What is bottom-up thinking in autism?

The regular way academic work is done, from middle school writing projects right on through the PhD dissertation prospectus, is exceedingly top-down oriented, which supports the way allistic people think, and which means that autistics are likely to do a lot more labor for the same end result.

Curious about this? Keep reading.

What is top-down and bottom-up thinking?

Top-down thinking, oversimplified

• “I create the big picture first, and then fill in the detail as I go.”
• Deductive approach

Bottom-up thinking, oversimplified

• “I collect information and detail I believe to be related, and then later sort it into a meaningful big picture.”
• Inductive approach

It’s a Brain Thing

It is common for autistics to be bottom-up thinkers, collecting information, including adjacent sensory detail that will need to be filtered out later, before a cohesive, shareable story emerges. There are differences between autistic and allistic brains in this regard with autistics having a so-called “connectivity imbalance” as noted in EEG studies in which autistic brains tend to display bottom-up signaling greater than top-down channeled flow (Tarasi et al. 2022; Ursino et al. 2022).

tl;dr autistic brains do autistic things. Those two cited articles are two of many. You can go through their reference lists for more, if you like reading.

Pros and Cons? Yes, of course.

Top-down thinkers have a linear thought process. This is generally efficient but tends to not be as innovative.

Bottom-up thinkers have an associative thought process, and tend to take in much more sensory information. This is generally innovative but tends to not be as efficient.

Does it take longer for a bottom-up thinker to solve a puzzle? Sometimes. Can they solve more complex puzzles and find patterns with elegant sophistication? Sometimes. One type of thinking is not inherently better than the other. Diverse teams are truly superior to those made up of just one type of thinker. But there are indeed pros and cons of each type of thinking.

What’s important to keep in mind is that people cannot change their fundamental learning or processing method. I don’t like phrases like learning style, learning preference, or preferred learning method, because these suggest we have options that we do not have. I have written thousands papers and earned many degrees.

I have taught hundreds of courses and given thousands of lectures. I have yet to be able to become neurotypical and approach a single learning task in a top-down cognitive learning style, because I cannot simply wish myself allistic, if even I wanted to.

I’m reminded of the illustration oft shared in disability justice circles of the little person seeking workplace accommodations and getting responses such as, “We’d love to support you and to provide all reasonable accommodations, but have you even first tried to be taller?”

These are not preferred learning methods or styles. This top-down versus bottom-up approach to learning refers to how our brains acquire new information.

An Example: The Dolphin Outline

When I was in the so-called talented and gifted program in sixth grade, they made us do a lot of writing. This was in 1985, and it was taught exclusively assuming that all of our brains operated with a top-down learning process.

For our first task, we were told to pick a topic to learn about. I chose dolphins. So far so good. Handled that. Then, we had to outline a paper in which we’d later share what we’d learned about our topic. That’s where it became tricky. How can I outline what I’m going to learn before I learn it?

I’m telling you, it’s been a while, and even as I write this article in 2023, I still couldn’t outline a paper about dolphins before I learn today’s science about dolphins. What kind of psychic trickery are these allistic students employing? That not-yet-in-existence paper is invisible to me. How can I outline it?!

They are using top-down processing, as is the TAG teacher and the majority of folks who can do these tasks without struggling.

Here’s what I did in the sixth grade. I asked the TAG teacher if I could learn about the dolphins first. She insisted that I could not do that, because completing this outline step was critical to my success in life. Finding this utterly nonsensical, what I ended up doing was going out and learning about the dolphins in secret, memorizing all the things, writing the paper in my mind, and then going into the little TAG closet armed with my memorized facts, which I could then sort into an outline performatively for the TAG teacher.

Then, she smiled and said I was ready to begin writing the dolphin paper (which I’d already written so that I could produce the outline or summary she was expecting).

In other words, I basically had to write the whole paper first. There is no way I could have written that outline first, neither then nor now. Writing the outline last is actually useful for me, as it is a good editing step to make sure I have filtered out the extra. As a bottom-up thinker, I tend to scoop up too much input, and often need to filter out the extraneous content. But that is definitely not the outline the TAG teacher valued.

Today’s Example: PhD Prospectus

Skipping ahead to today, I have a PhD dissertation prospectus. For the reader unfamiliar with the research-based doctorate, the prospectus is the outline of what the PhD Candidate’s doctoral research will be and will accomplish. The PhD Candidate has completed doctoral coursework and has passed their oral and written exams.

The prospectus has to go through the committee, and get refined and approved, prior to beginning dissertation research.

This is now at the outline step. Bring in the dolphins.

You can see where this is going, right? I will have completed the equivalent of about ten dissertations worth of research by the time I have completed my one PhD degree program. And I have to basically complete my actual dissertation specifically now, and first, so that I can then complete the outline of it and have the prospectus in to satisfy this requirement that the outline come first.

Why must I do things in this order? This does not align with how my brain works. I cannot outline an invisible paper about dolphins, which will exist in the future. My brain is hard-wired to do bottom-up processing. I’m sure this outline step is very useful for those top-down thinkers who work this way. This is so great for them. I’m glad.

For us neurodivergent scholars? It’s a hell of a barrier, and I’d assume this may be the breaking point for most autistics pursuing a PhD.

