You are standing at the kettle on a Thursday when your shoulder does something it has never done before, and only then does it occur to you that the shoulder has been doing that since Monday. You worked through it. You carried the shopping with it. You slept on it. Somewhere underneath all of that, quietly, the signal was there the whole time, and it did not get through to you until it got loud enough to shout.
Autism and pain meet at interoception, the internal sense that tells you what is happening inside your own body. When that channel runs differently, pain can arrive late, arrive as a vague wrongness rather than a location, or arrive all at once when it is already severe. Some of us register very little at first and a great deal later. Some of us feel pain constantly and are told we are exaggerating. The research does not support the myth that we do not feel pain. What it shows is a difference in how pain is detected, described and believed, which is a completely different problem, and a more dangerous one.
What the research shows
- In a small quantitative sensory testing study of 13 autistic adults and 13 matched controls, the autistic group showed higher light touch detection and mechanical pain thresholds, alongside a greater range of extreme scores and atypical responses including paradoxical heat sensation and mechanical allodynia. Vaughan et al. (2020)1
- In a heat pain study of 24 autistic and 19 non‑autistic adults, pain thresholds did not differ between groups, but the autistic group rated the same pain as more intense and scored higher on pain‑related anxiety and fear of pain. Failla et al. (2020)2
- In a study of 52 autistic adults assigned female at birth, 65% reported pain within the previous 24 hours, with high rates of pain interference in daily life. Haid et al. (2025)3
- In a survey of 507 autistic adults, 80% reported difficulty visiting a GP compared with 37% of non‑autistic respondents, 72% had trouble deciding whether their symptoms even warranted a visit, and 53% struggled to communicate with the doctor once they got there. Doherty et al. (2022)4
Pain does not always arrive as a message
Most health advice assumes pain works like a doorbell. Something is wrong, the bell rings, you go and look. For a lot of us it does not work like that. Interoception, the sense that carries hunger, thirst, temperature, bladder, heart rate and pain, is the channel that tells you what is happening inside your own skin. When that channel is noisy or slow or low‑resolution, pain does not arrive as a message. It arrives as weather.
So instead of “my knee hurts”, you get a general sense that today is a bad day and you cannot say why. Instead of “stop, that is an injury”, you get a faint background wrongness that you file under tired. Instead of a clean signal, you get a delay: the thing happened on Monday, and your body files the report on Thursday.
This is not stoicism. There is a real difference between deciding to push through pain and never receiving the interrupt at all, and only one of those is a choice. From outside they look identical, which is why so many of us present late with something that should have been caught early. If you cannot reliably feel a small problem, you cannot act on one. You act on large ones, because large ones are the only ones that get through.
“I finished the shift. I drove home. I made dinner. It was two days later, when I couldn’t lift the kettle, that I went in and found out it had been broken the whole time. The nurse kept asking how I’d managed. I genuinely didn’t know how to answer that.”
— Autistic adult, HeyASD community
Less sensitive, more sensitive, sometimes both in the same body
Here is where the research gets genuinely complicated, and where a lot of writing about this goes badly wrong. The old assumption, repeated for decades in clinical settings, was that we simply feel less pain. It has already caused real harm, and the evidence does not hold it up.
Moore’s review of the acute pain literature (2015) found the pattern splits by who is doing the reporting: self‑report and clinical observation tended to suggest reduced sensitivity, while direct observation during medical procedures and controlled experimental testing suggested normal or heightened responses5. Watched from outside and asked afterwards, we looked less sensitive. Measured directly, we often were not. The review’s own conclusion is worth quoting to anyone who repeats the old line: contrary to the classical reports, there is no evidence of systematically altered pain responses or thresholds.
The sensory testing work adds a layer. Vaughan and colleagues found higher detection thresholds for light touch, yet also paradoxical heat sensation and mechanical allodynia, where a stimulus that should not hurt does1. Failla and colleagues found no difference in the threshold at which pain begins, but autistic participants rated the same heat as hurting more2.
Both exist, then, and they can coexist in one body on one day. You can be the person who does not notice a fracture and also the person for whom a waistband seam is unbearable. That is not a contradiction. It is the same wiring doing both jobs, the wiring behind clothing sensitivity and sensory overload.
| What it looks like from outside | What is actually going on |
|---|---|
| High pain tolerance | The signal did not arrive on time, so there was nothing to tolerate yet |
| Overreacting to something minor | A genuinely amplified sensation, or a small pain landing on an already loaded nervous system |
| Vague, unhelpful history | You are being asked to locate and rank something your body reports as diffuse |
| Inconsistent from visit to visit | Interoceptive access genuinely fluctuates with fatigue, load and stress |
| Not distressed, so probably not serious | Flat delivery is your baseline, not a measure of severity |
Why “on a scale of one to ten” fails you
The numeric rating scale assumes three things you may not have. It assumes you can detect the sensation clearly. It assumes you can compare it against a remembered internal library of other pains. And it assumes ten means something stable to both you and the person asking.
