. The honest read is narrow and useful: in 80 adolescents, the lowest-use band — under two hours a day — tracked with slower symptom resolution than two to four hours, and the statistically solid half of that finding is the low end, not the high end. The coverage overstated it in one specific way, by converting a point on a dose-response curve into '141 minutes a day' as though that were a dose. It is not a dose. It is the shape of a curve fitted to a small, mostly white, single-center sample in which the sickest children were also the least able to look at anything. The authors say randomized trials are needed, and they are right; until one exists, the defensible takeaway is that the strictest restriction found no support here — the under-two-hours band tracked with slower resolution — not that 141 minutes is a dose.