Scientists have discovered why some people can’t shake their anxieties, even after the threat is long gone.
Researchers at the University of Cambridge identified a specific disruption in dopamine signaling that prevents certain individuals from unlearning their fears.
The study, published in Nature Communications, reveals that when dopamine release in a brain region called the nucleus accumbens is impaired, the ability to extinguish learned fears vanishes almost entirely.
This isn’t just about being anxious or cautious.
It’s about a fundamental biological mechanism that determines whether your brain can update its threat database when circumstances change.
The research team used genetically modified mice that lacked the ability to release dopamine properly in specific brain circuits.
When these mice were conditioned to fear a particular sound paired with a mild shock, they learned the association normally.
But here’s where it gets interesting: when the researchers stopped pairing the sound with the shock, typical mice gradually stopped freezing in fear.
The extinction-deficient mice never did.
They remained frozen in a fear response that no longer matched reality, trapped by a memory their brains couldn’t update.
This discovery matters because approximately 31% of adults will experience an anxiety disorder at some point in their lives, according to the National Institute of Mental Health.
For many of these individuals, the core problem isn’t learning to fear something dangerous, it’s failing to unlearn that fear when the danger passes.
Think about someone who develops a fear of driving after a car accident.
The initial fear response is adaptive and protective.
But when that person still feels paralyzed with terror two years later on empty residential streets in broad daylight, something has gone wrong in the extinction process.
Their brain knows intellectually that the situation is safe, but the emotional circuitry hasn’t caught up.
The Dopamine Connection Nobody Expected
For decades, neuroscientists understood that dopamine played a role in reward and motivation.
We knew it helped us learn what feels good and pursue those experiences again.
What the Cambridge researchers discovered challenges a much deeper assumption about how learning itself works.
Most people assume that forgetting a fear is simply the opposite of learning it in the first place.
If you can condition fear by pairing a neutral stimulus with something bad, surely you extinguish it by breaking that association, right?
But here’s what most people get wrong: extinction isn’t forgetting at all.
It’s learning something new that competes with the original memory.
Your brain doesn’t erase the fear association.
Instead, it builds a new memory that says “this stimulus is now safe” and that new memory has to be strong enough to override the old one.
Dopamine turns out to be the neurochemical that makes this competing memory stick.
Without proper dopamine signaling in the nucleus accumbens, the brain can observe that the feared stimulus no longer predicts danger, but it can’t encode that observation into a durable new memory.
The old fear memory remains dominant because no competing safety memory ever forms properly.
Dr. Amy Milton, who led the Cambridge research team, found that the timing of dopamine release was crucial.
The dopamine signal needed to arrive precisely when the animal encountered the feared stimulus without the expected negative outcome.
That moment of prediction error, when expectation didn’t match reality, was exactly when dopamine needed to stamp in the new learning.
In the extinction-deficient mice, dopamine release at that critical moment was severely blunted.
The prediction error happened, the mice surely noticed at some level that the shock didn’t come, but without the dopamine signal, that noticing never consolidated into learning.
This explains something puzzling about exposure therapy in humans.
Exposure therapy, where people gradually confront their fears in safe contexts, works beautifully for some patients and barely touches others.
The traditional explanation blamed motivation, effort, or the quality of the therapeutic relationship.
But this research suggests something more fundamental: some individuals may have dopamine signaling patterns that make it neurobiologically harder to form robust extinction memories.
Why Your Brain Holds Grudges
The nucleus accumbens sits deep in the brain’s reward circuitry, receiving dopamine input from an area called the ventral tegmental area.
Traditionally, scientists thought of this pathway as the brain’s “pleasure center” or “motivation hub.”
But the Cambridge findings reveal it’s also a critical gateway for safety learning.
When this dopamine pathway functions properly, your brain maintains flexibility.
You can learn that something is dangerous when it threatens you, and you can learn that it’s safe again when circumstances change.
This flexibility is adaptive in a changing world.
Our ancestors needed to learn that a particular water source was dangerous if a predator claimed it as territory.
But they also needed to learn when that predator moved on and the water was safe again.
Staying afraid forever would mean dying of thirst unnecessarily.
The extinction-deficient mice demonstrated what happens when this flexibility breaks down.
In behavioral tests, normal mice showed freezing behavior about 80% of the time when first re-exposed to the fear-associated sound after extinction training.
After several extinction sessions, their freezing dropped to around 20%.
