Aging and cognitive change

Research on aging in autism is recent. Studies of how thinking changes with age have included almost only autistic adults without an intellectual disability.¹ A 2026 meta-analysis of studies of people aged 50 and over found only studies that tested each person once, with few participants over 65.² For how brain structure changes with age, see Brain growth.

Many older autistic adults have never been diagnosed. In English primary-care records from 2018, 2.94% of 10- to 14-year-olds had an autism diagnosis against 0.02% of people aged 70 and over, and the authors estimate that 59 to 72% of autistic people in England are undiagnosed.³ Undiagnosed autistic people can end up in the comparison groups of aging studies, which would shrink any difference.^4,5 Autistic people are also more likely to die early: a meta-analysis of 12 studies found death rates about 2.4 times those of the general population.⁶ In a Swedish register study, the excess was partly moderated by sex and by intellectual ability, so who survives to old age shapes the make-up of older samples.⁷ Study authors suggest this could leave a healthier group among those who reach old age; that has not been tested.^4,5

Most studies find similar age differences in thinking skills in autistic and non-autistic adults. The largest longitudinal study followed 128 autistic and 112 non-autistic adults aged 24 to 85 in Amsterdam, testing them at two or three time points, on average 3.5 years apart. Change over time did not differ between the groups on any of 15 measures. The autistic participants were diagnosed in adulthood and had no intellectual disability, and a few years of follow-up could miss small or late differences.⁸ An earlier study from the same group, testing 236 adults aged 20 to 79 once each, found age differences that were similar or smaller in autistic adults, never larger.⁹ A French study that tested 56 autistic and 106 non-autistic adults aged 19 to 69 once each found similar age effects in most areas; memory for events was lower in older participants only in the non-autistic group.¹⁰

Two studies point the other way. A US study comparing autistic adults aged 30 to 67 testing each person once found larger age differences on executive-function tasks than in non-autistic adults.¹¹ A small longitudinal study of 25 autistic and 25 non-autistic adults aged 40 to 70, which its authors call preliminary, found faster decline on one test of long-term visual memory, though not of short-term visual memory.¹²

A 2025 systematic review of 36 studies, by the same French group, found no consistent pattern. Studies that tested people once mostly fit similar or slower aging in autistic adults; the few longitudinal studies fit similar or faster aging.¹³ A 2026 review of 56 studies found no support for generally faster or slower aging.¹ A 2026 meta-analysis of people aged 50 and over, some identified by autistic traits rather than a diagnosis, found autistic groups scoring slightly lower on processing speed, verbal fluency and working memory, and no differences on other measures; results varied widely between studies.² In an online study from the same London group, 13,390 people over 50 were followed for seven years. The 205 with many autistic traits, not necessarily diagnosed, did not lose spatial working memory faster.¹⁴

One proposal holds that the brain clears waste products through fluid movement that is mostly active during sleep, and that this clearance slows with age and could contribute to cognitive change; the mechanism is studied mostly in animals and through indirect measures, so it is a proposal rather than a finding.³⁰ Early imaging in autistic adults points the same way, but it uses an indirect proxy, and whether that proxy tracks clearance or something else is debated.³¹ In one cross-sectional study of 113 autistic and 90 non-autistic adults aged 18 to 71, autistic adults had lower values on the proxy and lower memory scores, and a second cross-sectional study, of 250 people in a shared dataset, found the same lower values in autistic adults but not in autistic children.^32,33 Neither measured clearance directly or followed anyone over time.

Autistic adults in mid-life and later report many more problems with memory and thinking than non-autistic adults, and the gap is large. On cognitive tests the gap is small and limited to a few skills. Reporting more problems is not the same as declining faster. This comes from a meta-analysis of 15 studies of people aged 50 and over, each testing people once; some counted people with many autistic traits rather than a diagnosis.² In one US sample of 210 autistic middle-aged and older adults, 30% screened positive on a short questionnaire used to screen for dementia. The most common items were losing interest in hobbies and more everyday problems with thinking, memory and judgment. The study had no non-autistic comparison group, and a positive screen is not a diagnosis.¹⁵ Positive screens were no more common in older than in younger participants.¹⁵ In a study by the same research group of 350 autistic and 350 non-autistic adults aged 40 to 83, autistic adults reported more memory problems, and it was the younger autistic adults who reported the most.¹⁶

In the Amsterdam group's sample of 202 autistic and 247 non-autistic adults, depression and being autistic predicted how many problems people reported much better than their test scores did; age did not.¹⁷ In the French study, older participants in both groups reported fewer problems with planning and control, while their test scores were lower.¹⁰

