0
0
Researchers from MIT have unveiled a groundbreaking model explaining how the brain’s hippocampus encodes not only spatial memories of places but also episodic memories of events, such as personal experiences. This discovery could revolutionize our understanding of how the brain links different types of memories.
Nearly 50 years ago, scientists identified specific cells in the hippocampus that are crucial for remembering locations, known as place cells. These place cells were thought to be key players in spatial memory, helping us navigate through physical spaces. However, the role these cells play in storing episodic memories—the vivid recollections of personal experiences—remained elusive until now.
The new model developed by MIT researchers offers a solution. It suggests that place cells, alongside grid cells found in the entorhinal cortex, act as a scaffold to anchor memories in the form of a linked sequence. This allows memories to be stored and retrieved in an organized manner, even when the memory has no spatial component.
Ila Fiete, a professor of brain and cognitive sciences at MIT and senior author of the study, describes this model as a foundational step in understanding episodic memory. “This model is a first-draft model of the entorhinal-hippocampal episodic memory circuit. It’s a foundation to build on to understand the nature of episodic memory. That’s the thing I’m really excited about,” Fiete said.
The model, published in Nature, shows that the hippocampus and the entorhinal cortex, which are known for their role in spatial memory, also support the storage of episodic memories by encoding them in a structured way, much like a memory “index.” Lead authors of the study, Sarthak Chandra and Sugandha Sharma, along with Rishidev Chaudhuri from the University of California at Davis, contributed to the paper’s findings.
A New Model for Memory Encoding
Place cells and grid cells, two types of neurons in the hippocampus and entorhinal cortex, play a crucial role in memory. While place cells are involved in remembering locations, grid cells fire in a regular geometric pattern that helps form a lattice of space representations. Together, they form a system that allows us to both navigate and remember spatial locations.
However, it’s not just about places. Episodic memories—like recalling a birthday party or a family vacation—are also encoded in these circuits, although they don’t always have a direct spatial component. The new study proposes that grid cells and place cells are integral to both types of memory, creating a unified network that supports both spatial and episodic recollections.
Rather than thinking of these cells as storing specific content, the researchers describe them as forming an “abstract scaffold” that points to where the details of a memory are stored in the sensory cortex. This allows for the pattern completion of fragmented memories, connecting partial memories with their full, detailed content stored elsewhere in the brain.
Memory Cliffs and Memory Palaces
One of the most fascinating aspects of the model is how it addresses long-standing issues in memory research. Existing models of memory, like Hopfield networks, have often struggled to replicate the way the brain stores and recalls memories. In Hopfield models, once memory storage reaches capacity, the system begins to “forget” previous memories, creating a “memory cliff.” The new model, however, shows that the biological brain is much more nuanced, with older memories gradually fading while new ones are continually added.
This insight also sheds light on the phenomenon of “memory palaces,” a technique used in memory competitions. Participants often visualize a familiar location—like a childhood home—and associate specific pieces of information, such as cards in a deck, with certain spots in the house. The new model suggests that this technique exploits the brain’s natural tendency to use a spatial scaffold, built from long-term memories, to store and recall new information more efficiently.
Next Steps and Implications
Looking forward, the team plans to refine their model to explore how episodic memories could be converted into more general “semantic” memories, which are detached from their specific contexts (like the fact that Paris is the capital of France). This research could also have implications for understanding how memory models could be applied in artificial intelligence, with brain-like memory systems integrated into machine learning models.
For now, Fiete and her team are excited about the potential of this new model to explain how the brain processes and stores memories across both space and time, offering new insights into the very nature of human memory.
The study, Episodic and Associative Memory from Spatial Scaffolds in the Hippocampus, was published in Nature on January 16, 2025.
