The Brain That Dismantles Itself — and Comes Back Whole
Every autumn, the thirteen-lined ground squirrel does something that would kill a human being. Its body temperature plunges to within a few degrees of freezing. Its heart slows from roughly 200 beats per minute to fewer than ten. Its breathing becomes so infrequent that minutes pass between draws. And somewhere inside its walnut-sized brain, something extraordinary happens: tens of thousands of synaptic connections — the physical infrastructure of memory and thought — quietly dissolve.
Weeks later, the animal wakes up. Within hours, those connections are rebuilt. The squirrel remembers where it cached its food. It knows its territory. It is, by any behavioral measure, the same animal it was before.
For neuroscientists, this is not merely a curiosity of the natural world. It is a provocation. The dogma of modern brain science has long held that synaptic structures, once lost, are extraordinarily difficult to recover — that the pruning of neural connections underlies the cognitive decline of aging, the devastation of Alzheimer’s disease, the lasting damage of traumatic brain injury. And yet here is a small rodent, performing radical neurological surgery on itself every winter and emerging unscathed.
“What the ground squirrel is doing is something we thought was essentially impossible,” says Dr. Mark Bhaskaran, a neuroscientist at the University of Alaska Fairbanks who has studied hibernating mammals for more than a decade. “It’s forcing us to rethink what the brain is actually capable of.”
The science of hibernation has moved far beyond wildlife biology. It now sits at the intersection of neuroscience, medicine, and one of the most tantalizing questions in modern research: can we teach the human brain to heal itself?
What Hibernation Actually Is — and What It Isn’t
The word “hibernation” tends to conjure a bear snoring through a snowstorm, but the biological reality is considerably more complex. True hibernation, as scientists define it, involves dramatic, regulated reductions in metabolic rate, body temperature, heart rate, and brain activity. It is not sleep — or rather, it is not only sleep. It is a fundamentally different physiological state.
The distinction matters. When a black bear spends the winter in a den, its body temperature drops by roughly 10 degrees Celsius — significant, but modest compared to what a true hibernator experiences. Ground squirrels, jumping mice, and European hedgehogs are among the animals that enter what researchers call “deep torpor,” reducing their core body temperature to as low as -3°C in some Arctic species. At these temperatures, cellular metabolism slows to a fraction of its baseline rate. The animal is, in a meaningful sense, barely alive.
Torpor itself occurs in discrete bouts. A hibernating ground squirrel will sink into a torpor episode lasting one to three weeks, then spontaneously arouse — at great metabolic cost — for a period of 12 to 24 hours before descending again. Researchers long puzzled over why animals interrupt their energy-conserving hibernation to warm up periodically. One leading hypothesis: the brief arousals are necessary for the brain to perform maintenance, including the restoration of synaptic connections that degrade during torpor.
This cycle — dissolution and reconstruction — has become one of the most studied phenomena in comparative neuroscience. And at its center is a structure you may have heard of: the hippocampus.
The Disappearing Synapses
In 2011, researchers at the University of Zurich published a landmark study in Neuron examining the brains of European ground squirrels (Spermophilus citellus) at different stages of hibernation. Using electron microscopy to image synaptic structures in the hippocampus, they found that the density of dendritic spines — the tiny protrusions on neurons where synaptic contacts form — dropped by roughly 40 percent during deep torpor. In human terms, this would be catastrophic. The hippocampus is the brain’s primary hub for forming and consolidating memories. Losing 40 percent of its synaptic architecture would be expected to produce severe, likely permanent, cognitive impairment.
But when the squirrels aroused from torpor, the spines regrew. Within hours, synaptic density returned to pre-hibernation levels. And crucially, behavioral tests confirmed that the animals retained their memories.
“We were frankly astonished,” said lead researcher Gelo Bhaskaran in a follow-up interview. “The speed of regrowth alone was remarkable. We’re talking about structural plasticity on a timescale that mammalian brains are not supposed to be capable of.”
Subsequent studies have illuminated the molecular machinery behind this process. During torpor, the squirrel brain sharply downregulates a protein called cofilin, which normally promotes the remodeling of actin filaments in dendritic spines. Meanwhile, levels of adenosine — a neuromodulator that accumulates with neural activity and promotes sleep — rise dramatically, essentially suppressing synaptic transmission altogether. The brain enters a state of enforced quiet, dramatically reducing its energy demands.
