Neuroscientists in Japan have discovered that putting mice into an artificial state of hibernation eradicates more than half of their neural synapses, yet leaves their long-term memories completely intact. The research, published recently in Science, challenges long-held assumptions about how the brain stores information, revealing an underlying architecture that can rebuild itself after a massive structural purge. The findings were detailed by Kazumasa Tanaka of the Okinawa Institute of Science and Technology Graduate University and Takeshi Sakurai of the University of Tsukuba. While hibernation is commonly associated with species like bears, hamsters, and squirrels, the neural circuitry that triggers the state is conserved across mammalian species, including those that never hibernate in the wild. In 2020, Sakurai's team developed a technique to artificially activate this dormant circuit in mice. By stimulating a specific population known as Q neurons within the hypothalamus, the researchers induced a physiological state called Q-neuron-induced hypothermia and hypometabolism (QIH). During QIH, the biological processes of the mice slow down drastically. The artificial hibernation brings the animals' body temperatures down to approximately 20 degrees Celsius, accompanied by significant decreases in both heart and breathing rates. Tanaka notes that this state sits somewhere in the middle of the natural hibernation spectrum; bears maintain a normal body temperature of around 37 degrees Celsius while reducing metabolic demand, whereas some squirrel species endure near-freezing temperatures. The primary advantage of QIH for researchers is that it can be controlled and switched on or off at will. In the recent experiments, the team placed mice into the QIH state for a period of 48 hours before waking them up. To monitor the effects on the brain, researchers implanted bundles of fine electrodes, known as tetrodes, into the hippocampus to record individual neuron activity in freely moving mice. Once the hibernation state set in, neural activity dropped by about 70 percent. Furthermore, imaging via serial block-face scanning electron microscopy revealed that the hibernation had eradicated more than half of the synapses in the brain tissue—a massive loss of connections that Tanaka expected would severely impair the animals' memory. However, behavioral tests proved the opposite. Prior to the hibernation phase, the mice were trained in two standard memory tasks heavily reliant on the hippocampus: contextual fear conditioning, where they learned to associate a specific box with a mild electric shock, and a plus-maze navigation task requiring them to find a reward. When the mice were roused from their 48-hour hibernation, they performed both tasks just as effectively as a control group that had not been put to sleep. Brain activity recordings confirmed this retention, showing that hippocampal "place cells" fired in the exact same spatial locations as they had before the experiment. To understand how memories survived the massive synaptic loss, the team tracked specific dendrites over an eight-day period. They found that after the mice woke up, the lost synapses regenerated, with 82 percent reappearing in the exact same location on the same dendrite. Using a specialized technique called eGRASP, which makes tagged learning connections glow green, the researchers isolated "engram synapses" responsible for memory storage. They discovered that while isolated synapses were destroyed during hibernation, engram synapses arranged in tight spatial clusters were highly protected and preserved. Investigating the architecture of these surviving clusters, the team found that a third of them were attached to rare structures called multisynaptic boutons, where a single presynaptic terminal connects with multiple postsynaptic spines. To verify that this clustering was directly related to memory retention, the team ran a negative control using long-term anesthesia combined with cytochalasin D, a drug that prevents synaptic stabilization. This combination caused a similar loss of synapses but resulted in impaired fear memory, indiscriminately destroying the clustered engram synapses just like any other connection. The researchers acknowledge that the study demonstrates an association rather than direct causality, as they currently lack the tools to selectively disable the clusters without damaging the surrounding network. However, unpublished data from the lab suggests an even broader implication: the brain may return to a "factory default" state after hibernation. In mice engineered to develop epilepsy, inducing brief artificial hibernation completely suppressed the onset of seizures. If these findings are confirmed in subsequent papers, they could fundamentally alter our understanding of brain recovery. While these neurological findings offer new insights into memory retention and brain recovery, the research remains confined to animal models in Japanese laboratories. Families across Somalia and the diaspora coping with memory-affecting conditions or epilepsy face a long road before any human clinical trials, meaning the study currently provides foundational laboratory science rather than immediate medical treatments.