Researchers uncovered how the brain balances excitatory and inhibitory neurons to organize and process memories.
May 28, 2025
How brain networks balance learning and memory
FMI researchers have provided new insights into how the brain organizes and processes memories, thanks to a study that looks at the balance between excitatory and inhibitory neurons. Memory networks have been thought to rely on stable patterns of activity to represent information, but the new findings suggest that certain brain regions function more dynamically, using balanced interactions between excitatory and inhibitory neurons for memory storage.
A big question in neuroscience is how the brain manages to store and process memories in a stable yet flexible way, especially when dealing with complex or overlapping information. Excitatory (E) and inhibitory (I) neurons are known to work together in so-called E/I assemblies, but how these assemblies stabilize memory storage and facilitate learning remained unclear.
To address this question, a team of FMI neuroscientists developed a computational model based on the activity of a brain region in zebrafish that is involved in odor memory. This part of the brain is similar to regions in mammals that are thought to build abstract mental maps of the world. Unlike previous models, which missed a precise balance between signals that excite and those that inhibit brain activity, their model captures key features of how the brain actually works, including activity that seems random at first but turns out to be stable and consistent when looked at more closely.
“By fine-tuning the balance between excitatory and inhibitory neurons, we can create a network that is both stable and flexible, allowing it to process sensory information,” says study co-author Friedemann Zenke, a group leader at the FMI.
The researchers further explored how these E/I assemblies influence tasks such as classifying odors. The findings reveal that networks with finely tuned E/I interactions performed better than networks with assemblies that are not balanced. Networks with a balanced E/I ratio also showed greater stability when new memories were introduced. “We’ve shown that these networks can manage complex tasks without becoming unstable, even when overlapping memories are added,” says study co-author Rainer Friedrich, a group leader at the FMI. “This suggests that E/I assemblies not only help store memories but can also support learning.”
The work contributes to our understanding of how the brain processes and stores sensory information, and it raises important questions for future studies, particularly in understanding how these principles may apply to other brain regions involved in memory and learning.
Original publication:
Claire Meissner-Bernard, Friedemann Zenke & Rainer W Friedrich Geometry and dynamics of representations in a precisely balanced memory network related to olfactory cortex eLife (2025) 13:RP96303.
Researchers uncovered how the brain balances excitatory and inhibitory neurons to organize and process memories.
About the first author
Originally from Strasbourg, France, Claire Meissner-Bernard completed her PhD at the Collège de France in Paris, where she used experimental approaches to investigate how the brain stores and retrieves information. During her PhD, she visited EPFL to work on a project in computational neuroscience — a field she has continued to pursue as a postdoctoral fellow in the Friedrich lab at the FMI. From 2019 to 2021, she co-organized the Computational Neuroscience Initiative Basel. Passionate about science communication, Claire enjoys participating in outreach activities. Outside the lab, she is an amateur pianist and regularly takes part in music contests and festivals.
