You don’t remember it, but you woke up at least 100 times last night. These spontaneous arousals, lasting less than 15 seconds each, occur roughly every five minutes and don’t seem to affect how well-rested you feel. They are unrelated to waking up from a bad dream or your partner tossing and turning. Instead, they seem to be linked to some internal biological mechanism.

Frequently waking up throughout the night may have protected early humans from predators by increasing their awareness of their surroundings during sleep. “The likelihood someone would notice an animal is higher [if they] wake up more often,” says Ronny Bartsch, a physicist at Bar-Ilan University in Israel. “When you wake up, you’re more prone to hear things. In deep sleep, you’re completely isolated.”

Sleep scientists, however, have been stumped as to what triggers these nocturnal disruptions. In a new Science Advances paper Bartsch proposes an innovative hypothesis that spontaneous arousals are due to random electrical activity in a specific set of neurons in the brain—aptly named the wake-promoting neurons.

Even when you are asleep your brain cells continuously buzz with a low level of electrical activity akin to white noise on the radio. Occasionally, this electrical clamor reaches a threshold that triggers the firing of neurons. The new paper suggests that when random firing occurs in the wake-promoting neurons, a person briefly jerks awake. But this is countered by a suite of sleep-promoting neurons that helps one quickly fall back to sleep.

Low-level electrical activity in neurons increases in colder temperatures whereas warmer temperatures flatten it. As a result, there should be fewer spontaneous arousals in hot weather. To test this theory, the researchers created computer models that mapped how neuronal noise should act at different temperatures and how the varying electrical activity could affect spontaneous arousals. They also measured sleep in zebra fish, which have similar day/night cycles to humans but are ectothermic, meaning their body temperature is controlled by the environment rather than by internal processes.

The researchers compared the fish’s sleep rates at four different water temperatures: 77, 82 (ideal for zebra fish), 84 and 93 degrees Fahrenheit. Across the board, the colder the water the more often the zebra fish woke up and the longer they stayed awake. The data from the zebra fish and the models of temperature, neuronal noise and arousal matched perfectly. “I think their theory is a perfectly good one and may even be correct,” says Clifford Saper, a neuroscientist at Harvard Medical School’s Division of Sleep Medicine and head of neurology at Beth Israel Deaconess Medical Center who was not involved with the study. “But the experiment they did doesn’t test that hypothesis.”

The zebra fish experiment shows the fish wake up more frequently and stay awake for longer in colder temperatures but reveals nothing about these animals’ neuronal noise—or humans’, for that matter. Bartsch says that, so far, no studies have figured out how to measure neuronal noise in a sleeping animal.

The idea that warm temperatures cause fewer nocturnal disruptions also seemingly flies in the face of conventional wisdom that a colder bedroom leads to better sleep. But waking up because you are hot and uncomfortable is different from these brief spontaneous arousals. In fact, our bodies are pretty good at regulating their core brain and body temperatures, so the difference of a few degrees outside would not alter neuronal activity. In contrast, zebra fish’s temperature varies quite a bit. Saper says because of this zebra fish “are probably the last animal that I would use to try to make this point.”

Bartsch emphasizes the study is not trying to make a claim about thermoregulation in adults but he says it may have implications for newborn babies. “Because very young infants are more ectothermic than endothermic, their arousability could scale similarly to fish for different ambient temperatures.”

Infants are not as good at regulating their own temperature and so are more vulnerable to changes in the environment. (This is why premature babies have to be kept in incubators.) Consequently, the researchers think newborns may be more susceptible to heat-related fluctuations in neuronal noise.

The theory may have important implications for infant sleep. Although they may be disruptive to parents, spontaneous arousals could help save a baby’s life. Sudden infant death syndrome (SIDS) has been a leading cause of mortality in children between one month and one year of age and yet largely remains a mystery. One idea is that SIDS is caused by a stoppage in breathing, often through accidental suffocation. Waking up during the night can prompt babies to shift or cry out, helping to ensure that they do not have anything obstructing their airways and are still breathing. “We came up again with a theory that the babies with SIDS have low neuronal noise and therefore they have lower arousals,” says Hila Dvir, a physicist at Bar-Ilan University who co-authored the paper.. “Because they have low arousals, they are less protected from any hypoxic event—a shortage of oxygen.”

Not everyone is convinced, though. “Over the years, people have come up with ideas to explain SIDS, like a single explanation for it, and they just keep hitting dead ends with it because it's probably a complex, heterogeneous situation,” says Rafael Pelayo, a clinical professor at the Stanford Center for Sleep Sciences and Medicine. “It is a cool idea that this neuronal noise is explaining the arousals. I just think they jumped a little bit when they got into SIDS. It has to be more complicated than that.”