The genetically modified mice with impaired dopamine signaling maintained freezing levels above 70% even after extensive extinction training.
The fear simply wouldn’t budge.
Brain imaging studies in humans with post-traumatic stress disorder and specific phobias show similar patterns.
These individuals often have altered activity in reward-related brain regions, including the nucleus accumbens, during extinction learning tasks.
Their brains are working hard to process the safe exposure, but something in the consolidation process falters.
Interestingly, the Cambridge researchers found that the deficit was specific to extinction learning, not to fear learning itself.
The mice with disrupted dopamine signaling learned initial fear associations just as quickly and strongly as normal mice.
Their problem was purely in the unlearning phase.
This specificity suggests that fear acquisition and fear extinction rely on partially separate neural mechanisms.
Understanding this distinction opens new therapeutic possibilities.
If we can identify people whose dopamine systems function normally for most types of learning but struggle specifically during extinction, we might be able to target that process directly.
The Chemistry of Letting Go
The practical implications of this research extend beyond understanding anxiety disorders.
They touch on how we design treatments, how we think about resilience, and how we understand individual differences in emotional regulation.
Current anxiety treatments often take a one-size-fits-all approach.
Cognitive behavioral therapy with exposure components is considered the gold standard, and it works well on average.
But “on average” masks enormous individual variation.
Some patients achieve full remission after a few months of treatment, while others show minimal improvement despite years of therapy.
The Cambridge findings suggest we might be able to identify extinction-resistant individuals before treatment even begins.
If someone has markers of impaired dopamine function in extinction-relevant circuits, they might benefit from augmentation strategies that enhance dopamine signaling during exposure therapy.
Several research teams are already exploring this possibility.
Studies have tested whether medications that boost dopamine availability, or that enhance the dopamine system’s responsiveness, can improve extinction learning in animals and humans.
One approach uses L-DOPA, a dopamine precursor, administered shortly before extinction training sessions.
Early results suggest this might help some individuals form stronger extinction memories.
Another strategy involves environmental factors known to enhance dopamine function naturally.
Physical exercise, for instance, increases dopamine receptor density and enhances dopamine signaling in reward-related brain regions.
Some preliminary studies indicate that aerobic exercise performed shortly before exposure therapy sessions may improve treatment outcomes, potentially by priming the dopamine system to better encode extinction learning.
The research also sheds light on why traumatic memories are so persistent.
Traumatic events often involve intense arousal, pain, or life threat, all of which create extremely strong initial memory encoding.
That strong encoding isn’t necessarily the problem, survival depends on remembering genuine dangers.
The problem emerges when the brain can’t later encode equally strong safety signals to compete with that traumatic memory.
If your dopamine signaling is already compromised by genetics, chronic stress, or other factors, you’re left with a brain that carved a deep groove for the danger memory but can only scratch a shallow mark for the safety update.
The traumatic memory wins by default, not because it’s more accurate, but because it’s neurochemically privileged.
Beyond Fear: What Else Can’t We Unlearn?
The Cambridge research focused specifically on fear extinction, but dopamine’s role in the nucleus accumbens likely extends to other forms of learning and unlearning.
Addiction, for example, involves learning powerful associations between cues and drug rewards.
Recovery requires extinguishing those associations, learning that formerly drug-associated environments and stimuli no longer predict the drug experience.
Many of the same brain circuits implicated in fear extinction also play roles in the extinction of drug-seeking behaviors.
People with substance use disorders often report that encountering drug-associated cues triggers intense cravings and urges even years into recovery.
A person who used cocaine in a particular friend’s apartment might feel overwhelming cravings when visiting that space, even if they haven’t used cocaine in a decade.
That persistent cueing effect suggests extinction failure, the brain never fully encoded the new reality that this environment is now drug-free.
Similar patterns appear in obsessive-compulsive disorder, where individuals struggle to update beliefs about danger even when they intellectually know their compulsions are excessive.
A person with contamination-related OCD might rationally understand that touching a doorknob won’t cause illness, but their emotional brain hasn’t extinguished the danger association.
Exposure and response prevention therapy, the most effective treatment for OCD, is essentially structured extinction learning.
The Cambridge findings suggest that individuals with OCD might have subtle differences in how their dopamine systems respond during these critical learning moments.
Even in non-clinical populations, individual differences in extinction learning might explain why some people are more adaptable than others.
