A specific neural pathway links the insula to the creation of new memories

A new study has identified a specific neural pathway that connects the brain’s processing of internal states to the formation of new memories. Researchers discovered that distinct populations of neurons within the insula, a region deep in the brain associated with emotion and bodily awareness, communicate directly with the hippocampus to help lock in memories of emotionally charged words. These findings, published in Nature Neuroscience, suggest that the insula is not a uniform structure but rather a mosaic of functional clusters that perform separate tasks.

The human brain is often discussed in terms of large regions with singular functions, such as the hippocampus for memory or the amygdala for fear. However, this view oversimplifies the complex reality of how different areas interact to produce the human experience. Neuroscientists understand that emotionally charged events are easier to remember than neutral ones. The mechanisms behind this phenomenon have remained somewhat opaque.

While the hippocampus is recognized as the central engine for episodic memory, it does not work in isolation. It relies on inputs from cortical regions to determine which information is worth saving. The insula has been a primary suspect in this process due to its role in processing affect and salient stimuli. Despite this, the precise timing and direction of the conversation between the insula and the hippocampus during the moment a memory is born have been difficult to capture.

To address this gap, a team of researchers led by Weichen Huang and Josef Parvizi at Stanford University School of Medicine undertook a detailed investigation using high-precision recording methods. The study aimed to map the functional architecture of the insula at a microscopic scale. The investigators sought to understand if the entire insula engages during memory formation or if specific sub-regions are responsible. They also wanted to determine if these potential memory-related sites were distinct from those that process the emotional tone, or valence, of a word.

The researchers recruited 16 participants who were already undergoing intracranial electroencephalography (iEEG) for the evaluation of focal epilepsy. This clinical procedure involves implanting electrodes directly into the brain to locate the source of seizures. The placement of these electrodes provided the scientists with a rare opportunity to record electrical activity from the insula and hippocampus with millisecond precision. This method offers much higher temporal resolution than non-invasive scans like fMRI.

Free daily newsletter

During the experiment, participants viewed a series of words with positive or negative emotional associations, such as “successful” or “loser.” They were asked to rate the emotional intensity and valence of each word. After a brief distraction task involving counting backward, the participants attempted to recall as many of the words as possible. This design allowed the researchers to separate the brain activity associated with simply viewing a word from the activity associated with successfully remembering it.

The analysis of the electrical signals revealed a complex landscape within the insula. The researchers found that the insula is not functionally homogeneous. Instead, it contains interspersed populations of neurons with distinct behaviors.

One specific subset of recording sites in the insula showed a particular pattern of activity that predicted whether a participant would later recall a word. In these locations, the researchers observed a decrease in the “aperiodic exponent” of the electrical signal. This technical metric reflects a shift in the broadband background activity of the brain rather than a specific rhythmic oscillation. When this shift occurred, the participant was more likely to remember the word.

The researchers termed these locations “INSDE” sites, standing for sites with decreased exponent. The activity in these memory-predicting sites appeared to be tightly synchronized with the hippocampus. Specifically, the electrical shift in the insula happened just before the hippocampus generated a “ripple.” A sharp-wave ripple is a high-frequency burst of brain activity that is known to be essential for consolidating memories.

The timing suggests a directional influence. The insula appears to signal the hippocampus to pay attention and store the information. This supports the idea that the insula helps tag salient information for long-term storage.

Simultaneously, the researchers identified a second, separate group of sites within the insula. These locations showed an increase in the aperiodic exponent, a pattern opposite to the memory sites. These sites, labeled “INSIE,” tracked the emotional valence of the words. The activity here changed depending on whether a word was positive or negative.

Despite their proximity to the memory sites, these valence-tracking sites did not predict memory performance. Their activity did not correlate with whether the participant successfully recalled the word later. This finding indicates a clear division of labor within the insula. One set of neurons processes the emotional content, while a neighboring set facilitates the storage of the memory.

To confirm the direction of the communication between these brain regions, the researchers used direct electrical stimulation. They delivered small electrical pulses to the different sites in the insula while recording the response in the hippocampus. This technique allows scientists to move beyond correlation and establish causal connections.

When the researchers stimulated the memory-related INSDE sites, they observed an immediate and sharp electrical response in the hippocampus. This evoked potential confirms a direct functional connection. The signal travels from this specific population of insular neurons to the memory center.

In contrast, stimulating the valence-related INSIE sites produced no such response in the hippocampus. Even though these neurons are located in the same anatomical structure, they do not possess the same hard-wired line to the memory system. This physical evidence reinforces the idea of the insula as a mosaic of functionally selective populations.

The team also tested the reverse connection by stimulating the hippocampus. They found that pulses originating in the hippocampus evoked slow responses across all sites in the insula. This suggests an asymmetric relationship. The hippocampus broadcasts information broadly to the insula, but only specific insular sites send rapid, direct signals back to the hippocampus to initiate memory encoding.

These results paint a picture of the insula as a highly specialized structure. It does not process experience as a single block. Instead, it utilizes intermingled clusters of neurons to handle different aspects of cognition. Some clusters process the “good” or “bad” of an experience, while others signal the importance of that experience to the memory systems.

The study does have limitations inherent to human intracranial research. The participants were patients with epilepsy, and their brain physiology might differ slightly from the general population. The number of participants was relatively small due to the invasive nature of the recording method. Additionally, the study focused on single words, so it remains to be seen if these same pathways apply to complex visual scenes or autobiographical events.

Future research will need to explore how these distinct insular populations interact with other brain networks. It is also unclear whether the memory-related sites are responding to a different dimension of emotion, such as arousal or intensity, which was not fully isolated in this experiment. Understanding these precise pathways could eventually inform treatments for disorders where memory and emotion are dysregulated, such as post-traumatic stress disorder.

The study, “Direct interactions between the human insula and hippocampus during memory encoding,” was authored by Weichen Huang, Dian Lyu, James R. Stieger, Ian H. Gotlib, Vivek Buch, Anthony D. Wagner, and Josef Parvizi.