Under-researched: extinction learning and RSD in ADHD

Ask ADHD adults about rejection and many will give you a name for it now: rejection sensitive dysphoria, or RSD. It’s the sudden, outsized pain that follows a rejection - real or imagined. A short reply, a cancelled plan, one bit of feedback from the boss; and the floor drops out.

Two things about RSD keep bothering me. The first is how much it looks like complex PTSD. The three symptom groups that turn PTSD into complex PTSD - big emotional swings, a sense of being worthless, trouble in close relationships - are almost word for word what people describe as RSD. Yet I can’t find anyone who has put the two side by side.

The second is a set of animal studies I spent time with this year in my neuroscience coursework. They show that the brain unlearns a fear through a small burst of dopamine; one that fires when something bad is expected and doesn’t happen. ADHD is, among other things, a condition of altered dopamine signalling.

Put the two together and there’s an obvious question. Does ADHD slow down the unlearning of social fear? If the dopamine “it didn’t happen” signal is weaker in ADHD, then every rejection that doesn’t come would teach us a little less than it teaches everyone else. Over years, the fear wouldn’t fade; it would pile up. And what it piled up into might look a lot like complex PTSD, without any single trauma behind it.

That’s my idea. It isn’t proven, and I’ll say where it could be wrong. ¹

1 RSD is a name, not yet a diagnosis

You won’t find RSD in the DSM or the ICD. The term came from the psychiatrist William Dodson, who first wrote about it in lay publications, based on what he saw in his adult ADHD patients. In a 2024 editorial, he and Edward Modestino describe it as episodes that start with a perceived rejection and move straight into a painful low mood. ² They also admit that, at the time, there was no scientific publication with evidence for it; journals often turned their own paper down because nothing on it existed yet.

So the research mostly uses an older idea: rejection sensitivity. Downey and Feldman defined it in 1996 as a tendency to anxiously expect rejection, to see it quickly in what other people do, and to overreact to it. ³ The part that matters here is where they think it comes from: early experiences of rejection leave a learned expectation that others will reject you. In their model, rejection sensitivity is learned.

The research linking rejection sensitivity to ADHD is small, but growing. Müller and colleagues surveyed 304 Hungarian college students and found that more ADHD symptoms went with higher rejection sensitivity. Students who were good at savouring positive experiences showed a weaker link. ⁴ Babinski and colleagues had 391 young adolescents play a computer game in which other players vote to keep them or kick them out. The more ADHD symptoms they had, the stronger their early brain response to being rejected - and the weaker their brain response to being accepted. ⁵

Not every study agrees. Canu and Carlson found no higher rejection sensitivity in 53 young men with ADHD than in 25 without, although the ADHD men did have lower self-esteem. ⁶⁷

The smallest study is the most telling. Rowney-Smith and colleagues ran focus groups with five ADHD undergraduates. ⁸ Three themes came out: pulling away from people, masking, and strong physical sensations. They often felt that expecting rejection was worse than the rejection itself. Some said masking left them feeling cut off from themselves. Five people can’t tell us how common this is, but they do describe what it’s like.

So why do we have it? The evidence points to two reasons.

The first is that we meet more rejection. Children with ADHD are less liked by classmates, more often rejected by them, and have fewer friends. ⁹ Adults with ADHD describe a lifetime of criticism that has worn down their sense of self-worth. ¹⁰ If rejection sensitivity is learned from rejection, ADHD gives us plenty of lessons.

The second is that our emotions are harder to regulate. Shaw and colleagues estimate that emotional dysregulation affects about 25 to 45 per cent of children and 30 to 70 per cent of adults with ADHD. ¹¹ A meta-analysis of 13 studies and 2,535 adults found much higher emotional dysregulation in adults with ADHD than in people without it. ¹²

But neither explains why the fear stays. Everyone gets rejected. Most people also get plenty of acceptance, and the fear fades. This essay is about why, for us, it might not.

