The complaint usually arrives at a dinner table. Someone puts down a fork, leaves the room, and comes back looking as though they have been in an argument nobody else was having. To everyone still eating, the reaction reads as bad manners, or a short fuse, or a person deciding to make a scene about soup.
For a subset of people, none of those readings fit. They describe something closer to a switch than to irritation building slowly into anger. It arrives ahead of any decision to be annoyed, and it attaches to a specific, often narrow set of sounds. Chewing. Sniffing. A pen clicking in a quiet office. The term for this is misophonia, and in the last ten years it has moved from a term of art among audiologists into a research literature with imaging data, prevalence estimates, and an expert consensus definition behind it.
We are writers, not clinicians. What follows is a reading of that research, not medical or psychological advice.
Where the imaging evidence comes from
The paper most often invoked when people say misophonia shows up on a brain scan is Sukhbinder Kumar and colleagues’ 2017 study in Current Biology, from Newcastle University and UCL. Inside an MRI scanner, 20 adults who described misophonic reactions and 22 age- and sex-matched controls listened to fifteen-second audio clips: trigger sounds, generally unpleasant sounds such as screaming, and neutral ones.
Two things separated the groups. Trigger sounds produced much stronger responses in the anterior insular cortex in the misophonia group, a region involved in registering what deserves attention and in reading the body’s internal signals. The reaction also showed up in the body, with heart rate and skin conductance rising more in that group, and the paper reporting that insular activity mediated those bodily changes.
Unpleasant sounds annoyed everyone; only the triggers produced the particular pattern.
A separate Amsterdam team led by Arjan Schröder, writing in Scientific Reports in 2019, found something comparable in 21 patients and 23 controls, using video clips rather than sound alone. They recorded heightened activity in the insula and in the anterior cingulate cortex, a neighbouring region that responds to conflict and discomfort, alongside raised heart rate.
Talk of seeing the condition on a scan does more work than the imaging can support. Both are group comparisons in samples of tens, and the earlier imaging work ran at similar sizes, with trigger groups of between seven and twenty people. No clinician can put one person in a scanner and read the condition off the pictures, and there is no brain test for it. The imaging still contributes something narrower and more useful. Trigger reactions carry a measurable bodily signature that separates them from ordinary annoyance, a claim people with misophonia have never been able to make on their own authority.
The largest analysis so far landed in February 2026. Heather Hansen and colleagues, writing in Human Brain Mapping, took resting-state scans of 162 adults recruited from the general population for a Welsh database, then scored their misophonia indirectly using a model trained on a separate Oklahoma sample of 777 self-reports. Insula connectivity tracked those scores, and the pattern did not appear when the same scans were sorted by anxiety, depression or autistic traits. Some coverage has reported this as 939 adults in a brain scanner, which adds the two samples together and puts 777 people inside a machine they never entered.
The chewing may not be the point
Kumar’s group followed the 2017 work with a different hypothesis, published in the Journal of Neuroscience in 2021. Most classic triggers come from other people’s mouths and faces. So the team looked at connectivity between auditory and visual areas and the ventral premotor cortex, the strip that controls movement of the mouth and face. They used scans taken while 33 participants lay still and did nothing, plus responses to sound from 42. Those 42 are not an independent sample; they are the 2017 cohort, reanalysed.
Auditory cortex responses to triggers did not differ between the groups. The difference sat in how strongly that mouth-and-face area was linked to the sound areas, and in how much it fired to triggers specifically. In the authors’ reading, the sound is a medium and not a cause, carrying an involuntary mirroring of somebody else’s mouth movement. Your own chewing rarely bothers you because there is no one else’s movement to mirror.
There is an obvious appeal to the model. It also rests on correlational imaging in small samples, within a mirror-neuron framework that remains contested in cognitive neuroscience generally. Nothing in the design establishes that mirroring produces the distress. The finding is worth taking seriously, but it should not be read as the final word.
