Diagnosed at 35,
burnt out long
before — what
autistic burnout
looks like in a
country that still
half-believes
Bettelheim

The same collapse — what 67 French adults described

Autistic burnout was named by the community long before research formalised it — a history this corpus has documented in the community-first concepts that would all have been coded “misleading” — and the formal literature that followed has been almost entirely Anglophone: Raymaker’s American participants, Higgins and Arnold’s Australian ones. A new preprint from Capucine Rebours and Jeanne Kruck at Université Toulouse Jean Jaurès takes the construct somewhere its institutional assumptions don’t hold: France.

Sixty-seven autistic adults — 48 women, 11 men, 8 non-binary or gender-diverse people, mean age 37.7 — were interviewed at length about what led to their burnout, with the first author herself an autistic researcher. The contributor map will be familiar to anyone who has read the five-stage model of autistic burnout: 70% described a cumulative build-up of demands across life domains rather than any single trigger; 69% cited persistent sensory overload with no environmental relief; 69% named camouflaging; 67% pointed to life transitions — childbirth, job changes, bereavement — landing on an already depleted system. The transitions were not necessarily negative. It was the adaptation that cost, even to welcome change: micro-disruptions, schedule changes and administrative churn functioning as tipping points inside an equilibrium already running on debt.

Underneath sit the mechanisms this corpus has traced separately: sensory processing load that never habituates on demand; monotropic attention taxed by exactly the fragmented, non-core demands participants kept naming — the teacher who loves teaching and is destroyed by everything around it; and 27% reporting that interoceptive or alexithymic differences meant exhaustion wasn’t detected until past the point of intervention, a dynamic the study reframes as an environmental design failure — settings that assume self-monitoring instead of providing rest — rather than a personal one. The insomnia-exhaustion loop is in there too, among the co-occurring strains 28% said raised the baseline load.

The model that emerges is chronic energy debt: environments that misfit autistic functioning drain resources continuously, camouflaging conceals the drain, systems fail to replenish, and then one more demand arrives that the account cannot cover. Collapse, in this frame, is arithmetic.

It is worth saying what that arithmetic looks like one level down, because there is a fifty-year-old model that fits these accounts better than anything in the burnout literature itself. W. Ross Ashby, the psychiatrist-cybernetician, described what a system does when a disturbance is too large for it simply to recover from: it reorganises, rewiring its own internal arrangement until it finds a configuration that works in the changed world. He called this ultrastability, and it is the deepest competence a living system has. But his model carried an assumption he never had to state in the 1950s, because his decade supplied it for free — that disturbances arrive occasionally, with quiet between them, and the reorganisation completes in the pause.

Now read what these 67 people actually described. Not one catastrophic disturbance but continuous low-grade demand: administrative churn, schedule changes, social niceties, sensory input with no relief at home or work, adaptation required even for welcome change. Seventy per cent said the accumulation mattered more than any trigger. That is a system given no pause in which to settle — reorganising continuously, mistaking each new arrangement for the one that will finally hold, and never arriving, because the interval in which arriving happens has been designed out of the environment. The exhaustion is not a failure of the reorganising. The reorganising is the most sophisticated adaptive capacity evolution ever built, running as a permanent emergency. What is missing is the quiet.

The national accent — Lacan, Bettelheim, and a 446-day wait

What makes this study more than replication is the institutional setting. France conceptualised autism for half a century through psychoanalysis — Lacan, Dolto, and Bettelheim’s refrigerator-mother thesis, which framed autism as childhood psychosis born of maternal failure. The Haute Autorité de Santé formally classified psychoanalysis as not recommended for autistic children only in 2026 — this year — and the paper’s authors note that the paradigm persists in university curricula, media discourse and some clinical practice, with some practitioners still treating adult autism as non-existent, or autism itself as curable. Participants met clinicians described as clueless about the condition; prior research the paper draws on records clinicians refusing to acknowledge a diagnosis they did not believe in.

The numbers carry the legacy. Mean age at diagnosis in this sample: 35.2 years — years later than even the late-identified cohorts in the UK’s silent bloom of adult diagnosis. National diagnostic waits averaged 446 days in 2016; an Ipsos survey found 56% of autistic adults had not even initiated state support applications because of procedural burden; diagnostic centres typically provide no post-diagnostic follow-up at all. And the gendered maths compounds it, in a sample that was 86% women and gender-diverse people. One participant put the whole diagnostic apparatus in a sentence: her brother was identified in childhood, while she — who liked “horses, not trains” — was diagnosed at 33, after her burnout. The criteria did the sorting, exactly as the evidence on structurally under-diagnosed girls and women, equal prevalence hidden by male-calibrated instruments and the extreme-male-brain theory’s collapse predicts. In France the male default was reinforced by a second layer: a theoretical tradition that, when it looked at an autistic child, saw the mother.

