According to evidence from a preclinical study published in
Nature, the hippocampus, the brain's memory center, may also reorganize memories to anticipate future outcomes.
The study, conducted by a research team from McGill University's Brandon Lab and their collaborators at Harvard University, reveals a learning process that had never been directly observed before.
"The hippocampus is often described as an internal model of the world nestled in the brain," explains Mark Brandon, the study's lead author, an associate professor in McGill University's Department of Psychiatry and a researcher at the Douglas Research Centre. "We find that this model is not static; it is updated day after day, as the brain learns from prediction errors. As outcome predictability increases, hippocampal neurons fire earlier because they learn what is going to happen next."
A new view of learning in action
The hippocampus builds maps of physical space and past experiences that help us understand the world. Scientists know these maps change over time as brain activity patterns evolve, a phenomenon currently considered random.
The study's results instead show that these changes are not random, but structured. This is what the research team concluded by examining the brain activity of mice while they learned a task associated with a predictable reward.
"We found something surprising," explains the professor. "The peaks of neuronal activity, which initially occurred at the time of reward, began to occur earlier and earlier, eventually appearing before the mice obtained the reward."
Rather than using traditional electrodes, which only allow observation of neurons over short periods, the team used new imaging techniques that make active neurons glow. The Brandon Lab is one of the first in Canada to use this technology. The team can thus examine cells over several weeks and detect slow changes that often go unnoticed with classical methods.
Learning and Alzheimer's disease
Simpler forms of reward learning have long been associated with more primitive brain circuits, as shown by Ivan Pavlov's famous experiments where animals associated a signal, like the sound of a bell, with food. The study's findings suggest a more sophisticated version of this process, in which the hippocampus uses memory and context to anticipate outcomes.
Furthermore, individuals with Alzheimer's disease often have difficulty remembering the past, as well as learning from their experiences and making decisions. By showing that a healthy hippocampus transforms memories into predictions, the study provides a new framework that could help us understand why learning and decision-making are affected in the early stages of Alzheimer's disease, and opens new research perspectives on the failure and restoration of this predictive signal.
The study
The paper "Predictive Coding of Reward in the Hippocampus", by Mohammad Yaghoubi, Mark Brandon, et al., was published in
Nature. This study was funded by the Fonds de recherche du Québec - Santé and the Canadian Institutes of Health Research.