Dementia is diagnosed more often in autistic adults than in other people in US public insurance records. Among Medicaid enrollees aged 30 to 64 (2008 to 2012), 4.04% of autistic adults without a recorded intellectual disability and 5.22% with one had a dementia diagnosis over five years, against 0.97% of enrollees with neither.¹⁹ Among Medicare beneficiaries aged 65 and over, "cognitive disorders" (a category that includes delirium and dementia) were recorded for 25.2% of autistic people against 4.9% of a comparison group matched on age and sex.²⁰ In Medicare records from 2013 to 2021, the same research group found new diagnoses of dementia and related disorders recorded for 25.7% of autistic older adults and 11.1% of matched non-autistic adults over the follow-up.⁴ A third team, with no comparison group, found recorded dementia among autistic adults aged 30 and over rising from 2.09% in 2011 to 8.11% in 2019, with an average age at first diagnosis of 59.²¹ A 2026 systematic review lists higher rates of neurodegenerative disease (a group that includes Parkinson's disease as well as dementia) among its more consistent findings in older autistic adults, though nearly all the studies it reviewed left out people with an intellectual disability.¹

Every cohort of autistic people with a dementia outcome found for this entry used US public insurance records. The authors of one note that this focus on insurance enrollees limits how far the results apply.²² One study using US research health records from about 600,000 adults looked the other way round. People with dementia more often had a neurodevelopmental condition (7.8% against 2.4%), a group that combined autism, ADHD and intellectual disability. For autism alone, the link was not statistically clear.²³

The two main studies disagree on how much of the higher rate goes with intellectual disability. In a matched study, 9,201 autistic and 18,356 non-autistic Medicare beneficiaries aged 65 and over were matched on known dementia risk factors such as high blood pressure, hearing loss, diabetes and depression. The autistic group had a 20% higher risk of a dementia-related diagnosis.⁴ Almost half of the autistic group (47.1%) had an intellectual disability. Their risk was 46% higher; for autistic adults without one, it was 4% higher, with a confidence interval that included no difference.⁴ Because the groups were matched on conditions such as high blood pressure and depression, any part of the risk that runs through those conditions is not counted in these figures.⁴ In Medicaid enrollees under 65, the picture differed: after adjusting for depression and heart-disease risk factors, the risk was about twice as high in autistic adults without an intellectual disability and nearly three times as high with one. People with an intellectual disability but no autism had the highest rate of all, 7.10%.¹⁹ Among autistic older adults in Medicare, cognitive disorders were more common in those with an intellectual disability.²⁵ The linked Medicaid and Medicare sample found the reverse among people over 64 (35% without and 31% with).²²

Other conditions are part of the picture. The gap between autistic and non-autistic adults was widest among people with high blood pressure, and also widened with high cholesterol, depression and type 2 diabetes.⁴ Epilepsy was recorded for 26.4% of autistic Medicare beneficiaries over 65 against 1.9% of the comparison group.²⁰ Neither of the two risk studies adjusted for epilepsy.^4,19 People with Down syndrome, who have a high genetic risk of dementia, were left out of the linked Medicaid and Medicare study, but not out of the matched Medicare study.^22,4

Insurance records show when dementia was diagnosed, which may be well after it began.⁴ The authors of the matched Medicare study suggest that autistic adults without an intellectual disability may notice changes and seek help sooner, while decline in those with one may be recognized later, because of communication differences and existing difficulties.⁴ The authors of the Medicaid study suggest errors could go either way. Overlapping features, such as difficulty with planning, could lead to dementia being diagnosed wrongly; poor access to care, and diagnostic overshadowing, could lead to it being missed.¹⁹

Overlap shows up from the dementia side too. In 142 people with dementia or mild cognitive impairment at a US Alzheimer's research center, none of them diagnosed autistic, higher ratings of autism-like behavior went with more behavioral symptoms of dementia and with more severe dementia.²⁶ In 56 people from the same center who were later autopsied, those with high autism-like ratings before death tended to have more tau tangles, a sign of Alzheimer's-type disease, in the frontal and temporal lobes. The study was small and the differences were borderline.²⁷ A 2025 international consensus report says standard dementia tests are not enough for autistic adults, especially those with an intellectual disability. It recommends assessments adapted to autistic traits, sensory needs and other ways of communicating, and a baseline assessment before any decline.²⁸

In Swedish registers covering everyone born from 1980 to 2013, including 88,343 autistic people, the parents of autistic people were more likely to be diagnosed with dementia (36% higher risk). Most parents were still in their 50s and 60s at the end of follow-up, so this mostly concerns dementia that starts earlier than usual. The rise was smaller in grandparents (8%), and in aunts and uncles it could have been chance.⁵ The link looked stronger for mothers than for fathers, though the ranges of likely values overlapped, and it differed little by whether the autistic person had an intellectual disability.⁵ In a smaller analysis it also looked stronger in families with more than one autistic member, again with overlapping ranges.⁵

The authors could not tell whether shared genes or a shared environment explain the link, though they lean toward genes, and they caution against reading it as a measure of autistic people's own risk.⁵ A genetic study using Mendelian randomization, in people of European ancestry, found little evidence that genetic liability to autism raises the risk of Alzheimer's disease; it did not test other dementias.²⁹

Almost half of autistic older adults in US Medicare records had a fall. In records from 2013 to 2021, 47.4% of 13,732 autistic adults aged 65 and over had a recorded fall, against 31.5% of 25,497 non-autistic adults matched on age, sex, race and health problems. Taking observation time into account, autistic older adults had about twice the odds of any fall (1.99), nearly twice the rate of fall-related injuries (1.93) and about a third more fall-related hospitalizations (1.43).³⁵ A fall appears in the records only when it led to care, and this is one dataset, so the size of the difference may not hold everywhere. The authors suggest coordination differences and the use of antipsychotic medicines as possible reasons.