2 Why then, is RSD a mirror of C-PTSD?

The World Health Organization’s ICD-11 splits trauma responses into two related diagnoses. PTSD has three core symptom groups: re-living the event, avoiding reminders, and a constant sense of threat. Complex PTSD (C-PTSD) has those three, plus three more. The ICD calls them disturbances in self-organisation: ¹³

1. Affect dysregulation - ongoing trouble managing emotions, including strong reactions, intense anger, emotional numbness or dissociation.
2. Negative self-concept - lasting beliefs of being worthless, defeated or diminished, often with deep shame, guilt or a sense of failure.
3. Disturbances in relationships - ongoing trouble forming or keeping close relationships, including pulling away, difficulty trusting, or avoiding closeness.

In a study of 302 people seeking treatment after interpersonal trauma, Cloitre and colleagues found a distinct group with this six-part profile. Long-running trauma, such as childhood abuse, predicted it more strongly than single events did. ¹⁴

Now set those three extra groups next to what the RSD research describes.

Table 1
The Three C-PTSD Self-Organisation Clusters Beside Reported Features of Rejection Sensitivity in ADHD

C-PTSD cluster (ICD-11)	Rejection sensitivity and RSD in ADHD
Affect dysregulation	Sudden, intense distress after perceived rejection 2; high emotional dysregulation in ADHD adults 1112; masking that leaves people feeling cut off from themselves 8
Negative self-concept	Lower self-worth after a lifetime of criticism 10; lower self-esteem in ADHD men 6
Disturbances in relationships	Expectation of rejection that strains close relationships 3; withdrawal and loneliness 8

Note: The match is between symptom shapes, not causes. C-PTSD requires exposure to an extreme and usually prolonged threat, plus the core PTSD symptoms. RSD has no such requirement.

Science treats them as two different things. They’re diagnosed differently, they sit in different parts of the manuals, and RSD isn’t formally a diagnosis at all. But the differences people usually point to get thinner the closer you look.

The first is trauma. On paper, C-PTSD needs one. The ICD-11 links it to sustained exposure to repeated or multiple traumatic stressors - childhood abuse, domestic violence, torture - and those are extreme events. ¹⁴ Being the kid who’s always corrected, left out or told they’re “too much” doesn’t meet that bar.

But the bar itself is shaky. Life events that don’t count as traumatic - a divorce, losing a job - have been found to produce as many PTSD symptoms as events that do. ¹⁵ In a study of 454 undergraduates, the people whose worst event didn’t meet the official definition actually reported more severe PTSD symptoms than those whose event did. ¹⁶ A review of 62 studies found that emotional abuse and emotional neglect in childhood - not beatings, not disasters - were linked to PTSD and C-PTSD symptoms, and just as strongly to C-PTSD as to PTSD. ¹⁷ And among trauma-exposed autistic adults, more than 40 per cent showed probable PTSD, whether or not their event met the official definition. ¹⁸ Rumball and colleagues concluded that treatment shouldn’t be held back just because the event looks unusual.

So no, an ADHD childhood isn’t a trauma by the manual’s definition. It’s thousands of small ones, one after another, with very little relief in between. I think the bar is set at what’s easy to count, rather than at what does the damage.

The second difference people point to is that RSD only matches C-PTSD’s three extra symptom groups, not PTSD itself. I don’t think that holds either. Look at PTSD’s own three core symptoms.

Re-living the event. In PTSD, this is the past breaking into the present: vivid memories, flashbacks, the feeling that it’s happening again. ¹³ The RSD version is quieter, but it works the same way. “Don’t say that - last time you said that, they laughed.” “Don’t ask - last time you asked, they said no.” The old rejection turns up uninvited and decides what we do next. Unwanted memories of our own past - the kind that pop into our heads without being called - turn up more often across a whole range of conditions, ADHD included. ¹⁹ In adults with ADHD, rumination - along with a mind that wanders too much - helps explain why so many of us end up anxious or depressed. ²⁰²¹ It isn’t a flashback in the textbook sense, but it’s still the past replaying in the present.

