Meet My Brain - A Field Guide to Autism

The Predictability Hypothesis

July 22, 2022 The Autistic Woman
Meet My Brain - A Field Guide to Autism
The Predictability Hypothesis
Show Notes Transcript

Scientists think one thing is responsible for autistic traits. Learn whether AI, robots and machine learning are the key to finding answers about autism. Get tips on handling the unpredictable.

Support the Show! Rate and review the podcast. It helps make this content available to more autistics. It only takes a few seconds to rate, less than a minute to review.

 This is the #1 autism resource for late-diagnosed autistic adults. Rated in the top 3% globally.

Slava Ukraine!

Support the show

The Predictability Hypothesis

Welcome to Meet My Brain - A Field Guide to Autism. I’m your host, The Autistic Woman.

Science has now found that the brains of autistics and non-autistics alike are predictability machines. The brain functions by observing the environment and concluding what may happen next. We learn this way. 

Some researchers suggest that autistics have both an impairment in predictive abilities and at the same time have a preference for predictability. Some say autistics have an aversion to things that are not predictable.

What’s going on here? Stay tuned to learn more in this episode.

Many of us are aware that research about autism is often focused on deficits, causes and cures. We find ourselves disagreeing with results on many occasions because the conclusions that are drawn are not our experience.

I recently came across an article written by researchers who suggests that all autistic traits are caused by or based on one thing, predictability. In this episode I’m going to summarize what the researchers said and give you an alternative viewpoint. I welcome your feedback.

Let’s first look at predictability itself which is the ability to know and understand what’s going to happen or come next. 

Many people feel less anxious when they know what to expect and how they’re going to navigate their day to day lives. 

There isn’t much research about predictability in autism and the limited studies that have been done relate to children. 

A few scientists are considering its role in autism based on a 2014 hypothesis that remains unproven.  

So what is the predictability hypothesis? 

Researchers need a hypothesis to be able to do a study, so they take a guess about something. For a hypothesis to be scientific it has to be something that can be tested. Scientists base a hypothesis on earlier observations that can’t be explained with current scientific theories.

Once they have data they can form a theory. It can take years of studies to form a theory. Occasionally we’ve seen an autism researcher make unsupportable conclusions using data from only one study. 

The predictability hypothesis is that virtually all autistic traits, issues, sensitivities are a result of the autistic need for predictability and an aversion to unpredictability.

Yes, that all autistic traits are due to a preference for predictability.

Researchers justify their position by citing a number of examples. 

They suggest that autistics do best at subjects like math because math has rules and is therefore predictable. 

I question whether being good at something with rules automatically proves its opposite, that we are bad at or averse to things without rules.

Autistics are said to like routine. No question routines make things predictable AND efficient and practical.

Some people eat the same food every day. The choice of food and whether a person will like it is somewhat predictable. Consider that most food doesn’t taste exactly the same every time. 

With sensory sensitivities having to eat food we don’t like is difficult. We protect ourselves from negative sensory overload by the food choices we make. Eating good food can be a positive sensory experience. I’m not sure this is just an autistic trait.

The hypothesis is that due to a predictability deficit we autistics inhabit a seemingly “magical” world where events occur unexpectedly and without cause. It sounds to me like they think we’re out of touch with reality.

When is it imagination and when is it magical? And why would either be a deficit and a criteria for autism?

Are autistics truly averse to things that are not predictable? Many autistics love history, writing, literature, music etc. The beauty of art is often found in its unpredictability.

The researchers hypothesize that an autistic’s reaction to loud noises is caused by the brain’s inability to determine the probability of what comes next.

Okay, so no. The reason I hate certain noises is because they literally affect my nervous system and cause pain throughout my body. I can predict that.

The sound of a leaf blower bothers me more and more over time. I know it will.

I’m not kidding when I say this: As support for the hypothesis the researchers reference torture techniques that use annoying, loud music to create stress and anxiety. 

They say the unpredictability of the noise is what makes torture effective. BTW from what I understand torture is rarely if ever successful. What’s more, I can’t believe researchers are even mentioning torture to support their hypothesis on autism.