If you are also a literal thinker, the question gets worse rather than better. Ten is described as the worst pain imaginable, and you can imagine quite a lot, so you will not say ten. Whatever you are in right now is presumably not the worst thing a human body can do, so you round down. You say six. Six gets you paracetamol and a review in a fortnight.
There is an accuracy problem in the other direction too. Some of us give a number that reads as disproportionate, because the sensation genuinely is that intense, and the clinician quietly reweights everything said afterwards. Failla and colleagues found pain‑related anxiety and fear of pain were elevated in autistic participants and positively associated with how intensely pain was rated2. Anxiety about pain does not mean the pain is imagined. It means the whole experience is louder. The scale is not going away, so do not refuse it. Give a number, then immediately give the functional information the number was standing in for.
“He asked me to rate it and I said seven, because seven sounded like a reasonable thing to say. It was not a seven. I had no idea what a seven was. I just knew ten meant screaming, and I was not screaming, so it couldn’t be ten.”
— Autistic adult, HeyASD community
The second problem: you report it, and your face gets in the way
Say the signal does get through, and you make the appointment, which is already significant work: 80% of autistic adults in Doherty’s survey reported difficulty just getting to a GP4. You sit down and describe the thing accurately. And you are not believed.
Clinicians are trained, formally and informally, to read pain off a body: the guarding, the wince, the catch in the voice, the way someone holds the affected part. Those cues are proxies. When your body language does not run the standard protocol, the proxies fail, and your credibility fails with them.
The flat, level delivery reads as “not that bad”. The precise language reads as rehearsed. The eye contact you are managing with effort reads as evasive. And if alexithymia is in the mix, and it often is, you may not be able to attach the right emotional word to what you feel, so you describe mechanics instead. Mechanics without affect sounds like someone reciting, not someone suffering.
There is a cruel second version too. Distress that does arrive can arrive all at once and at full volume, because it has been building unfelt for days, and that reads as disproportionate. Either way the conclusion in the room is the same: this report is unreliable. That is a failure of the reading, not of your reporting. There is more on surviving that room in our guide to why autistic adults avoid the doctor.
Diagnostic overshadowing: when autism explains everything
Diagnostic overshadowing is the name for what happens when an existing diagnosis becomes the default explanation for every new symptom. Once autism is on your file, chest pain can become anxiety, gut pain can become sensory sensitivity, joint pain can become somatisation, and exhaustion can become “well, autistic fatigue”. Sometimes those explanations are partly right. The problem is that they arrive before the investigation rather than after it.
When you are already labelled anxious
Many of us collected an anxiety diagnosis decades before anyone said the word autistic, and that label follows you. It makes a loop that is hard to argue your way out of from the inside: you are anxious, so the pain is anxiety, and when the pain does not resolve, that confirms the anxiety. Add the reasonable distress of not being believed, and you look more anxious than when you walked in. The way out is not to hide it, but to separate the two claims out loud. “I do have an anxiety diagnosis. I also have a symptom that is new, in one place, and changing. I would like that part investigated.” Naming the overshadowing before it happens is not rude. It is precise.
Pain, autism and being read as a woman
The credibility problem stacks. If you are read as a woman, your pain is already less likely to be taken at face value, and if you are also autistic you arrive carrying a lifetime of being mislabelled on top of that. Haid and colleagues asked 52 autistic adults assigned female at birth about the previous 24 hours, and 65% of them were in pain, with high interference across the group3. That is a small study, and it is not a population estimate, but living with pain in this body is clearly not the rare complication it gets treated as. If you were diagnosed late, you may have spent twenty years being handed psychiatric explanations for physical events, and if you are approaching menopause, there is now a second convenient explanation for anything that hurts.
“Eleven years of being told it was stress. When they finally scanned me, the radiographer said out loud, that must have been agony. I cried in the car park afterwards, not because it hurt, but because somebody had finally said the word for me.”