Consider two people who both had negative experiences with public speaking in high school.
One gradually becomes comfortable with presentations through repeated positive experiences in college and work settings.
The other remains anxious about public speaking for life, despite many objectively successful presentations.
The difference might not be in their initial fear conditioning or even in their objective experiences during extinction, but in how effectively their brains encode those new positive experiences into competing memories.
Some people’s dopamine systems might naturally create strong, flexible updates to their emotional associations, while others maintain more rigid connections to past experiences.
Measuring What Matters
One of the most exciting aspects of this research is its potential for translation into measurable biomarkers.
Currently, mental health diagnosis relies almost entirely on reported symptoms and observed behaviors.
We don’t have blood tests for anxiety disorders or brain scans that definitively diagnose PTSD.
The Cambridge findings point toward specific, measurable brain mechanisms that could eventually serve as objective markers of extinction capacity.
Functional MRI studies can already measure brain activity in the nucleus accumbens during learning tasks.
PET imaging can assess dopamine receptor availability and dopamine release in response to specific stimuli.
These tools are currently used mainly in research settings, but they could eventually inform clinical decision-making.
Imagine a future where someone seeking treatment for a specific phobia first undergoes brief imaging to assess their dopamine response patterns during a standardized extinction learning task.
If the scan reveals robust dopamine signaling during prediction errors, standard exposure therapy proceeds with high confidence of success.
If the scan shows blunted dopamine responses, the treatment plan might immediately incorporate dopamine-enhancing strategies, saving months of ineffective standard treatment.
This isn’t science fiction, the tools exist and the biological mechanisms are increasingly well-understood.
The gap between current practice and this future is primarily one of validation, standardization, and cost-effectiveness rather than scientific feasibility.
Genetic markers might eventually play a role too.
Dopamine function is influenced by numerous genes, including those that control dopamine synthesis, release, reuptake, and receptor sensitivity.
Variations in genes like COMT, which breaks down dopamine in the prefrontal cortex, or DRD2, which encodes a key dopamine receptor, have already been associated with differences in extinction learning in some studies.
As our understanding of the polygenic architecture underlying dopamine function grows more sophisticated, genetic risk scores might help identify individuals likely to struggle with extinction learning.
These biomarkers wouldn’t be deterministic, brain function is far too complex and plastic for simple genetic determinism.
But they could provide probabilistic information that helps optimize treatment selection and intensity.
The Flexibility We Take for Granted
Perhaps the most profound implication of this research is what it reveals about cognitive flexibility as a fundamental dimension of brain health.
We often think about intelligence, memory, or emotional regulation as key aspects of psychological function.
But the ability to update our emotional associations, to learn that what was once dangerous is now safe or that what was once rewarding is no longer worth pursuing, may be equally fundamental.
This flexibility isn’t a single trait but a collection of specific mechanisms, each vulnerable to disruption in different ways.
The Cambridge researchers identified one such mechanism involving dopamine in the nucleus accumbens.
Other research teams have identified additional mechanisms involving different neurotransmitters, different brain regions, and different time scales of learning.
Serotonin signaling in the prefrontal cortex influences how we represent and update predictions about negative outcomes.
Norepinephrine in the amygdala affects how arousal modulates emotional memory consolidation.
Endocannabinoid signaling in the hippocampus helps gate which memories get updated versus preserved.
Each of these systems contributes to our overall capacity to adapt our emotional responses to a changing world.
When they work well together, we experience appropriate fear that protects us without imprisoning us, appropriate caution that keeps us safe without paralyzing us, appropriate attachment to rewards without spiraling into addiction.
When any of these systems falters, our emotional lives become less flexible.
We may find ourselves stuck in patterns that no longer serve us, unable to fully embrace the present because our brains remain trapped in the past.
The beauty of research like the Cambridge study is that it transforms these subjective experiences into concrete, testable biological hypotheses.
Feeling stuck isn’t a personal failing or a character weakness.
For some individuals, it reflects specific neurobiological patterns that can be measured, understood, and potentially modified.
What This Means for You
You don’t need a genetics lab to think about extinction learning in your own life.
The principles the Cambridge researchers uncovered show up in everyday experience, even if we don’t usually think about them in terms of dopamine and brain circuits.
Consider where you might have formed strong emotional associations that haven’t updated even though circumstances have changed.
