The question the research was built to answer — and the answer it found
Abigail Dickinson, Declan Ryan, Judith Carroll and Catherine Lord at UCLA published the first systematic review of ageing in autism in Neuroscience and Biobehavioral Reviews in June 2026, synthesising 56 studies of cognitive, neural, and physical health outcomes in autistic adults across the lifespan. Catherine Lord’s name on the paper matters — she is among the most influential autism researchers alive, a principal architect of the diagnostic instruments the field runs on — and her turning to the question of how autistic people age signals that the question has finally become urgent. The global population of older adults is expanding, the first large cohorts of diagnosed autistic people are reaching later life, and almost nothing is known about what happens to them as they do.
The fear the review was built to test is the one that circulates widely: that autism accelerates ageing, that autistic brains and bodies wear out faster. It is an intuitive worry, and a frightening one for autistic adults and the people who love them.
The review’s central finding is that the evidence does not support it.
Across most objective cognitive measures, autistic and non-autistic adults showed comparable age-related patterns — memory, processing, executive function declining on roughly the same trajectory with age in both groups. The neuroimaging findings were similarly mixed and mostly null: most studies found no structural or functional difference in how autistic brains aged, with only some evidence of vulnerability in white matter microstructure. There is no signature of globally accelerated ageing in the autistic brain.
This is a genuinely reassuring finding, and it deserves to be stated plainly before the critique begins: if you are autistic, the research does not show that your brain is ageing faster than anyone else’s. The “accelerated ageing” narrative, like a great many frightening narratives about autism, turns out not to be supported when someone actually assembles the evidence. What the review found instead was not generalised decline but something more specific and, in its way, more damning — not a problem inside the autistic body, but a pattern in how the world around that body responds to it as it ages. The vulnerability is real. It is just not located where the fear assumed.
The review reframes the question accordingly. Rather than asking whether autistic people age faster, it identifies targeted vulnerabilities — specific conditions that do show up at elevated rates, against a background of otherwise typical ageing. Elevated rates of neurodegenerative disease. Sensory impairments. Musculoskeletal conditions. And, most tellingly, a cardiovascular picture that does not look like simple accelerated decline at all. It looks like something the healthcare system is doing, rather than something the autistic body is doing.
The danger isn't decline — it's detection
The cardiovascular findings are where the review stops being reassuring and starts being an indictment. Autistic adults showed lower recorded rates of hypertension than non-autistic adults — but higher rates of the severe outcomes hypertension causes: heart failure and stroke. Read those two findings together and the contradiction resolves into something specific. If the dangerous downstream events are more common but the upstream condition that causes them is recorded as less common, the condition is not actually less common. It is less detected. The hypertension is there. It is just not being caught, monitored, and treated early — so it proceeds, unmanaged, to the heart failure and the stroke that finally bring the autistic patient to medical attention at the severe end.
The review’s authors name this directly: the pattern suggests gaps in early detection and preventive care. That is the clinical apparatus failing autistic patients at the front end of their healthcare — the routine blood-pressure check, the preventive conversation, the early prescription, the ongoing monitoring that catches a manageable condition before it becomes a catastrophic one. These are exactly the encounters that depend on a neurodivergent patient being able to navigate a healthcare system built around neurotypical communication, sensory tolerance, and self-advocacy — the booking, the waiting room, the time-pressured appointment, the describing of symptoms to a clinician who may read autistic presentation as something other than what it is, et cetera.
The structural reading is one I’ve made across other domains: the apparatus catches a surface and misses what produces it. Here the missed cause is not located in the autistic person’s biology at all — their cardiovascular system is not failing faster. What is failing is the system’s capacity to detect and prevent in a patient it was not designed to serve. The autistic adult arrives at heart failure not because autism accelerated their heart disease but because the healthcare structure around them did not catch the hypertension that a non-autistic patient’s structure would have caught years earlier. The death-adjacent outcome is real. Its cause is a detection gap, not a decline. And a detection gap is a property of the system, not of the patient.
This is the difference between the two readings that matters most. The “accelerated ageing” story locates the danger inside the autistic body and implies there is little to be done but watch it decline. The detection-gap story locates the danger in the healthcare structure and implies that the danger is preventable — that the heart failure and the stroke are not the inevitable arc of an autistic body but the consequence of a system that stopped seeing the patient early enough. One story is fatalism. The other is an actionable failure with a name and a remedy.