— Autistic adult, HeyASD community
If you have spent years being told your body was fine while your body was quietly reporting otherwise, that is not a small thing to be carrying. The Unmasking Years sits with the misread decades and with the particular grief that arrives when you realise you were accurate the whole time and nobody was reading you.
The bodies this often comes with
None of this happens in isolation from the rest of your health. There is a documented association between autism, ADHD and connective tissue differences, which is why we have a whole guide to autism and hypermobility. Csecs and colleagues assessed 109 neurodivergent adults against 57 comparison participants and found generalised joint hypermobility in 51% of the neurodivergent group against a general population rate of about 20%, along with significantly more musculoskeletal pain, with hypermobility mediating that link6. Hypermobile joints subluxate, strain and inflame more readily, and if your interoceptive reporting is unreliable, you can accumulate a great deal of joint damage without a single clear alarm.
Long‑standing pain is common enough here that it deserves to be taken at face value rather than doubted. Asztély and colleagues (2019) followed up 77 Swedish women diagnosed in childhood with autism, ADHD or both, and 16 to 19 years later 76.6% reported chronic pain7. Only about half that group had autism as their main diagnosis, so read it as a neurodivergence signal rather than an autism‑specific figure. Gut pain, migraine, jaw pain from clenching, pelvic pain and widespread musculoskeletal pain all show up repeatedly.
There is also a loop worth naming honestly. Unmanaged pain wrecks sleep. Wrecked sleep lowers your threshold for everything, so the world gets louder and overwhelm arrives faster. Chronic sensory load and chronic pain together are an efficient route into autistic burnout, and burnout degrades interoceptive access further, so you notice the next problem even later. It is a closed circuit, and it does not open on willpower. None of this is meant to alarm you. It is so you can say accurately in a consulting room that pain here is common, under‑reported and frequently missed, and that you would like to be worked up rather than reassured.
Build a pain baseline before you need it
If you cannot rely on the alarm, build an instrument panel instead. The goal is not to monitor yourself anxiously. It is to gather enough external evidence that you are never forced to answer questions from memory in a room where you are already under load. Keep it small enough that you will actually do it. Three lines a day, at a fixed time, attached to something you already do.
- Location and quality. Where, and what kind. Sharp, dull, burning, gnawing, aching, pressure, electric. One word is enough.
- What it stopped. The functional line, and the most useful one you will write. “Could not carry the washing basket.” “Sat down twice on the stairs.” “Woke at 3am and did not get back to sleep.”
- What changed it. Better with movement, worse with movement, worse at night, better with heat, no change with painkillers.
Over four weeks that produces something no scale can: a trajectory. Trajectory is what moves clinical decisions. “It hurts a lot” is arguable. “Four weeks ago it stopped me once a week, now it stops me daily, and it wakes me” is not. If you already keep a structured journal, add these three lines to the template you have rather than starting a second system. Record your own normal while you are well, too. You cannot describe a change from baseline if nobody wrote the baseline down, and yours will not match the population average.
How to describe pain when the number fails you
Give the number, because the form wants a number, and then hand over the information the number was supposed to encode. Function, frequency, trajectory and interference are all things you can report accurately even when intensity is genuinely unavailable to you. Say what has stopped, how often it happens, and whether it is worse than a fortnight ago. Those three are enough to build a clinical picture on their own.
It also helps to rebuild the scale around your own history rather than an imagined worst. Decide in advance that your one is the background level you always carry and your ten is the worst thing you have personally experienced, then say out loud what that ten was. A scale anchored to something real is one you can use consistently, and consistency across visits is worth more than accuracy in any single visit.
| Instead of a number alone | Say this |
|---|---|
| “About a six.” | “I have stopped using that arm for anything above shoulder height. That started nine days ago.” |
| “It comes and goes.” | “Four episodes last week, each around twenty minutes, all in the evening.” |
| “It’s not that bad.” | “I want to flag that I under‑report pain. Please weight what it is stopping me doing, not how I sound.” |
| “Everywhere, sort of.” | “I cannot localise it. It is worse on the left, worse after eating, and it does not change with position.” |
| “I’m fine, honestly.” | “My face and voice stay flat regardless of pain. That is my baseline and not a severity signal.” |
The third and fifth lines matter more than anything else in that table. Saying out loud that your affect does not track your pain gives a clinician a piece of calibration information they otherwise have no way of getting, and most will use it. It is the same move as asking for instructions in writing. You are not requesting special treatment, you are correcting for a known measurement error.