Avoiding reminders. PTSD avoidance means steering clear of the people, places and situations that bring it back. ¹³ Rowney-Smith’s students pulled away from people and opportunities. ⁸ Some of the adults in Beaton’s study coped with criticism by avoiding it. ¹⁰ And the far end of avoidance is estrangement - cutting off the family who did the rejecting. It’s more common than most people think. In a large US study, 6 per cent of adult children reported a period of estrangement from their mother, and 26 per cent from their father, although most later reconnected. ²² It’s rarely about one event; it builds up over time, from several things. ²³²⁴ And adults with ADHD describe the families they grew up in as more rigid and less emotionally connected, with less parental warmth. ²⁵ I wrote about this in Take Me As I Am: nobody has measured how much family estrangement involves a neurodivergence that nobody recognised. I looked again for ADHD, and still found nothing. ²⁶ But from the inside, walking away from the people who kept rejecting you looks a lot like avoiding reminders.

A constant sense of threat. RSD is usually described in episodes: a trigger, then the crash. ² C-PTSD is described as persistent. But the episode is only the part you can see. Underneath it, Downey and Feldman describe an anxious expectation of rejection that leaves people constantly watching for signs of it - hypervigilant, in their word. ³ Rowney-Smith’s students said the waiting was worse than the rejection. ⁸ For a lot of us, the threat comes down to one feeling: that we don’t belong here. Wherever “here” happens to be. Researchers call it thwarted belongingness, and they take it seriously, because it’s tied to depression and suicide risk. In 217 college students with depressive symptoms, ADHD symptoms were linked to suicidal thoughts partly through it. ²⁷ In 341 early adolescents, it weighed most heavily on girls with ADHD. ²⁸ Medical students with ADHD described bullying, isolation and alienation, masking to avoid being pushed out, and a longing to belong. ²⁹ Not every study finds it: one US study found first-year college students with ADHD felt as much a part of their college as anyone else. ³⁰³¹ But never being quite sure you’re welcome is what a constant sense of threat feels like - it’s living on edge. The crash is just the part other people get to see.

So this is where I’ve ended up. RSD and C-PTSD aren’t diagnosed the same way, and they don’t start the same way. But they end up in the same place: the same replaying, the same avoiding, the same watching; and then the same emotional swings, the same sense of being worthless, the same trouble with the people closest to us. The cause is different, but the outcome is the same.

And that matters, because the diagnosis decides the treatment. Treatment for C-PTSD is built around the trauma. Treatment for ADHD is built around ADHD - and ADHD treatments often ease emotional dysregulation as well. ¹¹ Treat the outcome without knowing where it came from, and you can spend years treating the wrong thing.

For ADHD adults, this has been happening for years. Adult ADHD can look like bipolar disorder or borderline personality disorder, and that leads to mistaken diagnoses and treatment that doesn’t work. ³² Adults with ADHD more often walk into a clinic for depression, anxiety or alcohol & drug problems than for ADHD, and their ADHD gets put down to those other conditions. Leave the ADHD untreated, and outcomes stay poor even when everything else is treated. ³³ Women and girls are missed more often still, because anxiety and depression get the diagnosis instead. ³⁴

Autism, which has been studied more closely on this point, shows the same pattern. In a Dutch register of 1,211 autistic adults, one in four said they’d been given a psychiatric diagnosis before their autism diagnosis that they believe was wrong - and personality disorders topped the list. ³⁵ That’s the same wrong turn ADHD adults get sent down.

The confusion runs the other way too. In one NHS adult ADHD and autism service, four in five people referred didn’t meet the criteria - and the more childhood adversity someone reported, the less likely they were to be diagnosed. ³⁶ Trauma can look like neurodivergence, and neurodivergence can look like trauma. ³⁷ Either way, treatment has to aim at the cause.