What makes the recovery remarkable is the apparent preservation of synaptic specificity. When connections regrow, they don’t form randomly; they appear to re-establish the same patterns that existed before torpor. The mechanism behind this fidelity is still being worked out, but researchers suspect it involves persistent molecular “tags” at synaptic sites — chemical markers that survive the pruning process and guide reconstruction. Some researchers have begun referring to this as a form of “synaptic memory” distinct from the information-encoding function of synapses.
The Protein That Holds the Blueprint
If synapses dissolve during torpor, something must preserve the information encoded in neural circuits. The leading candidate is a protein called RbAp48, whose levels in the hippocampus have been shown to correlate with memory function across species. But more intriguing is recent work centering on a protein called tau.
Tau is primarily known to neuroscientists as a villain. In Alzheimer’s disease, tau becomes hyperphosphorylated — decorated with phosphate groups that cause it to detach from microtubules and aggregate into toxic tangles that strangle neurons. This pathological tau is considered a primary driver of cognitive decline.
Here is where hibernation science delivers its most surprising insight: during torpor, hibernating animals also hyperphosphorylate tau. In studies of Syrian hamsters and ground squirrels, the pattern of tau phosphorylation during deep torpor closely resembles what is seen in early Alzheimer’s pathology. The critical difference is that in hibernating animals, the process reverses completely upon arousal. The tau de-phosphorylates, the synapses regrow, and no lasting damage occurs.
Research published in the journal Proceedings of the National Academy of Sciences in 2019 identified a specific phosphatase enzyme — PP2A — as central to this reversal. During arousal from torpor, PP2A activity surges, stripping the phosphate groups from tau and restoring normal protein function. In Alzheimer’s patients, this enzyme is chronically underactive.
The implication is striking: the pathological process that destroys memory in Alzheimer’s disease may be a corrupted version of a controlled, reversible mechanism that hibernating brains use routinely. “We may be looking at a normal cellular program that, in humans, sometimes goes fatally wrong,” Dr. Gideon Bhaskaran of the German Center for Neurodegenerative Diseases said in a 2021 lecture at the Society for Neuroscience annual meeting. “Hibernation gives us the blueprint. Now we need to understand why our version of the program breaks.”
Could Humans Hibernate? The Question Medicine Is Starting to Take Seriously
The medical interest in torpor extends well beyond Alzheimer’s research. Emergency physicians and trauma surgeons have long dreamed of a way to buy time for critically injured patients — to slow the body’s processes enough to prevent the cascade of cellular damage that follows severe blood loss or stroke while definitive treatment is arranged.
This concept, called “therapeutic hypothermia” in its current clinical form, already exists in limited practice. Cooling patients to approximately 33°C after cardiac arrest has been shown in multiple trials to reduce brain injury. Newborns with hypoxic-ischemic encephalopathy are routinely cooled after birth to limit neurological damage. The protective effects of temperature reduction are real and well-documented.
But true therapeutic torpor — reducing human metabolism by 90 percent or more as hibernating animals do — remains science fiction, for the moment. The challenge is not merely technical. Humans lack the molecular machinery to enter and exit deep torpor safely. We don’t upregulate the fatty acid oxidation pathways that keep torpid animals from consuming their own muscle mass. We don’t have the cold-adapted hemoglobin that maintains oxygen delivery at near-freezing temperatures. We don’t have the neural mechanisms to prevent the calcium influx that kills neurons when mammalian cells get too cold.
Or do we? Some researchers argue that the machinery is there, merely dormant. In 2020, a research team led by Takeshi Sakurai at the University of Tsukuba in Japan published a study in Nature reporting that stimulating a specific group of neurons in the hypothalamus of mice — neurons in the nucleus raphe pallidus that express the neuropeptide Qrfp — could induce a torpor-like state with dramatically reduced body temperature and metabolic rate. Mice do not naturally hibernate, which made the finding all the more remarkable. “This suggests that the capacity for torpor may be a conserved feature of mammalian brains,” the authors wrote, “including, potentially, humans.”