Body scanning that works with an unreliable channel
Standard body scan instructions assume the signal is there and you are simply not attending to it. If the signal genuinely is faint, being told to notice harder is not a technique, it is a setup for failure. Two things before the alternatives: none of this involves creating physical discomfort to sharpen sensation, and none of it replaces having something looked at. What works is checking by function and external reference rather than by feeling.
- Move it, do not feel it. Take each joint or region through its normal range once a day and note what will not do what it did. Range of movement is observable from outside your own perception.
- Use a body map. A printed outline you shade converts a diffuse sense into a spatial one. Marking a page is a different task from finding a word, and often far easier.
- Use categories, not intensities. Ask “new, worse, same, or better than last week” instead of “how much does it hurt”. Comparison is a task you can usually do. Absolute rating often is not.
- Use objective markers. Photograph swelling, bruising or rashes with a date. Count the stairs before you need to stop. These do not depend on interoception at all.
- Anchor checks to routine. Attach the check to a fixed point already in your day, because a scan you have to remember is one you will not do.
An occupational therapist who works with adults can help you build this properly, and interoceptive work is one of the things they do. Our guide to occupational therapy covers what that looks like.
When to escalate, and what to take into the room
Because your alarm may be quiet, you need rules that do not depend on it firing. These are not judgement calls to make from sensation. They are thresholds you set in advance and then follow, the way you would follow a checklist.
Seek urgent medical care regardless of how the pain feels or how you sound describing it, and do not wait to build a record first, if you have chest pain or tightness, a sudden severe headache, difficulty breathing, abdominal pain with fever or vomiting, a head injury, new confusion, weakness or numbness down one side, difficulty speaking, a limb that will not bear weight after an injury, a fever with a stiff neck or a rash that does not fade under pressure, or bleeding you cannot account for. Add one more that is easy to miss when your signals run quiet: new back pain with numbness around the groin, inner thighs or buttocks, or with any change in bladder or bowel control. That combination needs assessing the same day, not next week. In Australia call 000, in the US 911, in the UK 999.
Book an appointment, rather than waiting to see, when anything has been present more than two weeks, when something is worsening rather than settling, when pain wakes you from sleep, when a symptom sits in one place and stays there, or when a part of your life has quietly disappeared because of it.
Take a single sheet of paper in and hand it over at the start. One page: what the symptom is, when it started, what it stops you doing, what you have already tried, the four‑week trajectory, and one line noting that your affect does not reflect severity. Written information bypasses the verbal performance problem entirely, and it survives your mask slipping into automatic politeness when someone asks how you are. Bring somebody whose job is to say the sentence you cannot, if you can. And if you are sent away and the thing does not resolve, go back. Being sent away once is not a verdict, it is a data point.
Key points
- Pain often reaches you late or as a general wrongness rather than a located signal, which means small problems become large ones before you get the chance to act.
- The research does not support the myth that we do not feel pain; it shows differences in detection, description and belief, and reduced and heightened sensitivity can exist in the same body.
- The one to ten scale assumes clear detection, an internal library of comparisons and a shared meaning for ten, and you may have none of those.
- Flat affect, precise language and no visible distress are routinely misread as evidence that the pain is mild, which is a failure of the reading and not of your reporting.
- A three‑line daily log of location, what it stopped and what changed it produces a trajectory, and trajectory moves clinical decisions in a way that intensity ratings do not.
- Decide your escalation thresholds in advance rather than from sensation, and get urgent care for red‑flag symptoms without waiting to see whether they get worse.
Questions about autism and pain
Do autistic people feel pain differently?
Pain is processed and reported differently, and the direction varies. Reviews of the acute pain literature found that self‑report and outside observation suggested reduced sensitivity, while measured responses during procedures and experimental testing suggested normal or heightened responses. Sensory testing has found higher thresholds for light touch alongside paradoxical responses where something that should not hurt does. Practically, this means you might not register an injury for days and still find a seam or a waistband unbearable. Neither of those is inconsistent. They are two outputs of the same interoceptive and sensory wiring, and both are real.
Is high pain tolerance an autism trait?
What gets recorded as high pain tolerance is usually a delay in detection rather than an ability to endure. Tolerance implies you felt it and chose to continue. If the signal did not reach you until three days later, there was nothing to tolerate at the time. Be careful with the phrase, because once “high pain tolerance” is on your notes it can be used later to justify not investigating something. If a clinician says it to you, a useful reply is that you do not detect pain reliably, which is a different problem and a more concerning one.
Why can't I tell where the pain is coming from?