That’s why I think the overlap between RSD and C-PTSD deserves more than a passing mention. Picture an ADHD adult who walks into a clinic with all three of C-PTSD’s extra symptom groups, and no single trauma to point to. What are they likely to be told? And what are they likely to be treated for?

So why hasn’t anyone joined RSD and C-PTSD together? The closest I found was in passing: Dodson and Modestino bring up complex PTSD only to say that rejection sensitivity is common there too, before arguing that RSD is something separate. ²

What has been studied is ADHD and ordinary PTSD. A meta-analysis found that people with ADHD were 2.9 times more likely to have PTSD, and people with PTSD 1.7 times more likely to have ADHD. ³⁸ The link runs both ways. And one group went a step further. Spencer and colleagues put 27 young adults with ADHD, and 20 without, through a two-day fear-learning task in a brain scanner. None had experienced trauma, and none had ever taken ADHD medication. The ADHD group still showed different activity in the brain areas that handle unlearning fear - and some of those differences matched what’s seen in PTSD. ³⁹

That’s the closest anyone has come to my idea, and it points the same way. But it used a non-social threat, it didn’t look at dopamine, and it didn’t ask about rejection.

3 How the brain unlearns a fear

The laboratory version is simple. An animal hears a tone, followed by a mild shock to the foot. After a few pairings, it freezes as soon as it hears the tone. That’s fear conditioning.

Then the tone is played again and again, with no shock. The freezing slowly fades. That’s extinction.

Extinction isn’t forgetting, though. The original fear memory stays put. The animal learns a second, newer memory - “the tone is safe now” - that competes with the first. And because that new memory is tied more tightly to the place and time it was learned, the old fear can come back with time, in a new setting, or after a reminder. ⁴⁰ Anyone who has felt fine for months and then been flattened by one cold email knows what that’s like.

What drives the new learning is surprise. The animal expects a shock, and it doesn’t come. That gap between what was expected and what happened - a prediction error - is what the brain learns from. Here, things turned out better than expected.

So which part of the brain notices that something went better than expected? Most of the evidence points to a small cluster of dopamine neurons in the midbrain, in the ventral tegmental area (VTA). They’re best known for firing when a reward is better than expected, and dipping when it’s worse. Two rat studies showed that these bursts and dips actually cause learning, rather than just going along with it.

Steinberg and colleagues trained rats to expect sugar water after a cue, and then took the sugar away. Normally the rats soon stop responding to the cue. But when the researchers used light to switch on the rats’ dopamine neurons at the moment the sugar should have arrived, the rats kept going for longer. The artificial burst told the brain “this is still good”, and it slowed the unlearning. ⁴¹ Chang and colleagues did the reverse. They briefly switched the dopamine neurons off at the moment of a reward. That short dip alone was enough to make the rats value the cue less afterwards, as if the reward had been smaller than they’d expected. ⁴²

That’s the reward side. In 2018, two groups working separately asked whether the same signal teaches safety.

Luo and colleagues used rats whose dopamine neurons could be silenced with light. The rats learned that a tone meant a shock. The next day they heard the tone without the shock. In one group, the dopamine neurons were silenced for the few seconds when the shock should have come. In another, they were silenced during the tone itself. Silencing during the tone made no difference. Silencing at the moment the shock failed to arrive greatly reduced the unlearning - on the day, and when the rats were tested again the next day. ⁴³

Two things stood out to me. First, some unlearning still happened. It was slowed, not stopped. Second, silencing the neurons alongside a tone, with no shock at all, didn’t create fear. The silencing wasn’t acting as a punishment. It simply took the safety lesson away.

Salinas-Hernández and colleagues did the same kind of work in mice. They first listened in on the dopamine neurons and found that about a quarter of them fired more when the expected shock didn’t come. That happened early on, while the missing shock was still a surprise, and the size of the signal closely tracked how quickly each mouse stopped freezing. ⁴⁴ Then they silenced those neurons, right at the moment of the missing shock. The mice kept freezing through the whole session, and the next day they froze as much as they had at the start. As far as the researchers could tell, no real unlearning had happened at all. And when they boosted the same neurons at the same moment, the mice unlearned faster.