The complaints the tests can't hear
There is a second finding in the review that pulls in the same direction, and it concerns the gap between what the tests measured and what autistic adults reported. On objective cognitive measures, autistic adults aged comparably to everyone else. But subjective cognitive complaints — autistic adults’ own reports of their cognition declining, of memory and focus and mental sharpness slipping — were consistently elevated. Autistic people reported their minds ageing worse than the tests said they were.
The standard interpretation of a subjective–objective gap like this is to trust the tests and discount the reports: the person feels worse than they objectively are, so the feeling is the error to be managed. But that interpretation assumes the objective test is measuring the thing that matters, and there is good reason to doubt it here. The standardised cognitive battery measures performance under controlled conditions on defined tasks. It does not measure the lived cognitive load of navigating a demanding, under-accommodating world with an ageing autistic nervous system — the accumulating exhaustion of masking, the cost of sensory environments that do not get easier with age, the executive burden of a life that was always more effortful and does not become less so.
It is at least as plausible that the autistic adults are reporting something real that the test cannot capture as that they are misperceiving a cognition the test has correctly measured. The subjective complaint may be the more accurate instrument — picking up a genuine decline in functional capacity that the decontextualised laboratory task is structurally unable to detect, because the task strips away exactly the environmental demands that make autistic cognition costly in the first place. My corpus has argued that objective measures calibrated to a neurotypical default routinely miss what they were not built to see. A cognitive test that finds “normal ageing” while the person living that ageing reports otherwise is a candidate instance — the measure capturing what it can measure, and missing the lived reality it was never designed to register.
The honest position is that the review cannot resolve which reading is correct, and neither can I. But the dismissive interpretation — that autistic adults simply over-report — is the one that requires the most assumptions and serves the apparatus best, because it converts a possible measurement failure into a patient perception problem. The more careful position holds the gap open and treats the subjective report as data, not noise.
The people the research forgot
The review’s most important limitation is one the authors foreground rather than bury, and it reframes everything above. The entire literature on ageing in autism — all 56 studies — almost exclusively involves autistic adults without intellectual disability. The autistic people with the highest support needs, who often have co-occurring intellectual disability, who may be non-speaking, who depend most heavily on the systems that age alongside them — are essentially absent from the evidence base. The research that exists describes the most able, and generalises to no one else.
This is not a minor sampling caveat. It means that everything the field currently knows about how autistic people age is drawn from the subset of autistic people best able to participate in research — to consent, to complete the cognitive batteries, to attend the appointments, to self-report. The population for whom ageing poses the greatest risks, and who can least navigate the detection gaps the review identifies, is the population about whom there is almost no data at all. The cardiovascular detection gap is likely worse for them. The healthcare-navigation barriers are certainly higher. And the research that might document it, and force something to change, has not been done.
The pattern that runs through this review is the same one that runs through the broader field, and it is worth naming as the through-line: the danger to ageing autistic people is concentrated not in their biology but in the structures around them — the healthcare system that detects too late, the cognitive tests that measure the wrong thing, the research base that studies only the most able and leaves the rest invisible. None of these is a fact about autistic ageing. All of them are facts about how the apparatus around autistic ageing is built. Dickinson and colleagues have done the field a real service by assembling the evidence and refusing the easy frightening story — but the story they found in its place is, in its quiet way, harder. Autistic people are not ageing faster. They are ageing inside systems that were never built to see them, and some of them are dying of it.
The reassurance and the indictment are the same finding. The autistic body ages much like any other. The world around it does not treat it that way — and the gap between those two facts is where the harm lives, caught too late, measured wrong, and for most autistic people, not yet studied at all.
Citations
Dickinson, A., Ryan, D., Carroll, J. E. & Lord, C. (2026) — Aging in autism: A systematic review of cognitive, neural, and physical health findings — Neuroscience and Biobehavioral Reviews
Croen, L. A., Zerbo, O., Qian, Y. et al. (2015) — The health status of adults on the autism spectrum — Autism
Mason, D., Stewart, G. R., Capp, S. J. & Happé, F. (2022) — Older Age Autism Research: A Rapidly Growing Field, but Still a Long Way to Go — Autism in Adulthood
Weir, E., Allison, C. & Baron-Cohen, S. (2021) — The sleep and physical health of autistic adults — Autism Research