These researchers say that if we had no aversion to unpredictability our autistic traits would vanish. Noise wouldn’t be a problem nor would any sensory issue and we could suddenly engage in social communication with ease.

Let’s talk about something funny. The hypothesis suggests that the reason autistics don’t understand humor is because it’s based on unpredictability.

We’ve probably all heard this joke: 

Why did the chicken cross the road? 

To get to the other side. 

In fact we may have heard it so often that it isn’t terribly funny any more.

Now let’s make it unpredictable:

Why did the chicken cross the playground?

To get to the other slide.

If you’re smiling it might be because you like corny jokes as much as I do. I’m not alone. I know a number of autistics who enjoy this kind of play on words.

I understand there may be autistics who don’t get jokes but neither do some non-autistics. Otherwise all we would have to do to test for autism is tell a really good joke.

If you’re not tired of the hypothesis yet, believe me very little was overlooked. 

The hypothesis suggests that an aversion to unpredictability is the cause of social communication deficits.

Many autistics know quite a bit about human behavior and theory of mind in part from years of counseling and reading psychology and self-help books.

It’s one reason I can observe and analyze others, in essence I can predict behavior. I’m less able to apply that knowledge to myself when I’m overwhelmed. 

Even if I knew 100% in advance what was going to take place in a social context, in essence if it were 100% predictable, it would not necessarily change my reaction.

What about unexpected questions? You can predict there will be some whether or not you’re autistic. We’re led to believe that only autistics are unable to know in advance how a conversation or job interview will go.

There’s nothing wrong with me if I’m not perfect at conversing yet I feel a pressure that I have to be. On a conscious level it’s that feeling that causes me to script. 

I want to do things well, okay I’ll say it, perfect, which I know is an autistic thing. I’m driven to have the perfect convo so it’s logical to me to rehearse or script in advance. 

I write scripts for the podcast not for predictability. I want to communicate effectively so that someone listening can understand my meaning.

The fact that we can predict our reactions is one reason we isolate. We know what we might feel and that it’s possible that those feelings, sounds, temperatures, tastes, will be more than our nervous system can comfortably handle.

Guessing that there was only one thing that explains all autistic traits, the researchers decided that autism could be diagnosed using artificial intelligence using a small percentage of the criteria defined in the DSM, the Diagnostic and Statistical Manual of Mental Disorders.

The researchers said that since autism currently is diagnosed by behavior AI or artificial intelligence can do it better by removing the human subjective factor. The DSM is highly subjective imo and we know it has a deficit-based bias.

This idea is exciting to these researchers who believe algorithms, artificial intelligence and machine learning can be used to determine if someone is autistic. 

We do know that AI is used to predict and to influence purchasing or clicking behavior. 

If we click on links for sales on lamps we probably are thinking of buying a lamp. If we then zero in on a link for a particular lamp it’s likely we’ll buy that one. 

AI learns the likelihood of future behavior and how to influence it. What we want and how to motivate us to buy becomes predictable. It takes a lot of data to make general and personal predictions about human behavior. 

The researchers decided to test the hypothesis and developed an algorithm. 

I admit I don’t understand how testing a preference for predictable things, assuming you can, proves that a preference is the explanation for symptoms, traits, features of autism.

The researchers built robots to diagnose autism using their algorithm. I’m not kidding. 

Drum roll, here are the results: The robot was correct 1 out of 3 times. Must not have been an autistic robot.

Undeterred they next created a medical device using machine learning, The medical device correctly diagnosed autism in 31% of the cases. 

After repeated studies since 2014 and continuing today that yield poor results it seems that autism can’t be diagnosed by AI.

It’s likely an autistic could have predicted that AI wouldn’t work. We can also accurately predict that the hypothesis has and will create the opportunity for many funded studies.

The need for predictability must apply to non-autistics if we believe that the brain is a predictability machine. 

We’re all taught to make predictions about possible partners or spouses. It may be one way we choose friends.

When we apply for and accept a job we’re predicting it will go well. Most human beings get up in the morning with an expectation of how the day will go. Often these predictions are wrong whether you’re autistic or not.

