How the brain switches between older and newer memories

 

As humans and other animals experience new things, their brains continuously update their memory of past events. These updates allow them to adapt to changing environments, all while preserving older memories that could still help them to make decisions in some situations.

Many past neuroscience studies have investigated the neural circuits involved in the encoding and retrieval of memories. However, the mechanisms via which it decides whether to retrieve older or newly updated memories remain poorly understood.

Researchers at Korea Advanced Institute of Science and Technology (KAIST) recently carried out a study involving mice that was aimed at better understanding how the brain switches between older and newer memories.

Their paper, published in Nature Neuroscience, delineates a specific neural pathway that appears to support the flexible switching between the retrieval of old and updated memories in mice, involving brain regions known as the medial septum (MS), the medial entorhinal cortex (MEC) and the CA1 segment in the hippocampus.

"We have been studying a group of cells called engram cells, in which physical memory traces are thought to be located," Jin-Hee Han, senior author of the paper, told Medical Xpress.

"Previously, we found that repeated learning recruits a new set of cells to encode updated memory (strengthening) instead of using the same set as commonly thought.

"Despite this reshuffling of engram cell ensembles during memory updating, artificially activating the old set can still induce memory recall, indicating that it retains memory traces, even though it is not reactivated by natural environmental cues."

The researchers' earlier observations ultimately led them to hypothesize that individual events leave their own "memory traces" each time the brain updates memories. They also proposed that the brain can access previously formed engram cell ensembles (i.e., cells encoding specific memories) that remain unaltered during the updating process, but it predominantly retrieves memories encoded by newly formed engram cell ensembles.

"This idea strongly motivated us to search for neural mechanisms that may control shifts and selections between new and old memories during memory retrieval," said Han.

A neural mechanism responsible for flexible memory retrieval

To test their hypotheses, Han and his colleagues carried out a series of experiments involving adult male mice. The mice were trained to form specific associations between stimuli and rewards, which ultimately guided their behavior.

The researchers then exposed the mice to new experiences that led them to update their original memory associations. Using neuroimaging techniques, they looked at what cells became active while the mice's brain was retrieving newer, updated memories.

"Our approach was to investigate whether neurons are selectively activated during retrieval after memory updating, compared with a single learning episode as a non-memory-updating control," explained Han.

"To do this, we mapped active cells throughout the brain using c-Fos protein expression as a proxy for neuronal activity. c-Fos is an immediate early gene whose expression is rapidly induced upon neuronal activation. From this whole-brain screening, we identified a specific subset of GABAergic neurons in the MS that project to the MEC as a candidate."

The MS and MEC are known to play a key role in the generation of brain rhythms, the encoding of memories and spatial navigation. The team found that when they inactivated projections from neurons in the MS to those in the MEC using optogenetic tools, the mice reverted their behaviors to those they used to perform before their memories were updated.

"This reversal was not merely behavioral," said Han. "When we examined neuronal activity patterns in behaving animals, those in the hippocampal CA1 region, a key memory hub, also shifted back to those associated with the earlier memory version.

"The activity of MS-MEC GABAergic neurons follows brain states (online versus offline), predominantly active during the online state, while less active during the offline state. Consistently, the duration of online state correlated with memory performance."

Interestingly, the researchers found that when this "online" activity lasted longer after the animals' memories were updated, mice performed better on memory tasks. This suggests that they uncovered a neural signature of effective memory updating and retrieval.

"Our findings reveal that the MS-MEC GABAergic pathway functions as a neural switch that organizes retrieval between new and old memories, thereby enabling memory updating," said Han.

Towards a richer understanding of memory processes

Han and his colleagues identified a neural pathway that appears to play a central role in the adaptive switching between older memories and newer, updated ones.

Future studies could further explore this newly uncovered mechanism and investigate its possible role in disorders associated with memory loss.

"I remember the day Mujun, a postdoc in my lab and the first author of the paper, told me that when he inactivated the MS GABAergic projections to the medial entorhinal cortex, the animals' behavior reverted to pre-update patterns rather than simply impairing memory retrieval," said Han.

"I was super excited and I find this flexible switching between the new and old patterns the most intriguing finding, raising many questions. It suggests that this GABAergic-dependent selection mechanism contributes to organizing episodic memories in time in the brain."

Overall, the team's results suggest that the retrieval of memories does not entail the passive reactivation of stored memory traces, but it instead involves an active selection between various available memory traces. This selection process appears to be controlled by a GABAergic neural pathway connecting the MS and the MEC.

Notably, the inability to update older memories and form new memories is a characteristic of various neurodegenerative diseases, psychiatric disorders and neurodevelopmental conditions, including Alzheimer's disease (AD), schizophrenia and, in some cases, autism spectrum disorder (ASD).

Further research could thus explore the possible contribution of the neural pathway identified by the researchers in the memory deficits associated with these conditions.

"Our findings provide insights into the potential neural basis of these memory disorders and could contribute to the development of interventions to treat them," added Han.

"We are now trying to understand the detailed circuit mechanisms underlying memory switching control by medial septum GABAergic neurons and to identify the cells and connections involved in this process. Another important question is how the engram architecture is reorganized by memory updating.

"Finally, it will be important to study this mechanism in disease models such as AD, schizophrenia, and autism, and also to investigate whether the same mechanism is conserved in human brains."