Neuroscientists discover biological mechanism that helps the brain ignore irrelevant information

New research published in Nature Neuroscience sheds light on how the brain filters out distractions during memory formation. The study reveals that a type of synaptic plasticity—long-term changes in inhibitory connections between neurons—helps the brain suppress irrelevant sensory input during memory replay, a process believed to support long-term learning. This filtering mechanism allows the hippocampus to focus on the general structure of past experiences, rather than the noisy or unpredictable details.

The researchers wanted to understand how the brain distinguishes between useful and irrelevant information when storing memories. While it is well established that the hippocampus replays neural activity to consolidate memories, what actually gets replayed—and how the brain decides what to keep or discard—remains unclear. This study aimed to uncover the neural rules that shape the content of these replay events, particularly how distractions are suppressed while core patterns are preserved.

To investigate this, the research team used a combination of computational models and experimental methods in mice. First, they built three types of models—a basic spiking network, a detailed biophysical model of hippocampal neurons, and an abstract mathematical model. Each model simulated memory replay in the hippocampus, focusing on an area called CA3, which is known for generating sharp-wave ripples—brief bursts of brain activity linked to memory consolidation. These models were trained on sequences of experiences containing both relevant spatial patterns and irrelevant, randomly placed cues.

In each simulation, the models learned to represent the environment using a Hebbian learning rule. This rule strengthens connections between neurons that activate together, and in this study, it was applied not only to excitatory synapses but also to inhibitory ones—a departure from most previous models.

The team found that including inhibitory plasticity allowed the network to suppress the activity of neurons tuned to unpredictable, non-generalizable stimuli, while still reinforcing patterns that reflected stable aspects of the environment. In contrast, networks without inhibitory plasticity replayed noisy and disorganized sequences that did not resemble meaningful memories.

To test the predictions of their models in real brains, the researchers conducted optogenetic experiments in mice. They implanted tiny optical fibers and injected viruses to target specific neurons in the hippocampus with light-sensitive proteins. By pairing random sensory cues with artificial stimulation of a sparse group of neurons, they effectively created artificial “distractor cells.” These neurons became active during irrelevant stimuli.

Later, during periods of rest when memory replay typically occurs, the researchers measured whether these neurons were suppressed. As predicted, neurons artificially linked to unpredictable stimuli showed reduced activation during memory replay, consistent with the idea that the brain learns to inhibit irrelevant representations.

The experimental data also confirmed that this suppression was linked to increased inhibitory input onto these neurons. Cells that responded to random cues received stronger inhibitory signals than place cells, which are known to represent consistent features like spatial location. This supports the idea that inhibitory synapses adapt based on experience and help shape the content of memory replay in a way that supports generalization.

In both simulations and biological experiments, when inhibitory plasticity was removed or disrupted, the brain’s ability to distinguish signal from noise broke down. Replay became cluttered with irrelevant information, and the networks failed to maintain clean, structured representations. This suggests that inhibitory learning is not only important for filtering distractions but also essential for the formation of coherent memories.

The study also introduced a simplified model that abstracted real-world learning into a sequence of observations with both consistent patterns and random distractions. In this model, excitatory learning alone could not filter out noise. But when inhibition was allowed to adapt based on the timing of neural activity, the network learned to suppress unpredictable inputs over time. The more often a distractor occurred in different contexts, the more inhibition it accumulated, eventually blocking it from being replayed. This mechanism aligns with the idea that memory consolidation prioritizes generalizable information—patterns that remain stable across multiple experiences.

While the study provides strong support for the role of inhibitory plasticity in memory consolidation, it also has limitations. The researchers used relatively simple environments and focused on a specific hippocampal circuit. In real life, experiences are far more complex, and many different brain areas are involved in processing and storing memories.

Additionally, the study treated inhibitory neurons as a uniform group, while in reality, there are many subtypes with different roles. Future research could explore how different types of inhibitory neurons contribute to memory selection and whether this mechanism applies in other brain regions or during sleep.

The study, “Inhibitory plasticity supports replay generalization in the hippocampus,” was authored by Zhenrui Liao, Satoshi Terada, Ivan Georgiev Raikov, Darian Hadjiabadi, Miklos Szoboszlay, Ivan Soltes, and Attila Losonczy.