When we’re faced with a problem, let’s say at work, autistics typically approach it differently than many non-autistics.

A non-autistic might say: We have a problem, we’ve always done it this way, so the solution is let’s keep doing it this way. That’s an aversion to the unpredictable.

An autistic might say: We have a problem, let’s look at it with an open mind, figure out what causes it, and choose something with the potential to solve the problem. We’re not afraid of the unpredictable.

I understand the desire for answers because I seek those myself. I also am well aware that the right answer comes from the right question.

One example suggested that if a person can’t tell what is inside an unlabeled box they have an impairment in predictability. Yet to be valid a thing has to be inherently predictable and guessing what’s in a box is not.

Consider that the very nature of scientific research is predictability!

I’m not suggesting that autistics don’t ever struggle with it. I question it as the way to explain autism.

There was a time when I planned everything. I found myself frustrated and angry when another person got involved and things turned out differently than I anticipated. I would be so emotionally upset by the change in plans that I didn’t enjoy what we were doing.

 One day I found this advice, “take one day and make no decision for yourself.” I happened to have a trip coming up and already planned how each day would go. 

The people I was traveling with had their own agenda so I decided to face my resistance and make that the day I made no decision for myself. Don’t get me wrong, it was HARD!

I started to let others decide. Since I’d promised myself I’d do this, I felt my frustration slowly fade. After all, I chose this, I planned to let others decide.

Something amazing happened. I did things I didn’t expect to do and I had a great time! Sure there were moments when I felt panicky and felt the urge to resist but I fought it. Maybe my belief that I needed to plan was just that — a belief. Maybe even a habit.

Over time I’ve realized that in life very few things, if any, go exactly the way we plan, anticipate, predict or expect. Has anything gone just the way you imagined it? It’s not really likely. Knowing that takes off the pressure.

If you understand in advance that things can’t go precisely the way you imagine then you can let go of the idea that you need them to.

When I went on a trip a few years later we were told at the airport that the flight home was unexpectedly delayed for 5 hours. The airline gave us vouchers for lunch and we smiled and said “no thank you.” The woman at the desk was astonished. At that point it wasn’t a bad thing that the unpredictable happened. 

We’re going to have one more meal at our favorite restaurant about 5 miles from here, we told her. We couldn’t be more excited that our expectations, of leaving at a certain time, had not been met. And the best part was that we were okay with it, in fact, more than okay.

We might convince ourselves we need predictability more than we do. Oh sure, for some things it’s important. There are some we can let go of. It’s a question of how we handle it.


First accept that things hardly ever turn out exactly as planned or imagined.

Second, understand that the unpredictable will happen and know you can handle it. Not always, pick and choose. Start small.

Third, start noticing how many things in your environment are actually unpredictable, like the weather or what you’ll get in the mail or when you might get a flat tire for example.

Fourth, unpredictable things, just like jokes, can be good things.

Finally, know that when the unpredictable happens it’s neither your fault nor your inability to predict it that’s the cause. “You should have known” is baloney.

IMO all humans prefer predictability most of the time. I can’t say what non-autistic motivation is for it nor whether it’s different than mine.

I don’t think a scientifically sound hypothesis would attribute every symptom and trait of autism to a predictability impairment. I don’t believe deficits in social communication are caused by a need for predictability. I think it has to do with brain processing. It may be better explained as a double-empathy problem.

I refer to the problem-solving example I gave earlier. The non-autistic will solve a problem by saying “we’ve always done it that way” because they need predictability even when it makes no sense. 

The autistic can do the unpredictable because we’ve learned that doing it the way it’s always been done does not work.

  Neither does the predictability hypothesis.

I hope you’ve enjoyed this episode. I’m on Twitter at @anautisticwoman or you can email me at Don’t hesitate to contact me. You can support the podcast by subscribing, liking, tweeting, rating, reviewing, telling your friends and becoming a patron on Patreon. I’ll put a link in the show notes. I appreciate hearing from you. Thank you for listening.

This has been Meet My Brain - A Field Guide to Autism, I’m the Autistic Woman. Slava Ukraine!