Boosting noradrenaline during learning makes the brain link memories more broadly, study finds

By Vijay Kumar MalesuReviewed by Susha Cheriyedath, M.Sc.Mar 16 2026

Scientists show that boosting noradrenaline while people learn does not strengthen memory itself but changes how the brain connects related experiences, revealing how arousal can expand the hippocampal “cognitive map” and increase memory overgeneralization days later.

Study: Noradrenaline causes a spread of association in the hippocampal cognitive map. Image Credit: beast01 / Shutterstock

In a recent study published in the journal Nature Communications, a group of researchers investigated whether increased noradrenaline during learning influences how the hippocampus links related memories within a cognitive map.

Cognitive maps organize relationships between memories

Imagine remembering a place and suddenly recalling other related experiences connected to it. The brain uses “cognitive maps” to organize relationships between events, objects, and places. These maps are largely supported by the hippocampus, which is required for both memory and navigation. Cognitive maps help people to predict results and connect ideas, but can also lead to memory distortions when associations spread too widely.

Noradrenaline is a neuromodulator that has been shown to influence an individual’s learning, attention, and synaptic plasticity, but the mechanism by which it affects cognitive map formation remains unknown. Therefore, if researchers can better understand how neuromodulators influence memory networks, it will help further develop theories of both flexible thinking and memory errors. However, additional research will be needed to determine how neurochemical signals regulate memory creation through associative processes.

A double-blind experiment tested the effects of increased noradrenaline

The researchers conducted a double-blind, randomized, placebo-controlled experiment involving 44 healthy adults. Participants were randomly assigned to receive either a 10 mg dose of atomoxetine (a drug that increases noradrenergic signaling by blocking noradrenaline reuptake) or a placebo capsule. To get unbiased results, neither the participants nor the researchers knew which treatment was administered until the study ended.

The participants took the capsule and waited 90 minutes to ensure that the drug reached peak concentration. They then completed a learning task designed to form a structured network of associations between visual stimuli. The task involved pairs of bird images presented within different room contexts. Each bird appeared in two different pairings, creating a hidden ring-like structure of relationships among the stimuli. These associations were learned by the participants through a three-alternative forced-choice task with feedback until they recorded a performance accuracy of more than 90%.

Functional magnetic resonance imaging was used to measure brain activity during later testing, and pupil diameter was used as a physiological marker of noradrenergic activity. Additionally, the levels of GABA+ (gamma-aminobutyric acid plus macromolecules) and other brain metabolites were measured using magnetic resonance spectroscopy in the right lateral occipital complex, a brain region involved in visual object processing. Participants returned 4 days later to test explicit and implicit memory for the visual associative network and how its elements were linked.

Elevated noradrenaline linked to broader memory associations

Participants in both the atomoxetine and placebo groups successfully learned the bird associations during the training phase. Accuracy levels during the learning task were similarly high across both groups, suggesting that increasing noradrenaline did not improve the basic ability to encode individual memories.

However, differences emerged during later memory tests designed to measure how participants linked related information. Four days after learning, participants were tested on their ability to remember relationships between birds and contextual cues. While overall memory accuracy remained similar across groups, the error patterns revealed an important difference. Individuals who had learned the task under elevated noradrenaline were significantly more likely to make “overgeneralization” errors. Instead of selecting the correct contextual cue, they frequently chose a cue from a nearby association within the task's ring structure.

This pattern suggested that noradrenaline was associated with a broader spread of associations between related memories. For example, if a bird had originally been paired with a specific room, participants exposed to higher noradrenaline were more likely to mistakenly associate it with a neighboring contextual cue that had never been directly linked to it.

The hypothesis of increased noradrenergic activity in the atomoxetine group was supported by physiological measurements. Pupil responses to unexpected “oddball” stimuli were significantly larger in these participants, indicating heightened physiological arousal consistent with noradrenergic signaling. Magnetic resonance spectroscopy also revealed reduced levels of GABA+ in the lateral occipital complex of the brain, which processes visual objects. This reduction is consistent with decreased inhibitory neural tone and increased cortical excitability.

Functional magnetic resonance imaging analyses further provided evidence consistent with a spread of associations across the cognitive map in the right hippocampus and nearby medial temporal regions, including the parahippocampal cortex. Neural activity in these regions showed stronger overlap between representations of stimuli that were close together within the learned network.

Importantly, the strength of this neural spread predicted later behavioral outcomes. Participants who showed stronger neural spreading immediately after learning were more likely to display greater overgeneralization errors several days later. Physiological measures such as pupil dilation and reduced GABA+ were associated with this neural spreading rather than directly predicting behavioral errors.

Computational modeling provided further insight into the mechanisms behind these findings. A spiking neural network model simulated how reduced inhibition during learning could allow excitatory connections between memory representations to spread more broadly. When inhibitory balance was weakened, nearby memory nodes became increasingly linked, producing neural activity patterns that resembled the experimental observations.

Findings suggest arousal-related neurochemistry shapes memory networks

This study suggests that elevated noradrenaline during learning may promote the spread of associations within the hippocampal cognitive map. By potentially reducing inhibitory balance and allowing graded strengthening of nearby representations, noradrenaline may enable nearby memories to become more strongly linked, even when they were not directly experienced together.

Although this process may support flexible thinking and making inferences, it can create distortions and lead to overgeneralizations in the formation and retrieval of memories. These findings also provide insight into how heightened arousal-related neurochemical signals during learning might influence the way memories become interconnected over time. Understanding how noradrenaline organizes memory systems may help inform future research into conditions involving memory overgeneralization, such as anxiety disorders and post-traumatic stress disorder.