Scientists unlock the hidden workings of memory using neurofeedback in rats

A new study published in Neuron has shown that rats, like humans, can intentionally recall specific memories without being prompted by external cues or needing to act on those memories. Researchers developed a neurofeedback system that trained rats to mentally “jump” to remembered locations in a maze, revealing that memory retrieval can be studied as an internal brain process distinct from cues or behaviors. This breakthrough allows scientists to investigate how the brain recalls information without relying on physical movement or sensory triggers.

Humans often remember people, places, or events spontaneously and without needing to do anything in response. For instance, we might think about a past vacation or choose where to go for dinner based on past experiences. These kinds of memory retrievals happen without specific prompts or observable actions.

In contrast, memory studies in animals typically rely on external cues and physical behavior to assess whether a memory has been recalled. For example, rats may be tested on whether they can find a reward in a maze or react to a previously learned signal. The challenge with these methods is that it becomes hard to separate memory retrieval itself from the sensory information, decision-making, or behavior involved.

“We all know that memories can be retrieved without directly causing an action, as when we reminisce about the past. Despite that, virtually all studies in animals assess memory retrieval based on whether an action is triggered,” said study author Loren Frank, a professor at UC San Francisco and investigator at the Howard Hughes Medical Institute.

“Similarly, most studies of memory use specific cues to cause memories to be retrieved, but we know that sometimes memories just bubble up without a clear cue. As a result, we did not have a way to directly investigate memories separate from the actions they cause or the cues that drive them. Our first goal was to see if animals could spontaneously retrieve specific memories.”

The research team focused on the hippocampus, a brain region known for storing spatial memories, and particularly on “place cells”—neurons that become active when an animal is in or thinks about a specific location. These cells not only represent where the animal currently is but can also reflect places the animal has been in the past.

The researchers developed a brain-machine interface that used real-time recordings from the rats’ hippocampal neurons. Six rats were implanted with devices that could detect and interpret the activity of these place cells. During training, the rats explored a Y-shaped maze with rewards at the end of two arms. Their brain activity was recorded to create a map linking neural patterns to specific locations.

In a later phase, the lights on the maze arms were turned off and the rats remained near a central reward port. At this point, the system would provide a reward only when the rat’s brain activity indicated it was mentally representing a remote target location—one of the maze arms—even though it was physically nowhere near it.

Importantly, the rats could not see the target location from where they were. Nor were they given any cues to remind them of it. Instead, they had to generate the representation of that place using only internal memory. When the system detected that their hippocampal activity matched the pattern associated with the target location, a tone played, and if the rat responded quickly, it received a reward. Over multiple sessions, the rats learned to reliably produce these remote memory representations, proving that they could intentionally recall specific places in their minds.

In most cases, the rats’ brain activity skipped directly from their current location to the target, rather than moving through intermediate points in space. This is similar to how people can mentally “teleport” to a memory without reliving every moment in between. The neural activity was specific: it typically reflected only the goal location and not the path to it. Moreover, the rats could switch between different target locations when the researchers changed the memory required for the reward, showing that the animals had flexible control over what they remembered.

Frank and his colleagues also found that these memory recall events involved groups of neurons working together. The researchers identified distinct “cell assemblies” that became active when a rat generated a remote memory. These coordinated patterns of firing mirrored the activity seen when the rat physically visited the location earlier. Even when the rats were not moving or not receiving a reward, these memory-linked assemblies were still active, suggesting a spontaneous and internal process of recall.

Interestingly, most of these remote memory events happened while the rats were still but not during a well-known brain state called a sharp wave ripple, which is often linked to memory consolidation. Instead, they occurred during quieter periods without strong external brain signals. This finding suggests that memory retrieval may operate through brain mechanisms that are different from those typically associated with storing or consolidating memories.

In some cases, the recalled representations occurred while the rats were moving, and in those instances, the activity aligned with a specific rhythm in the brain called the theta rhythm, which has been tied to memory-guided planning and decision-making.

“We expected to see that these events occurred during a specific hippocampal activity pattern called a sharp-wave ripple (SWR) where we and others had hypothesized that memory retrieval took place,” Frank told PsyPost. “This was not the case, and our results suggest that while reactivation of past experience can happen during SWRs, those are not the times of intentional memory retrieval in our animals.”

By training animals to retrieve memories without external prompts or visible cues, the study offers a powerful new approach to studying how the brain accesses past experiences. Unlike prior research where memory could only be studied indirectly through behavior or cue-based retrieval, this method allows researchers to observe memory recall as a distinct and intentional brain process.

One of the most important implications of this work is that it creates a model for understanding memory in ways that are closer to human cognition. It shows that animals can recall past experiences based on internal motivation, not just sensory input or habit. This opens new doors for exploring how memory works in the brain, how memories are used to make decisions, and what goes wrong in disorders where memory is impaired.

“The main result is that our rats, like people, can intentionally activate representations of places they have visited (e.g., memories), which means we can study how it is that the brain retrieves memories and how they could be used to guide decisions, as when we deliberate over where to go to dinner based on past experiences,” Frank explained.

However, the study has some limitations. The findings are based on a small number of rats and a specific type of spatial memory. It remains to be seen whether similar patterns would emerge for other types of memories or in different brain regions. Additionally, while the researchers could infer that the rats were recalling a location, it’s impossible to fully know the subjective experience of the animal. Future research will likely explore how these findings relate to other forms of memory, including those that involve emotion, time, or complex associations.

By identifying the neural signatures of internally driven memory recall, the researchers hope to eventually understand how the brain supports imagination, planning, and decision-making—all of which rely on the ability to retrieve past experiences.

“Our long term goal is to understand how the brain can generate and use memories,” Frank told PsyPost. “More broadly, we want to understand how the brain can think, where internally generated patterns, typically based on past experience (e.g., memories) are used to understand the world and make decisions. By isolating the patterns of brain activity that support intentional memory retrieval, we are one step closer to that goal.”

The study, “Closed-loop modulation of remote hippocampal representations with neurofeedback,” was authored by Michael E. Coulter, Anna K. Gillespie, Joshua Chu, Eric L. Denovellis, Trevor Thai K. Nguyen, Daniel F. Liu, Katherine Wadhwani, Baibhav Sharma, Kevin Wang, Xinyi Deng ,Uri T. Eden ,Caleb Kemere, and Loren M. Frank.