The 18 per cent figure needs handling
Prevalence work has arrived more recently. Silia Vitoratou, Jane Gregory and colleagues at King’s College London and Oxford surveyed 772 UK adults, mean age 46.4, using a questionnaire they had developed, which scores five separate aspects of the reaction. The cut-off came from interviews with 26 self-identified individuals and 29 people drawn from the survey pool. Their estimate for symptoms causing a significant burden was 18.4 per cent.
Coverage has since turned that into roughly one in five adults having a disorder. The paper makes no disorder-level prevalence claim. Its authors are explicit that misophonia appears relatively common, and that further research is needed to establish at what point it becomes disordered in terms of distress, impact and need for treatment. The 18 per cent is a threshold on a severity scale, calibrated against a small interview sample, in one country.
A different figure in the same survey interests us more. Only 13.6 per cent of respondents had met the word at all, and 2.3 per cent applied it to themselves.
Why the missing diagnostic code matters
Misophonia does not appear in either of the two manuals clinicians use to classify conditions: the American Psychiatric Association’s DSM-5, or the World Health Organization’s ICD-11. Arjan Schröder, Nienke Vulink and Damiaan Denys proposed diagnostic criteria in PLOS ONE in 2013, and argued that misophonia was a distinct condition sitting within the obsessive-compulsive spectrum, on its impulsive side. That placement did not take hold. Later reviewers countered that much more work on diagnostic validity was needed before misophonia could be treated as a category at all, and the consensus definition that eventually arrived does not put it on that spectrum.
The nearest thing to shared ground came in March 2022, when Susan Swedo and sixteen co-authors published a consensus definition in Frontiers in Neuroscience. The definition came out of four rounds of voting by a fifteen-member committee spanning audiology, psychiatry, psychology and neuroscience, with most statements needing 80 per cent agreement to survive. It covers decreased tolerance to specific sounds or the stimuli associated with them. It is a tool for researchers, not an entry in either manual, and the practical difference is that clinicians have nothing standardised to record and patients have limited grounds on which to request accommodation.
What the research does not settle
Anyone hoping to use this research to win a household argument will be disappointed, whichever side they are on. The scans support the idea that the reaction starts on its own, before the person has decided anything. They say nothing about what happens after that, in a kitchen or an open-plan office, and none of the papers try to.
On whether anything helps, the evidence is thinner than the brain research, though it has grown. Inge Jager, Vulink, Denys and colleagues ran the first proper randomised trial of talking therapy for misophonia, published in the journal Depression and Anxiety in 2021. Patients at one Amsterdam clinic were given group cognitive behavioural therapy, and their results were compared against patients put on a waiting list instead. About 37 per cent improved, counting everyone who started rather than only those who finished, and they were still better a year on.
Three more trials have followed. Tamar Podoly and colleagues tested a phone app aimed at angry thinking in 85 people, in the Journal of Affective Disorders in 2025, setting immediate use against a delayed start. Adam Lewin and colleagues treated 43 children and teenagers across two clinics, reported in Behavior Therapy in 2026, weighing a flexible form of CBT against sessions that taught families about the condition and how to relax.
One of them stands out. Michael Twohig’s team at Utah State gave 60 adults one of two treatments: acceptance and commitment therapy, which trains people to sit with an uncomfortable feeling instead of fighting it, or a structured relaxation programme. Earlier in-person trials with adults had measured a treatment against a waiting list. This one measured it against another real treatment.
Both worked about equally well.
That is a more awkward result than a clear win, because it does not show that the first treatment did anything the second one did not. The authors are also blunt that plenty of people in these trials did not get better. Four trials, most of them small, is not much to go on. And the Newcastle finding, that these triggers seem to involve mirroring someone else’s mouth movement, points somewhere else again, toward treating how the brain handles movement instead of sound, which nobody has tested in a trial we have seen.
Anyone whose work, meals or relationships are genuinely being shaped by this will get more from an audiologist, a GP, or a psychologist who knows about sound sensitivity than from a summary of the research. The field is new enough that plenty of clinicians will not have come across it.
The scans, the surveys and the trials have mostly produced a vocabulary. People who spent years being told they were oversensitive now have a word for it, a set of measurements, and a small pile of papers describing a reaction that starts before they have any say in it. All of that is less than a diagnosis, and considerably more than they had.