Localising pain requires reasonably high‑resolution interoceptive input, and if that channel runs low‑resolution for you, pain arrives as a region or a general state rather than a point. This is common and it is not you being unobservant. A body map you shade in is often much easier than a verbal answer, because marking a page uses spatial processing rather than the naming that is failing. Take the marked page in with you. “It is somewhere in this shaded area, worse on the left” is genuinely useful clinical information, and far more accurate than guessing at a point to satisfy the question.
Why does my doctor not believe my pain?
Because clinicians are trained to read pain partly off the body, and the cues they look for are guarding, wincing, a strained voice and visible distress. When your delivery is flat and your language is precise, those cues do not appear, and the unconscious conclusion is that it cannot be that bad. It is rarely deliberate. The most effective correction is to name it at the start of the appointment: your face and voice do not track your pain level, so please weight what it is stopping you doing. Most clinicians will adjust once they are told, because they have no other way of knowing.
What is diagnostic overshadowing in autism?
Diagnostic overshadowing is when an existing diagnosis becomes the automatic explanation for new symptoms, so investigation stops before it starts. With autism on your file, gut pain becomes sensory sensitivity, chest pain becomes anxiety, joint pain becomes somatisation, and exhaustion becomes burnout. Some of those explanations may eventually be right, but they should be the conclusion of a process rather than the opening move. You can pre‑empt it in the room: acknowledge the existing diagnosis, then separate out the new physical symptom, its location, and how it has changed over time, and ask specifically for that part to be investigated.
Is chronic pain more common in autistic adults?
The evidence points that way, though most of it comes from small studies. A comparison of 109 neurodivergent adults with 57 others found significantly more musculoskeletal pain in the neurodivergent group, with joint hypermobility appearing to mediate it. A 16 to 19 year follow‑up of 77 women diagnosed in childhood with autism, ADHD or both found 76.6% reporting chronic pain, and a study of 52 autistic adults assigned female at birth found 65% in pain within the previous 24 hours. Associations with connective tissue differences, gut conditions and migraine come up repeatedly. If you have lived with pain for years and been told it is stress, that history deserves a proper workup rather than another reassurance.
How do I answer the one to ten pain scale as an autistic adult?
Give a number so the form is satisfied, then immediately add the information the number was standing in for. Try: “I will say six, but I under‑report, so here is what it is actually doing.” Then give function, frequency and direction. What have you stopped doing, how often does it happen, and is it worse than a fortnight ago. If it helps, anchor your own scale to your own history rather than to an imagined worst pain: one is the background you always carry, ten is the worst you have personally experienced, and say out loud what that ten was.
Why do I only notice an injury days later?
Interoceptive signals can be delayed, faint, or drowned out by whatever else your nervous system is processing, and the busier the day the more likely they are to be missed entirely. Load matters: on a heavily masked or heavily overstimulating day, internal signals lose to external ones. The signal frequently arrives once you finally stop, which is why so much of this lands in the evening or on a day off. The practical answer is not to try to feel harder, but to run a short functional check at a fixed time each day, so detection does not depend on the signal being loud.
Can autistic burnout make pain worse?
It works in both directions, which is what makes it hard to break. Burnout degrades interoceptive access further, so you detect problems later than usual, and it lowers your threshold for sensory input, so pain that is present feels louder. Meanwhile unmanaged pain destroys sleep, and poor sleep accelerates burnout. If you are in the middle of that loop, treating the pain is not separate from recovering from burnout; it is part of it. Getting pain properly assessed is one of the few interventions that can loosen the circuit from the outside rather than through willpower.
What should I take to a doctor's appointment about pain?
One page, handed over at the start rather than read aloud. Put on it: what the symptom is and where, when it began, what it now stops you doing, how that has changed over the last four weeks, what you have already tried and what happened, any medication you take, and one line stating that your affect does not reflect severity. Add photographs with dates if there is anything visible. Bring a support person if you can, whose only job is to say the sentence you will not say. Ask for what is written down to be added to your notes.
Is it normal to feel pain and not be able to name the feeling?
Yes, and it has a name. Alexithymia, difficulty identifying and describing internal emotional states, is common alongside autism, and it does not stop at emotion. If you cannot easily attach words to feelings, you may also struggle to attach words to sensation, so you end up describing mechanics instead: what moved, what stopped, what it was like structurally. That is a perfectly valid report, and mechanics are often more useful clinically than adjectives. The only risk is that a description without emotional colour sounds detached to the listener, so it is worth saying explicitly that the flatness is not the severity.