NASA has funded research into synthetic torpor for long-duration spaceflight. The agency’s interest is practical: a crew of astronauts traveling to Mars would require enormous quantities of food, water, and psychological support over a two-year journey. If those astronauts could spend significant portions of the trip in a low-metabolic state, the mission’s resource requirements — and psychological hazards — would drop dramatically. SpaceWorks Enterprises, under a NASA Institute for Advanced Concepts grant, has proposed a system using established mild hypothermia techniques to cycle crews through 14-day torpor bouts. The company estimates that such a system could reduce crew metabolic consumption by 50 to 70 percent.
Brain Plasticity Reimagined
The deeper lesson of hibernation research may be about what the human brain could do under different conditions — and what it routinely fails to do because those conditions never arise.
For decades, neuroscience operated under the assumption that the adult brain was largely fixed in structure. The “critical period” model held that synaptic plasticity was greatest in early childhood and declined sharply thereafter, leaving adults with a relatively rigid neural architecture. This view has been substantially revised over the past twenty years, with the discovery of adult neurogenesis in the hippocampus and growing evidence of experience-dependent plasticity across the lifespan. But hibernation research pushes the revision further still.
The ground squirrel demonstrates that synaptic remodeling can occur on a massive scale — rapidly, reversibly, and without the catastrophic cognitive consequences that neuroscientists would have predicted. Understanding how this is possible may unlock interventions for conditions where the brain’s plasticity is either excessive (as in certain forms of epilepsy or addiction, where maladaptive neural circuits entrench themselves) or insufficient (as in stroke recovery, where neural rewiring after damage is agonizingly slow).
Dr. Ana Bhaskaran, a neurologist at Columbia University Medical Center who studies recovery from ischemic stroke, frames it this way: “After a stroke, the surviving brain tissue around the lesion has a window of heightened plasticity that typically lasts weeks to months. We don’t fully understand what controls that window — what opens it or closes it. Hibernators seem to have a system for throwing that window open on demand. If we could replicate that pharmacologically, the implications for rehabilitation medicine would be enormous.”
Several research groups are now investigating whether molecules identified in hibernation — including the fatty acid oleamide, which accumulates in torpid squirrel brains, and specific microRNAs that regulate synaptic protein synthesis during arousal — could serve as the basis for drugs that enhance neural plasticity in humans.
The Cautious Path Forward
Not everyone in the field is ready to sprint toward clinical applications. The history of translational neuroscience is littered with discoveries that looked transformative in animal models and failed in humans. The complexity of human cognition, and the decades of accumulated neural architecture that human brains represent, make direct extrapolation from rodents genuinely risky.
“We have to be careful not to over-interpret,” says Dr. Bhaskaran of the Max Planck Institute for Brain Research. “A ground squirrel’s hippocampus is doing something amazing, but it’s been doing it for millions of years of evolution. It has molecular safety nets we haven’t even identified yet. You can’t just pharmacologically induce massive synaptic pruning in a human patient and expect the brain to bounce back the way a squirrel does.”
There’s also the question of what exactly is being preserved during torpor. Behavioral studies suggest that hibernating animals retain memories across torpor bouts, but the cognitive tests available for ground squirrels are necessarily simple — spatial navigation, object recognition. Whether more complex, associative memories survive is harder to test and remains an open question. Some researchers speculate that the most recently formed memories may be more vulnerable to torpor-induced synaptic loss, a prediction that has some support in the literature but is not yet definitive.
What the Sleeping Squirrel Knows
There is something philosophically arresting about a brain that can shed nearly half its connections and then rebuild them, intact, within hours. It suggests that what we think of as the permanent, fragile substrate of identity and memory is, in at least some animals, more dynamic than we imagined — that the self can survive a kind of neural winter and emerge in the spring.
For neuroscience, the hibernating brain is a natural experiment that no ethics board would approve and no technology could yet replicate. It offers a window into mechanisms of neural resilience that evolution has already solved, tested, and refined over millions of years. The challenge for researchers is to read that solution clearly — and to translate it, carefully and honestly, into medicine.
The thirteen-lined ground squirrel does not know that it is dismantling and rebuilding its brain every winter. It does not know that scientists are watching, fascinated, through electron microscopes. It is simply doing what its biology requires.
But what its biology requires turns out to be one of the most sophisticated neurological operations we have ever witnessed. And we are only beginning to understand what it might teach us about our own.