The result fits with what is known about balance. In two small studies, autistic adults swayed more than non-autistic adults while standing still, and one found that the difference persisted into middle and older age.^36,37 Neither measured falls. No study of fracture rates or of fear of falling in autistic adults was found as of October 2026.

Autistic writers say that aging has been left out of the conversation. Daniel Jones opens a chapter on the future with "Autism doesn't stop at eighteen", and writes that there is plenty of support for autistic children and teenagers "but there's less to none for the older, aging autistic population."^c1 Devon Price writes that what is known about growing older as an autistic person comes from a handful of researchers and clinicians, "but more than that, the efforts of older Autistic adults themselves", who have had to invent their own solutions.^c2

Price also writes about the many older autistic people who were never identified. In Unmasking Autism, Price says older autistic people "never had the opportunity to be assessed" because so little was known when they were children.^c3 Price recounts clinicians describing senior autistic men they see as "clean hoarders": socially isolated, polite, with large collections tied to their interests, and slow to ask for help. Price reads this as a masked form of autism.^c2 Price adds that masking can grow harder to live with over time: "As we grow older, the separation between our true personalities and the false selves we project to the world only widens".^c4

Older writers describe finding out late. In Alex Ratcliffe's collection Our Autistic Lives, Michael learned about being autistic at 68, after someone in an online support group pointed it out: "I realized that while autistic people are often invisible to others, they're also often invisible to themselves."^c6 Harry, 74, was diagnosed at 70 and describes a hard adjustment that ended in a meltdown, an arrest and the loss of a home; Harry writes that it would have been easier with some support for autistic adults, which "just doesn't exist around here."^c7 Pin, another contributor in their seventies, received only an informal opinion of "probably autistic" after a year of therapy; Pin and the psychologist agreed that a formal assessment would cost too much and bring no help or financial support at that age.^c8

On ability, Price writes that "As they age, many Autistic people do appear to experience losses in their abilities", including tiring more easily and finding it harder to mask, while some social skills keep developing.^c2 Orion Kelly describes more periods of autistic burnout with age, as responsibilities and expectations pile up: "As you get older as an Autistic person, you need more time to regulate yourself to avert Autistic burnout."^c9 The research above finds that thinking test scores mostly change in parallel with those of non-autistic adults, while autistic adults report many more problems with memory and thinking than tests show. Daniel Burford, writing for adults diagnosed late, warns that dementia figures can frighten readers, and argues that the raised risk "sits almost entirely with autistic people who also have an intellectual disability"; Burford points instead to treatable conditions such as high blood pressure, depression and diabetes.^c10

Much of the community writing is practical. Price lists ways autistic people adapt: retiring early, moving into shared or multigenerational homes, converting vans or tiny homes, and pooling resources with other disabled people.^c2 Price interviews Vera, a semi-retired autistic woman with ADHD who lives in a van and relies on an online community of van dwellers, and Craig, who moved into the spare room of friends he has known since childhood: "It is very bittersweet to be slowing down, and there are certain things I may never do again."^c2 Price suggests deciding early who will provide company, handle legal matters, act as a medical advocate and count as close family in later life.^c5 Jones suggests similar groundwork, such as a friend or relative who checks finances each month.^c1 Pin names a fear about care itself: "I now worry, though, that when in the care of others, I will be disregarded and not taken seriously when I fear something is wrong."^c8

Many autistic writers describe a fear that skills will slip away with age, often in the words of regression or losing skills. In Alex Ratcliffe's collection, Alice, in her seventies, writes that "with aging comes a slow and steady unravelling of our coping skills", and describes dropping out of a senior group when the social effort stopped working.^c11 Sol Smith, an autistic writer and coach, gives the worry a name of its own, skill loss, and describes autistic adults in burnout who feel their memory, attention and concentration have gone, adding that the skills usually return once the strain eases.^c12

Some of this is well described and some is not. The skill loss these writers describe most often sits inside autistic burnout, which has its own entry and usually eases when demands lift. Whether skills also slip away with age, on their own, is the open question. No study found as of October 2026 has tracked whether autistic adults lose previously acquired skills as they age. The test-based studies above find that thinking skills mostly change in step with those of non-autistic adults, while autistic adults report far more problems than the tests show. One other condition comes up often and has its own entry: catatonia, a treatable condition that can begin with a loss of movement, speech or everyday skills, and can look like autism itself getting worse. Both are reasons to seek assessment rather than to assume any decline is permanent.