Peter Allen Mann

Autistic Regression in Adults

Losing skills you had last year

In March, I could make a phone call without staging it. By October, I was writing down what I was going to say first, and then not making the call.

Nothing had happened to my competence. I was running a company the whole time, and the work held. What had gone was a set of small automatic things that had been automatic for four decades, and their going was far more frightening than the tiredness that came with them.

The word people reach for is regression, and it is a bad word for this. It is also the word everyone types into a search bar, so it is the word this essay uses.

Before anything else: I am not a clinician. I am an operator with a late diagnosis and thirty-seven years in leadership. What follows is published research plus my own accounting.

What does autistic regression look like in adults?

In adults, it usually looks like the loss of specific everyday capacities that were previously automatic, while the ability to do difficult professional work is still intact. Speech, phone calls, cooking, driving a familiar route, following a conversation with three people in it. The pattern is that the hard things survive and the easy things stop, which is the reverse of what anyone expects.

That reversal is why it goes unreported for so long. No one asks a competent adult whether they can still use a telephone.

It is not a general decline and it is not evenly distributed. It arrives in named capacities, and for most people the same capacities go each time, in roughly the same order. Mine are speech and unstructured conversation, and speech goes first.

Skill loss is in the definition, not an extra

When Raymaker and colleagues published the first formal definition of autistic burnout in 2020, loss of skills sat in it alongside chronic exhaustion and reduced tolerance to stimulus. It is one of three defining features rather than a complication or a downstream effect. The definition was built from interviews with autistic adults and from accounts those adults had written in public before any researcher asked.

Higgins and colleagues found the same thing in 2021 through a grounded Delphi study, iterating a panel of autistic adults toward consensus. Loss of function again, named by people with no reason to name it except that it was happening to them.

So, the first useful thing to know is that this is documented. It has a literature, it has a construct, and it is not you failing at burnout in some unusual private way.

The community had the description long before the journals did. The word regression was in use among autistic adults writing to each other, and the researchers reached it by going and reading what those adults had already written down. The insight is theirs. What I can add is the vantage point, which is that I was watching it happen from a chief executive’s chair while nobody around me noticed a thing.

Why this frightens people more than exhaustion does

Exhaustion reads as temporary. A skill you have lost reads as permanent, and the mind supplies a story about decline before you have finished noticing. That story is what sends people to a search bar at eleven at night, and the search results it returns are written mostly about children.

Exhaustion is a quantity. You have less of something you had before, and the math of that is legible.

A lost skill is categorical. Last year the thing worked and this year it does not, and there is no obvious dial. It feels like losing a piece of yourself rather than running low on fuel, and it arrives with the particular terror of a person doing the arithmetic forward: if this went, what goes next, and when.

I will not talk you out of the fear. I will point out that it is doing its reasoning on a bad model, because a capacity that is offline under load is not a capacity that has been deleted.

What this is not

Three things, and the third is the one to take to a doctor.

It is not the regression the clinical literature describes. Search for autistic regression and a good deal of what comes back is about a different phenomenon in a different population. The most-cited paper on late regression in autism is Ghaziuddin’s 2021 brief report in Frontiers in Psychiatry, and it studied twenty adolescents with a mean age of fifteen and a half. Its central finding is that catatonia was the most common comorbid diagnosis in that group. That is a real clinical phenomenon and it is not what a forty-five-year-old means when they say they have lost their words after a bad quarter. The vocabulary for this was built for someone else and handed to adults who have nothing better.

It is not evidence that you were faking the skill before. This is the one people say out loud least and believe most. The reasoning goes: if it can go away, it was never really mine. What is going away is the capacity to run a process that was always costing more than anyone could see, and the cost is why it stopped. Not the ownership.

It is not automatically burnout. Skill loss in adults also belongs to thyroid conditions, sleep disorders, depression, medication effects, neurological conditions, and the catatonia literature above. I cannot tell you which of those you are in, and no article can. If this is new, if it is getting worse, or if it came without a period of sustained load in front of it, that is a clinician’s question.

What is skill regression, and why am I experiencing it

Skill regression in autistic adults is the loss of previously automatic capacities during a period of sustained mismatch between what an environment demands and what supports exist. It is a documented feature of autistic burnout rather than a separate condition. The mechanism people describe is load: the capacities that go first are the ones that were being run manually, at cost, for years.

That last clause is the whole answer to the second half of the question, and it took me a long time to arrive at it.

A conversation with three people in it is not effortless for me and never was. It was being computed, in real time, under a process I had no name for. For four decades the process ran and the output was fine, which is exactly the trap: as long as the output holds, nobody asks, including you. When the load rose past what the process could carry, the process stopped, and what I noticed was the missing output.

So, the skill did not degrade. The subsidy did.

If the question underneath yours is whether this is autistic burnout at all or the ordinary kind, ordinary burnout does not do this, and that comparison is its own essay.

What came back, and what did not

Most of it came back. Speech came back first. Phone calls took longer and are still worse than they were in my thirties. Unstructured group conversation has not returned to where it was, and I have stopped expecting it to.

I want to be careful here, because the honest version of this ending is partial and the tempting version is not.

The tempting version says it all comes back once you rest. It may not, and telling you otherwise would set you up to read an incomplete recovery as a personal failure, which is the same error that got most of us here. Some capacity returns. Some returns in a reduced form. Some of what I lost I have simply designed around, which is a real outcome and not a consolation prize. On why there is no sequence to follow here at all, recovery is not a protocol.

Designing around it is the part nobody sells, because there is no product in it. I schedule the calls I used to make spontaneously. I ask for agendas. I put things in writing that I used to settle in a hallway. Every one of those was a workaround for me first and turned out to make the room work better for a lot of people who were never going to say why.

What to do with a partial answer

Write down which capacities go, in what order, and how long each one takes to come back. Do it across months. That record is the single most useful object you can own here, for three reasons.

It tells you what your own early warning is, which is worth more than any general list of symptoms.

It converts a terrifying categorical event into something with a shape and a history, and a thing with a history is a thing you can reason about.

And it is what you take to a clinician, who will get further in ten minutes with your record than with your recollection.

A skill that goes offline under load has not been deleted. It has been repossessed, and what you are paying is the arrears.

Sources

Peter Allen Mann is the author of The Perception Revolution (Fast Company Press, 2026). He spent thirty-seven years in leadership, including four years as a Navy Surface Warfare Officer and two companies founded and run, and was diagnosed autistic at fifty-five. He writes for executives and system designers at SHIELD Perception.