Scientists uncover brain’s “reset button” for splitting memories into distinct events

New findings published in the journal Neuron suggest that the locus coeruleus—a tiny region buried deep in the brainstem—may play a key role in helping the brain separate one memory from the next. This cluster of neurons, long known for its role in arousal and attention, appears to become especially active when meaningful changes occur, such as a shift in context or environment. That activity seems to serve as a kind of neural punctuation mark, telling the brain where one event ends and another begins.

The study, led by researchers at UCLA and Columbia University, used a combination of functional brain imaging and pupil tracking to monitor moment-by-moment changes in brain activity as people encoded and later recalled sequences of information. The results indicate that bursts of activity in the locus coeruleus tend to signal important transitions and promote the creation of distinct memory representations—possibly by “resetting” memory circuits in the hippocampus.

“Life unfolds continuously, but to make sense of it, our brains naturally break it up into meaningful events. We do not remember the past as one long stream. Instead, we recall it as distinct chapters that shape our personal stories and understanding of the world. We wanted to determine how our brains accomplish this remarkable feat,” said study author David Clewett, an assistant professor at UCLA.

Earlier research has pointed to the hippocampus as a critical player in episodic memory. This brain region is thought to support both the integration of events that occur close together in time and the separation of those that differ in context. Yet scientists have questioned what signals the hippocampus to shift between these two modes. The team behind the new study proposed that the locus coeruleus might be the trigger.

This small nucleus, which releases the neurotransmitter norepinephrine, is known to help the brain respond to novelty and unexpected changes. It also has strong connections to memory-related regions such as the dentate gyrus—a subregion of the hippocampus involved in distinguishing between similar experiences. The authors theorized that phasic bursts of locus coeruleus activity might act as a “reset” signal that prompts the hippocampus to update its memory representations when significant changes are detected.

Free daily newsletter

The researchers recruited 36 healthy young adults to participate in a memory task while undergoing brain scanning. After data exclusions due to technical issues or participant attrition, 32 individuals were included in the final analysis. Of these, 28 had valid eye-tracking data.

While lying in an MRI scanner, participants viewed series of neutral object images. Before each image, they heard a tone in either their left or right ear, which indicated which hand they should use to make a size judgment about the object. The tones were used not just for task instructions, but also to manipulate perceived context. Each sequence was organized into four blocks of eight images. Within each block, the same tone repeated in the same ear, creating a sense of continuity. At transition points, the tone changed in both pitch and ear—producing a perceptual event boundary.

After each sequence, participants completed memory tests outside the scanner that asked them to judge the order and temporal distance between pairs of items. The researchers also collected high-resolution brain scans to locate the locus coeruleus in each participant, and they used neuromelanin-sensitive imaging to estimate long-term activity levels in the region. Additionally, eye-tracking data were used to measure changes in pupil size, a known proxy for arousal and locus coeruleus activation.

The behavioral results showed that participants had more difficulty remembering the order of items that spanned an event boundary than items that occurred within the same context. This suggests that boundaries introduced by tone switches effectively segmented the experience into separate memories. At the same time, participants were slower to respond to objects that followed a boundary, which may indicate increased attention to contextual change.

Importantly, bursts of activation in the locus coeruleus during these boundaries were linked to reduced memory for the temporal order of item pairs—an indicator that the brain was separating rather than integrating those events. This trial-by-trial relationship between locus coeruleus activity and memory separation did not appear during stable stretches of the sequence.

“We were surprised that transitions between events, known as ‘event boundaries,’ didn’t always activate the locus coeruleus (LC), a core hub of the brain’s arousal system,” Clewett told PsyPost. “Even so, this pattern revealed something important: when the LC did respond, memories became separated. This suggests that only certain cues or level of engagement may be important for structuring new memories.”

To examine the effects on memory circuits more directly, the researchers looked at activation patterns in hippocampal subfields. They found that items separated by an event boundary were associated with reduced pattern similarity in the left dentate gyrus—indicating that the hippocampus encoded them as more distinct events. This pattern separation effect was more pronounced when locus coeruleus activation at the boundary was stronger. By contrast, another hippocampal subregion, CA2/3, showed the opposite trend, with increased pattern similarity during boundaries, although the meaning of this effect remains less clear.

The link between pupil dilation and brain activity also aligned with the hypothesis. Tones marking boundaries elicited larger changes in pupil size compared to repeated tones, and those dilations tracked with locus coeruleus activation.

Specific components of the pupil response predicted memory performance and attention to boundaries. Participants who showed stronger early pupil responses to boundary tones also had worse temporal order memory across those boundaries, supporting the idea that enhanced local processing comes at the expense of integrating events over time.

Beyond these phasic effects, the researchers explored what might happen when the locus coeruleus is in a heightened state of chronic activation. Using neuromelanin imaging, they found that individuals with higher structural signal intensity in the locus coeruleus—thought to reflect sustained norepinephrine output—showed weaker pupil responses to boundaries. This suggests that under conditions of hyperarousal, the system becomes less sensitive to meaningful changes, potentially impairing the segmentation process that supports memory organization.

Functional MRI analyses reinforced this idea. Participants with greater low-frequency fluctuations in locus coeruleus activity, a proposed marker of background activation, showed reduced responses to boundaries both in terms of pupil dilation and phasic neural activation. In other words, when the “alarm system” of the brain is always on, it may fail to ring when it matters most.

“Our key takeaways were that arousal processes, the brain’s alertness system, help us detect important changes and use those moments to organize information into distinct memories,” Clewett explained. “However, individuals who experience chronically elevated stress may fail to register these changes, making it harder to perceive and remember discrete events. This insight opens the door to new interventions that could target hyperarousal in conditions like PTSD or Alzheimer’s disease to help improve how memories are formed and remembered.”

The study offers compelling evidence for the role of the locus coeruleus in segmenting memory, but the findings are correlational. The researchers cannot definitively conclude that locus coeruleus activity causes the observed memory effects. They also note the technical challenges of imaging such a small and deeply located brain structure, although they took several steps to ensure accuracy.

“One of the biggest caveats is that the LC is tiny and difficult to image in humans,” Clewett noted. “While we cross-validated LC activation using both pupillometry and structural imaging, fMRI signals from this region are still noisy and should be interpreted with caution. Additionally, the limited spatial resolution of our fMRI scans makes it challenging to fully segment the hippocampus into its distinct subregions.”

“Future work could use techniques with higher spatial and temporal resolution to image these brain regions. It would also be helpful to manipulate arousal states more directly to determine the causal influence of the LC on memory structuring.”

Another open question concerns the relationship between stress, tonic locus coeruleus activation, and memory. The current findings suggest that chronic hyperarousal can blunt memory segmentation, but further research is needed to explore how interventions that regulate arousal—such as mindfulness, breathing exercises, or pharmacological treatments—might restore normal memory function in people experiencing stress-related disorders.

“Ultimately, our goal is to design behavioral or neuromodulatory interventions that help ‘quiet’ elevated LC activity and reduce hyperarousal in conditions like Alzheimer’s disease and PTSD,” Clewett said. “We are also interested in exploring whether these approaches can complement memory-based therapies aimed at improving how people perceive and remember events.”

The study, “Locus coeruleus activation ‘resets’ hippocampal event representations and separates adjacent memories,” was authored by David Clewett, Ringo Huang, and Lila Davachi