Scientists at Stanford’s Bio-X Center have proven that memories last as long as the synapses that store them. Until now, this concept has been a long-standing unproven theory in the scientific community.

The team of Stanford researchers tested the theory by studying the duration of episodic memories in mice. Episodic memories are memories of specific autobiographical episodes that can be explicitly stated using a time and location. These memories include events such as remembering the first day at a new job or the date and time of your birth. They are stored in a region of the brain called the hippocampus for about a month. Researchers observed that, if there was damage done to the hippocampus within 30 days of collecting episodic memories, the memories disappeared. However, if the hippocampus was damaged after 30 days, the memories remained because the brain relocated the them to another part of the brain called the neocortex.

“The thought is that memories are gradually moved around the brain,” Mark Schnitzer, a faculty member at the Stanford Neurosciences Institute and an associate professor of biology and applied physics, said in an interview for Stanford News. “The neocortex is a long-term repository, whereas considerable evidence indicates that memories stay in the mouse hippocampus only about a month.”

The hippocampus is located deep within the brain making it difficult to study, whereas the neocortex is located closer to the surface of the brain and can be visibly studied without disturbing the brain itself.

Studies from other institutes have been valuable for Schnitzer and his team. Researchers at Cold Spring Harbor Laboratory in New York and elsewhere observed the spines in the neocortex, or the projections that form at the tip of neural connections. They recorded their behavior to identify when connections between neurons were broken or created. It was shown that about half of neocortex spines remained while the other half were constantly turned over.

“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,” Schnitzer said.

Schintzer hypothesized that if the spines in the hippocampus are undergoing the same process as the spines in the neocortex, they should be turning over, too. Initially, it was impossible to study individual spines because they are densely packed within the neurons of the hippocampus located deep in the brain.

Schnitzer and his team overcame these obstacles using a three-part solution. The first technique the scientists employed allowed them to create a stable image of a single neuron in a living mouse over a long period of time. Next, they used a microendoscope, an optical needle that produces high-resolution images of structures that are located deep inside the brain. They were still not able to distinguish a single spine from a small clump. The problem was finally solved with a third component: a mathematical model that analyzed how the limitations of the optical resolution would affect how images depicted the appearances and disappearances of spines.

The spines in the hippocampus of a mouse brain have a turnover frequency of three to six weeks — about the same lifespan of an episodic memory in mice. The results confirmed the idea about how the brain stores memories. The findings and technological advances associated with this study have created a huge potential for research.

“It opens the door to the next set of studies, such as memory storage in stress or disease models,” Schnitzer said.