A new theory suggests that a memory's utility for future situations determines its location in the brain, either in the hippocampus or neocortex. This challenges the traditional view, which emphasizes that the consolidation of memory in the neocortex is based on its generality, rather than age.
According to a new theory proposed by researchers at the HHMI Janelia Research Campus and their colleagues at University College London, how useful a memory is for future situations determines where it is located in the brain.
The theory offers a new way to understand systems integration, a process that moves certain memories from the hippocampus (where memories are initially stored) to the neocortex (where memories reside long-term).
According to the classic view of systems consolidation, all memories are transferred from the hippocampus to the neocortex over time. But this idea doesn't always hold water; research shows that some memories stay permanently in the hippocampus and never transfer to the neocortex.
In recent years, psychologists have proposed a number of theories to explain this more complex view of systems consolidation, but no one has yet mathematically figured out what determines whether a memory stays in the hippocampus or is consolidated in the neocortex.
Now, Janelia researchers have proposed a new, quantitative view of systemic consolidation to help address this long-standing problem, proposing a mathematical neural network theory that memories consolidate to the neocortex only if they improve their generalization capabilities.
Generalization is built from reliable and predictable components of memory, allowing us to apply it to other situations. We can generalize certain features of memory to help us understand the world, such as a canyon that predicts the presence of water.
This is different from episodic memories, which are detailed recollections of the past that have unique characteristics, such as a single memory of us hiking into a certain canyon and encountering a body of water.
According to this view, memory consolidation is not about copying a memory from one area of the brain to another, but about creating a new memory that is a generalization of a previous memory. The degree to which a memory is generalized—not age—determines whether a memory is consolidated or retained in the hippocampus.
Using neural networks, the researchers showed how the amount of consolidation changes depending on how generalizable the memory is. They were able to reproduce previous experimental patterns that could not be explained by the classical view of system consolidation.
The next step is to test this theory experimentally to see if it predicts the extent of memory consolidation. Another important direction is to test the authors' model of how the brain distinguishes between predictable and unpredictable components of memory to mediate memory consolidation. Revealing how memory works could help researchers better understand a component of cognition, leading to potential benefits for human health and artificial intelligence.