Maybe you still feel a flutter of anxiety when your phone rings, even though you’re no longer waiting for stressful news.
Maybe you still avoid a particular restaurant, even though the person you associate with that place is long gone from your life.
Maybe you still feel defensive in meetings with authority figures, even though you’ve developed expertise and confidence in your field.
These persistent emotional reactions aren’t irrational, they’re the product of learning mechanisms that prioritize caution over flexibility.
From an evolutionary perspective, failing to learn about danger is often more costly than failing to unlearn it.
If your ancestor missed the signs that a particular berry was poisonous, they might die.
If they remained cautious about that berry even after the toxic variety went extinct in their region, they just missed out on some extra calories.
Our brains evolved to be better at acquiring fears than extinguishing them because the costs of these two types of errors were asymmetric.
But in the modern world, where many of our learned fears involve social situations, performance scenarios, or past traumas that won’t recur, this asymmetry can leave us burdened with outdated defensive reactions.
The good news is that extinction learning, even if it requires more effort for some people than others, is still learning.
It can be facilitated by the right conditions.
Repeated, safe exposure to feared stimuli in contexts that feel controllable helps build extinction memories even in people whose dopamine systems don’t make this easy.
Social support during exposure, reframing the experience as gathering evidence rather than testing courage, and celebrating small victories all help strengthen extinction learning.
Some people find that physical exercise before confronting a feared situation helps, possibly by enhancing dopamine availability.
Others find that mindfulness practices, which may improve interoceptive awareness and reduce defensive arousal, create better conditions for extinction learning.
The key is recognizing that if extinction learning feels harder for you than it seems to be for others, that difference is real and biological, not a personal inadequacy.
The Research Continues
The Cambridge study opens more questions than it answers, which is exactly what good science should do.
One major question involves the specificity of the dopamine deficit.
The researchers used genetic modifications that affected dopamine release broadly in the nucleus accumbens, but this brain region contains multiple types of neurons with different connectivity patterns and functions.
Do all dopamine signals in the nucleus accumbens contribute equally to extinction learning, or are specific subpopulations of neurons particularly important?
Another question concerns individual differences in humans.
While the mouse model created an extreme deficit in dopamine signaling, human variation is more subtle and continuous.
Most people fall somewhere on a spectrum of extinction learning capacity, not into discrete categories of capable versus incapable.
Understanding the full range of genetic, developmental, and environmental factors that shape this capacity remains an ongoing challenge.
The interaction between dopamine systems and other neurotransmitter systems also needs further investigation.
The brain doesn’t operate as a collection of independent chemical systems but as an integrated network where each neurotransmitter influences others.
Dopamine in the nucleus accumbens interacts with glutamate, GABA, serotonin, and other signaling molecules to produce its effects on learning.
Targeting dopamine alone might not be sufficient to fully restore extinction learning in all individuals who struggle with it.
Some researchers are investigating whether the timing and context of dopamine signaling can be manipulated to enhance extinction learning.
If dopamine needs to arrive at precisely the right moment during extinction training, could we use technologies like transcranial magnetic stimulation or focused ultrasound to modulate brain activity and optimize that timing?
These questions will occupy researchers for years to come, gradually refining our understanding and expanding our therapeutic toolkit.
A Final Thought
The mice in the Cambridge study couldn’t tell the researchers how they felt, trapped in a fear response they couldn’t update.
But anyone who has struggled with persistent anxiety or intrusive memories can imagine it.
That feeling of knowing intellectually that you’re safe while your body remains convinced of danger, of wanting desperately to move forward while some part of your brain insists on scanning for threats that no longer exist.
What this research gives us is not just an explanation for that feeling, but a path toward changing it.
By identifying the specific biological mechanisms that allow us to update our emotional memories, scientists are creating new possibilities for helping people who have been stuck.
The work is still early, still mostly in laboratory animals, still far from being a standard part of clinical care.
But the trajectory is clear, from mystery to mechanism to medicine.
The next time you successfully overcome a fear, take a moment to appreciate the complex biological dance happening in your brain.
Dopamine molecules are being released at precisely the right moment in precisely the right location, binding to receptors that trigger cascades of cellular changes, ultimately updating your emotional memory systems.
It’s remarkable that something so subjective, so personal as letting go of a fear, has such specific molecular machinery behind it.
And for those who find letting go harder than it should be, knowing that machinery exists is the first step toward learning to work with it more effectively.