This is demand-structure-as-variable run as a natural experiment. The substrate is the same on both sides of the Channel. The mechanisms of depletion are the same. What differs is the institutional configuration doing the depleting — which theories trained the clinicians, how long the queue is, whether the law’s promised accommodations exist anywhere but the law. France’s 2005 Disability Rights Act mandates adjustments; participants described workplaces where none materialised even after explicit requests, and disclosure operating as a double-bind — infantilisation if you tell, no accommodations if you don’t. The round hole, it turns out, has national accents. Psychiatry has at least occasionally audited its own founding assumptions; French autism practice is doing so decades late, with the bill itemised here in ruined careers and diagnoses at 35.

Camouflaging as compulsory — and a discrimination this corpus owes

Sixty-nine percent named camouflaging as central to their burnout, describing it as a necessary evil rather than a choice — suppressing stimming, running rehearsed scripts, monitoring faces on constant alert. And camouflaging came with a cruel amplifier: it worked. Participants who masked well were seen as too capable to need help, so the strategy that managed the environment also destroyed the evidence that management was costing anything — the same invisibility economics documented in twice-exceptional adults who camouflage at clinical levels and stay unseen.

This corpus owes a discrimination here, held in the open. The Directory has argued that masking is situational navigation, not an exhaustion symptom — a skilled, context-read response, not itself pathology. This study’s participants describe camouflaging as a central driver of collapse. Both are true, and the reconciliation is in the duty cycle, not the strategy. Navigation is costly and sustainable when it is situational — deployed, then set down. What these accounts describe is navigation with no off-duty: French institutional environments demanded it across work, healthcare, family and administration simultaneously, with no accommodated spaces in which it could be dropped. A strategy is not an exhaustion symptom; a demand structure that requires the strategy continuously converts it into one. The pathology sits in the duty cycle the environment sets — as the wider evidence that sustained masking worsens outcomes also shows — and the corpus position survives with a sharpened edge: defending masking as competence never meant pretending competence is free at 100% utilisation.

A preprint, read as one

Provenance, stated plainly. This is a preprint — not yet peer-reviewed, carrying no journal and no DOI at the time of writing — and it should be weighted accordingly. The sample is late-diagnosed, verbally fluent and highly educated (58% hold a Master’s or higher); 18% self-identified without formal diagnosis; accounts are retrospective; the coding was done by one researcher with lived experience, audited by a supervisor, without formal inter-rater reliability or member checking. The authors declare all of this themselves, and the percentages are prevalence counts within a self-selected sample, not effect sizes. Nothing here generalises to autistic people with intellectual disability or limited speech, whose burnout pathways remain almost entirely unstudied.

Read within those limits, the study’s structural contribution stands. Burnout research built in Anglophone systems risks mistaking its local institutions for the human condition; this is the first substantial look at the construct inside a system with a different pathological history, and the finding is that the mechanisms travel while the pathways localise. That is precisely what an environmental account of autistic collapse predicts and what an individual-deficit account cannot explain — because the deficits did not change at the border, and the outcomes did. If burnout were a property of autistic people, France would have the same burnout with the same fingerprints. Instead it has the same burnout with French fingerprints: later diagnosis, psychoanalytic residue, symbolic rights, absent follow-up. The variable is the system. It always was.

Which raises the question of whether “autistic burnout” will survive as a term, and here the history is not encouraging. In 1869 George Miller Beard named neurasthenia — weakness of the nerves, the nervous system overdrawn — and was explicit about the cause: the telegraph, the railways, the press, the speed of business, a world arriving faster than a nervous system could process it. For fifty years it was everywhere. Then it died, completely, and the reason it died is the reason worth attending to: it never acquired an apparatus. No school category ran on it, no benefits system asked for it, no gateway assessed against it. Words that locate the fault in what the world is doing to a person do not get hired, because a fault filed inside a person produces a fileable person and a billable pathway, and a fault filed in the environment produces neither.

“Autistic burnout” is currently a community word doing exactly what neurasthenia did — naming the strain and pointing at its source. It has no diagnostic code, no assessment instrument, no commissioned service. That is its integrity and also its precarity. The likeliest futures are that it stays uncommissioned and stays honest, or that it is taken up, given criteria, and quietly relocated into the person, at which point the environmental half of its meaning will drop off and nobody will notice the loss within a generation. On the evidence of every word this field has minted, I would not bet on the first. Which is the strongest argument for reading a study like this now, while the construct still says what its coiners meant — and for every jurisdiction currently capping, delaying and rationing its own diagnostic pathways to read a mean diagnosis age of 35.2 not as a French curiosity, but as a preview of the invoice.

Citations

Rebours, C. & Kruck, J. (2026) — Pathways to Autistic Burnout: A Qualitative Study of French Autistic Adults’ Experiences — preprint, not yet peer-reviewed; no DOI at time of writing

Raymaker, D. M. et al. (2020) — “Having All of Your Internal Resources Exhausted Beyond Measure and Being Left with No Clean-Up Crew”: Defining Autistic Burnout — Autism in Adulthood

Higgins, J. M., Arnold, S. R., Weise, J., Pellicano, E. & Trollor, J. N. (2021) — Defining autistic burnout through experts by lived experience — Autism