Marius Comper

Biophysics · two published studies

Why brain freeze hurts your forehead

You eat a spoonful of ice cream too fast. A few seconds later, a sharp pain shoots across your forehead, even though the cold was on your tongue. It's a real reflex, tested in a clinical trial: eat ice cream in under 5 seconds and you get this pain 2.2 times more often than if you take over 30 seconds.

The reason is a reflex inside your brain. Cold on the tongue suddenly triggers a surge of blood flow in a cerebral artery, and a single nerve carries signals from both the tongue and that artery. Your brain mixes them up and puts the pain in your forehead.

A direct comparison, from the same trial

12.5%cautious · over 30 seconds
27.4%fast · under 5 seconds
2.2×relative risk
cautious 12.5% fast 27.4%

Bars are proportional to the published percentages. Kaczorowski & Kaczorowski, BMJ, 2002. 145 schoolchildren, 100 ml of ice cream each.

2.2×relative risk, fast pace versus cautious
20 smedian latency to pain
12 smedian pain duration

What you control

Eating faster increased headache risk

In the ICE-H trial, 145 middle-school students (mean age 12.7) each got the same portion: 100 ml of ice cream. The only variable tested was consumption time: one group ate it over more than 30 seconds, the other in under 5.

Headache incidence, by group
GroupParticipantsWith headacheIncidence
cautious (>30 s)72912.5%
fast (<5 s)732027.4%

The relative risk computed from these numbers is 2.19, published as 2.2 (95% confidence interval: 1.03–4.94). The number needed to harm is 6.71: for roughly every seven people who switch from cautious to fast eating, one extra headache occurs. Of the 29 total headaches reported, 17 (58.6%) lasted under 10 seconds.

What happens inside the artery

The stronger the reflex, the more likely the pain

In a second study, 77 healthy adults drank 200 ml of ice water (0°C) through a straw pressed against the roof of the mouth, while transcranial Doppler ultrasound measured, second by second, blood flow velocity in both middle cerebral arteries. In 39 of the 77 participants (50.6%), the cold stimulus provoked a typical headache, on both sides of the head, with a stabbing sensation.

The pain didn't appear right away: median latency was 20 seconds, and median duration was 12 seconds. The cold stimulus was identical for every participant. What differed was how hard the blood vessel reacted, measured by its flow velocity.

Rise in blood flow velocity, by group

4.6%without headache
9.3%with headache
without 4.6% with 9.3%

Bars are proportional to the published percentages (p=0.003). Hensel et al., Frontiers in Neurology, 2019.

Those who also had tearing (19 of 39, or 48.7%) showed an even larger increase in flow: 12.8% versus 5.9% for those without tearing (p=0.006). In a quarter of those with a headache (10 of 39, 25.6%), a second headache followed shortly after the first one resolved, with a latency of 12 seconds (range 3 to 26 seconds).

Why you feel it in your forehead

One nerve links the roof of your mouth to the vessels in your brain

The phenomenon has a familiar precedent. In a heart attack, pain can show up in the left arm, because the nerves from the heart and the nerves from the arm send signals to the same neurons in the spinal cord: the brain gets one combined signal and misplaces it. Doctors call this pattern referred pain.

Something similar happens with ice cream. The roof of the mouth and the blood vessels in the brain are served by the same nerve, called the trigeminal nerve. The fibers that sense cold on the roof of the mouth and the fibers that sense the cerebral artery stretching reach the same neurons, at one point in the brainstem. The brain gets a single mixed signal and places it where it usually expects news from those fibers: the forehead and temples.

The authors of the Doppler study call this pattern a "trigeminal-parasympathetic vasodilator reflex," the same circuit suspected in other vascular headaches. The cold on your tongue stays on your tongue. What travels on, to your forehead, is only the nerve signal getting mixed up along the way.

Limits

What these two studies don't cover

  • The samples are small by epidemiological standards: 145 schoolchildren in one trial, 77 adults in the other. The confidence interval on the relative risk (1.03–4.94) is wide.
  • The ICE-H trial used schoolchildren and ice cream. The Doppler trial used adults and ice water. The populations and stimuli of the two studies differ, though the described mechanism is the same.
  • Neither study directly measured trigeminal convergence. It is the accepted model in the neurology literature, not an observation made in either experiment.
  • Headache was reported by participants in both studies, not confirmed by a separate objective marker.

Sources

Studies cited

  • Kaczorowski M, Kaczorowski J. "Ice cream evoked headaches (ICE-H) study: randomised trial of accelerated versus cautious ice cream eating regimen." BMJ, 2002. pmc.ncbi.nlm.nih.gov/articles/PMC139031
  • Hensel O, Burow P, Mages S, Wienke A, Kraya T, Zierz S. "Increased Blood Flow Velocity in Middle Cerebral Artery and Headache Upon Ingestion of Ice Water." Frontiers in Neurology, 2019. pmc.ncbi.nlm.nih.gov/articles/PMC6611440