Sixteen captive black bears fitted with implanted sensors that recorded their brain activity have given researchers the clearest picture yet of what actually happens inside a hibernating bear. Across a hibernation season, the bears spent close to 65 percent of their time asleep, roughly double the amount they slept during summer, even though their metabolism during hibernation had dropped to about a quarter of its normal rate.
The study was carried out by a team led by Øivind Tøien at the Institute of Arctic Biology, University of Alaska Fairbanks, with collaborators at Stanford University, the University of Oxford, and the Florey Institute of Neuroscience and Mental Health in Melbourne. It appeared as a preprint in March 2025 and was published in PLOS ONE in August 2026.
What the bears were actually wearing
“Wired with brain sensors” is a fair plain-language description, but it understates what the procedure involved. The 16 bears, all housed in outdoor enclosures near Fairbanks rather than living wild, were surgically implanted with telemetry devices recording electroencephalography through bone-screw electrodes, along with eye movement, muscle activity, and core body temperature. This is the same combination of signals sleep researchers use to distinguish REM sleep, non-REM sleep, and wakefulness in humans, applied for the first time at this scale to bears through a full hibernation season.
The “five years” in the framing refers to the span over which this telemetry system operated across the bear colony, accumulating more than 3,500 days of recordings in total. The sleep analysis itself draws on a smaller set of recordings from each bear, typically three separate 24-hour windows capturing the peak of hibernation, a later point in hibernation, and a summer comparison period, rather than five unbroken years of continuous brain monitoring on every animal. That is a meaningful distinction for anyone picturing five years of nonstop recordings, though it does not undercut the reliability of the comparisons the study draws between hibernation and summer.
The specific sleep-stage figures that follow come from six of those sixteen bears, each contributing three sampled 24-hour recordings, rather than the full implanted colony.
The numbers behind “two-thirds”
During hibernation, the bears were asleep for approximately 64.9 percent of the time, split between 42.9 percent non-REM sleep and 22 percent REM sleep, with the remainder classified as wakefulness. In summer, total sleep dropped to about 31.7 percent, with REM sleep specifically falling to less than half its hibernation level. The doubling holds up closely across the data: bears slept roughly twice as much during hibernation as they did during their active season, and REM sleep rose by a similar margin.
Tøien has drawn a specific distinction about what fills the remaining hibernation hours that are not spent asleep. “Most of those awake hours are what we call quiet wake, which means they are not quite asleep but hardly moving at all,” he has said. That detail matters because it means hibernating bears are not simply awake in any ordinary sense during the time they are not sleeping.
Why more sleep during a metabolic slowdown is the interesting part
The metabolic side of hibernation is well established: black bears drop their metabolic rate to roughly a quarter of normal, without the deep drop in body temperature seen in smaller hibernators like ground squirrels, whose body temperature can approach freezing. Black bear body temperature falls only from around 38 degrees Celsius in summer to somewhere between 30 and 35 degrees during hibernation, a comparatively modest change that keeps the brain warm enough to remain active and readily arousable throughout the season.
That combination, a warm and functioning brain alongside a sharply reduced metabolism, is what makes the sleep finding worth attention rather than a foregone conclusion. A colder, more profoundly shut-down hibernator does not need to sleep in the conventional sense, because its brain is barely functioning regardless. A black bear’s brain stays capable of producing ordinary sleep architecture, and the data show it doing exactly that, for longer stretches than in the animal’s active season. The paper frames hibernation in this species as an extended period built from real, scoreable sleep mixed with quiet, low-movement wakefulness.
What the automated scoring adds
Part of what made an analysis at this scale possible was a machine-learning tool the team used to score sleep stages automatically, rather than relying solely on humans reviewing every recording by hand. Human scorers agreed with each other about 94.5 percent of the time on how to classify a given stretch of brain activity, and the automated tool matched that standard closely, with performance scores between 0.90 and 0.98 depending on the sleep stage being identified.
Tøien has said the team was “surprised how well the machine learning worked compared to manual scoring,” a result that matters less for what it says about bears than for what it suggests about scaling this kind of hibernation research to more animals and longer recording periods in future work.
What this study does not establish
These are captive research bears in Alaska, not a wild population, and the enclosure environment, food access before denning, and lack of real predation risk could all plausibly affect sleep patterns in ways a wild study might not replicate. The sleep-stage comparisons also rest on a handful of sampled days per bear rather than continuous coverage of an entire hibernation season, so shifts in sleep patterns across the full winter, rather than at the specific points sampled, remain unmeasured.
Tøien has described the current results as “an early snapshot from the coldest period during mid-hibernation,” and the team has signalled that more data collection is planned.
The natural next step is broader sampling across more of the hibernation season, and ideally in wild bears fitted with less invasive monitoring, to see whether the two-thirds figure holds steady from early denning through emergence in spring.
A comparison with other large hibernators using the same automated scoring method would also help establish whether black bears are unusual among hibernating mammals or whether warm-brained, sleep-rich hibernation turns out to be more common than the deep-torpor model has suggested.