So dopamine doesn’t only mark good things. It also marks bad things that fail to happen - and that mark is how the brain learns that a threat has passed. Take it away, and the fear stays.

Human studies point the same way. Haaker and colleagues gave people (and mice) L-dopa, a building block of dopamine, after an unlearning session; the safety memory held up better in new settings, and less fear came back. ⁴⁵ Esser and colleagues gave L-dopa during unlearning and found it changed the brain’s response at the exact moment an expected threat didn’t arrive. ⁴⁶ People with PTSD show a version of the same problem: Milad and colleagues found they unlearned a fear about as well as others on the first day, but held on to the lesson less well the next. ⁴⁷ And in mice, methylphenidate - the stimulant many of us take for ADHD - helped the safety memory stick, depending on the dose and timing. ⁴⁸⁴⁹

4 What I think is going on

Rejection works like a learned threat. A frown, a short reply, a pause before someone answers - for a lot of ADHD kids, those cues were often followed by being corrected or left out. ⁹¹⁰ Downey and Feldman’s model already treats rejection sensitivity as learned: the cues come to predict rejection. ³

If it’s learned, unlearning it is extinction. Every time the feared rejection doesn’t come - the friend replies, the boss was only busy - the brain should register a small “better than expected” moment, and the link should weaken a little.

Rejection sensitivity already looks like a problem with that unlearning. Olsson and colleagues taught people to fear angry faces, neutral faces and non-social images, and then let the fear fade. People high in rejection sensitivity were slower to let go of their fear of angry faces - but not of the neutral faces or the non-social images. ⁵⁰ That wasn’t an ADHD sample. But it’s the pattern I’d expect: a fear that won’t fade, and only for social threats.

That fading depends on dopamine, as the rat and mouse studies show. ⁴³⁴⁴ And dopamine signalling is different in ADHD. Adults with ADHD who weren’t taking medication show lower dopamine transporter and receptor binding along the brain’s reward pathway. ⁵¹ A meta-analysis found a weaker response in a key reward area when people with ADHD anticipate a reward. ⁵² Tripp and Wickens have argued that in ADHD the dopamine response doesn’t move forward in time from the reward to the cues that predict it, the way it should. ⁵³ And remember Babinski’s game: the adolescents with more ADHD symptoms showed a weaker reward-related brain response to being accepted. ⁵ That response is thought to reflect the kind of reward signal that dopamine prediction-error theories describe. ⁵⁴⁵⁵

So this is what I think is happening. We learn rejection at a normal rate, or faster, and we meet more of it. But we unlearn it more slowly, because the dopamine signal that should mark “it didn’t happen this time” is weaker, or less reliable. It’s not missing, just slower - remember that some unlearning still happened in Luo’s rats. But over decades, the gap between what we learn and what we unlearn adds up.

I think RSD in ADHD is, at least partly, a failure to unlearn social fear. The brain doesn’t get the “you’re safe” message as often as it should.

If that’s right, a lot about RSD starts to make sense.

It would explain why the bias runs one way. In two studies of adults with ADHD, early brain processing of positive images was reduced, but not of negative ones. ¹¹ In people high in rejection sensitivity, images of rejection strengthen the startle reflex, but images of acceptance don’t calm it. ⁵⁶ Babinski’s adolescents reacted more to rejection and less to acceptance. ⁵ A weak “better than expected” signal is what you’d expect to produce that lopsided pattern.

It would also explain why the reaction comes before any thought. RSD episodes are described as going straight from a perceived rejection to distress. ² Learned fear is fast and automatic; in the startle research, rejection cues set off a defensive reflex before anyone has had time to think about them. ⁵⁶

Reassurance doesn’t stick, either, because a safety lesson is tied to the setting where it was learned. ⁴⁰ What you learned with one friend doesn’t automatically carry over to the next.

