Understanding the storage model for human memory will enable a range of medical technologies relevant to repair and augmentation of the brain, but at present there is only a general consensus on the online of that model. Researchers believe that the data of memory is stored in the architecture of synapses, and here researchers provide more evidence that synapses are indeed the relevant location in mammals. This is perhaps of interest due to recent research in lower animals that seemed to rule out synaptic structures as the location of memory.
Our memories are as fleeting as the brain structures that store them, or so the theory goes. When the connections - called synapses - between neurons break, the memories they hold are thought to evaporate along with them. The idea seemed good, but has been hard to test. Now a team has taken on the challenge, studying a brain region called the hippocampus, which stores "episodic" memories. These are the memories of events or conversations that might be forgotten over time if the memories aren't used. The challenge to studying synapses in this region is that the hippocampus is so deep and the connections so densely packed that no microscope could easily monitor the synapses' longevity.
When mice experience a new episode or learn a new task that requires spatial navigation, the memory is stored for about a month in a structure at the center of the brain called the hippocampus (it is stored slightly longer in people). If mice have hippocampus-disrupting surgery within a month of forming a memory - a memory of meeting a new cage-mate or navigating a maze - that memory is lost. If the disruption occurs after more than a month, then the mouse still retains the memory of a new friend or location of food. That's because the memory had been relocated to a different region of the brain, the neocortex, and is no longer susceptible to disruption in the hippocampus.
In the past, scientists had monitored connections between neurons in the neocortex, nearer the brain's surface and therefore visible with little disruption to the brain. They watched not the connections themselves, but the bulbous projections called spines that form connections at their tips. Watching the spines come and go serves as a proxy for knowing when excitatory connections between neurons are created and broken. Those scientists found that about half of the spines in the neocortex were permanent and the rest turned over approximately every five to 15 days. "The interpretation was that about half the spines in the neocortex are long-term repositories for memories while others retain malleability for new memories or forgetting."
If the same line of thinking held true for the hippocampus as it did for the neocortex, spines in the hippocampus should turn over roughly every 30 days along with the memories they hold. Verifying that idea had been challenging, however, because the hippocampus is deeply buried in the brain and the spines in that region are so densely packed that multiple spines can appear to merge into one. The team overcame that problem with new techniques that allow stable imaging of a single neuron in a living mouse over long time periods, an optical needle, called a microendoscope, that provides high-resolution images of structures deep within the brain, and a mathematical model that took into account the limitations of the optical resolution and how that would affect the image datasets depicting the appearances and disappearances of spines. The researchers found that the region of the hippocampus that stores episodic memories contains spines that all turn over every three to six weeks - roughly the duration of episodic memory in mice.