Good stretches don’t last, because fear that has been unlearned can come back with time, or after a reminder. ⁴⁰ Relapse after months of feeling fine is what you’d expect.

Expecting rejection hurts more than the rejection itself, as Rowney-Smith’s students said, ⁸ because a threat cue that was never unlearned causes fear whether or not the bad thing then happens.

And avoidance makes it worse. Unlearning needs moments where the feared thing could happen and doesn’t. If you never ask, never apply and never reach out, there are no such moments. Avoiding rejection feels safer in the moment, but it keeps the fear in place.

It would even explain why RSD mirrors C-PTSD. C-PTSD builds its three extra symptom groups from a big, sustained threat. The same three could build up from thousands of small ones, each never quite unlearned.

There’s plenty here I can’t prove, though.

Nobody has measured that dopamine safety signal in anyone with ADHD. The ADHD dopamine findings I’ve leaned on are about receptor binding, anticipation and theory; dopamine in ADHD isn’t simply “low”, and I shouldn’t pretend it is. Until someone records what happens at the moment an expected rejection fails to arrive, this stays an idea.

The best clinical result for RSD looks, at first, like it points somewhere else. Dodson and Modestino report that about 60 per cent of their patients responded well to guanfacine or clonidine when both were tried. ² Those drugs act on noradrenaline receptors, not dopamine ones.

But the two chemicals are much closer than that makes them sound. Noradrenaline neurons make their noradrenaline out of dopamine, and some of them release dopamine as well. ⁵⁷⁵⁸ In the prefrontal cortex, the pump that clears away noradrenaline also clears away dopamine. ⁵⁹ Guanfacine’s main known effect is to strengthen the prefrontal networks that regulate attention and emotion ⁶⁰ - the same area where Spencer’s ADHD adults showed differences when recalling what they’d unlearned. ³⁹ And noradrenaline has its own part in unlearning fear. In rats, the noradrenaline cells that connect to the prefrontal cortex help extinguish fear, while a separate group connecting to the amygdala strengthens it. ⁶¹

So guanfacine helping with RSD doesn’t count against my idea. It may be working on the same unlearning circuit, from a different side. It’s still clinical observation rather than a trial, and it still suggests dopamine isn’t the whole answer. ⁴⁹

And rats fearing a shock aren’t people fearing rejection. It’s a big jump from one to the other. A meta-analysis of 23 experiments found that social anxiety wasn’t reliably linked to slower unlearning in standard lab fear tasks. ⁶² That might mean the link isn’t there, or that the usual tasks are the wrong tool. Olsson’s angry faces suggest the second, ⁵⁰ but I can’t rule out the first.

Trauma muddies things too. ADHD raises the risk of PTSD, ³⁸ so some ADHD adults with C-PTSD-like struggles may simply have had trauma. Spencer’s group got around that by screening it out, ³⁹ and anyone looking at this properly would need to as well.

And the problem might not be the teaching signal at all. Spencer’s ADHD adults differed in the brain areas that store and recall what was unlearned. ³⁹ Learning the lesson poorly and remembering it poorly are two different problems. My guess is it’s both.

Still, every piece of this is already in the published research. As far as I can tell, nobody has put them together. I’d love to see someone put ADHD adults through Olsson’s angry-face task and watch what happens at the moment the expected rejection doesn’t come.

What this changes for me

If RSD is a character flaw, the advice is “toughen up” or “don’t take it so personally”. If it’s a mood problem, the advice is medication. But if it’s even partly a learning problem - a “you’re safe” message that doesn’t get through often enough - then the target is the lesson itself. More moments where the feared rejection doesn’t happen, and a brain in a state to notice them. That’s close to what exposure therapy already does for other fears. I haven’t found it adapted for an ADHD brain and a social threat.

It also changes the story we tell ourselves. “I’m too sensitive” sounds like a judgement on who we are. “My brain is slower to register that I’m safe” is something we can work with - and that means sending the text, asking the question, and letting the answer be yes more often than we expect.

⁂

---