And it’s unfortunate, really. Taking a step back should reveal that the required outline-first model is not serving the same purpose for everyone. As is true in all social systems, because people are people, PhD programs in the U.S. are not designed for the thriving of all people.

If anyone is looking for a side research project, look into point in time where autistic PhD students drop out before defending their dissertations, and whether it relates to top-down processing requirements such as this prospectus. Please contact me if you have time to do this; I could also collaborate on analysis if you have data.

If you’re a current student struggling or especially if you’re feeling like low grades reflect your character, please read my reflection on grades, offered from the perspective of an autistic veteran professor with some wisdom. I’m expecting that Dear Students write-up on grades to replace the current write-up on bottom-up processing as the most popular thing on this website. If you haven’t read it yet, please do!

Examples Everywhere

It’s the same in all of science. In order to present at the Population Association of America in the Spring of 2023, I had to submit an extended abstract (it’s like the outline of the dolphin paper). Well? I can’t write an extended abstract until I’ve done all the things, so I’ve done more studies than will ever be published, and the labor that is required to write those “abstracts” takes more time, limiting the number of submissions one can make per year. Fewer chances means fewers successes. It does not mean less work, nor less skill or knowledge… but it may look like all that!

If you’re ever hiring an autistic PhD researcher, their CV will have a few publications, but they have likely done way, way, way more research than that. And that’s real experience, just not the kind that earns citation credit and esteem, due to the extra labor it takes for submissions and abstracts, a burden not everyone experiences.

I have so many folders on my drives, filled with rather elegantly programmed Stata do-files and statistical models that nobody will ever see. And the graphics I produced to model the outcomes? They’re beautiful. Google my Stata graphics tutorials; they’re pretty popular actually. None of that is on my CV though…

HR: Hiring an Autistic Researcher?

If you’re ever interviewing an autistic data scientist or researcher, ask them about the most elegant graphical representation of results they have ever produced, published or not, and just see what they show you. (Oh just ask. Why not?)

Autistics: What can you do?

Learn about top-down and bottom-up processing. You can probably figure out whether this applies to you. Most of academia is oriented to support top-down processing. If that’s not you, you’re doing a lot of extra work.

And you might have to continue doing a lot of extra work, but knowing that is what you’re doing can at least give you some structure. You cannot change how you fundamentally learn new information, so if you are a bottom-up thinker then determine what strategies work for you within that learning process and do what you can.

Mostly, try to remember you really are doing #LifeOnHardMode with this stuff. Personally, I have estimated that it can take me 800% of the time it would have taken me to complete a task if I didn’t have to go through all the extra, and backwards, steps that I need to go through to learn and to check the performative boxes on things that intended to assess my learning.

Whether it’s your tenure clock or just some other capitalist, productivity-oriented alleged measure of your value, consider the extent to which it is aligned with neurotypical standards and ways of thinking and living. If you are ever in a position to improve accessibility there? Please do. In my experience, nobody listens to us, hence the compounding, de facto gatekeeping.

Teachers and professors: What can you do?

Learn about top-down and bottom-up processing. You can probably make two versions of assignments and allow students to self-select into the one that makes sense to them. Help students learn about themselves for this, and you’ll be helping them and you.

Bottom-up thinkers: You will start by rummaging through facts and archives and your first goal is to identify main categories or buckets into which to sort your growing collection of facts. Does this sound like the way you want to begin? If so, then choose this adventure.

Top-down thinkers: You will start with a main topic and your first goal will be to come up with sub topics under that main topic, which you will then use to guide your collecting of facts. Does this sound like the way you want to begin? If so, then choose the adventure.

Try to modify assignments that require top-down thinking. As in the “dolphin outline” example described earlier, think about how that could have been a different assignment for bottom-up thinkers like the autistic author you’re currently reading.

If they don’t need the outline, what it is that you actually need in order to satisfy program’s assessment and overseeing objectives really? Can you work that out?

I’m not suggesting that you chuck all your regular assignments, because the dolphin outline is useful for top-down thinkers, which is most students. You might keep it as an option for them. But it’s absolutely useless at that point in the process for bottom-up thinkers, creating a ton of additional burden which is solely performative. And in today’s world, most of the students are going to lean on AI for parts of this kind of work anyway.

You need to rethink what you intend to measure with your assessments, and what you want students to actually learn through your assignments. Think through your assignments for both top-down and bottom-up thinkers, with your learning objectives in mind, and try to do right by students and, in general, remove barriers.

And if you care about accessibility, then you’re probably heading in the right direction. I’m glad you took the time to read this article – from one educator to another, thank you! I know you’re statistically one click away from burnout yourself, and I appreciate you. Hearts and solidarity.

What is inductive versus deductive logic?

I made this graphic a few years ago for undergrad methods students, and I thought I would share it here since it’s showing the same sort of thing. Inductive approach to research uses inductive logic, which is bottom-up thinking.

Deductive approach to research uses deductive reasoning or deductive logic, which is top-down thinking.

Explaining the Infographic, for Accessibility

The inductive approach has a cartoon text bubble which says: “I’ve observed a phenomenon. I don’t know what’s up. Let’s get some data and we’ll go from there to make sense of it.”

The abbreviated example of the inductive research is: It looks like we have more older students this year. I wonder why? I’ll conduct a survey, asking age and why enrolled. From that, I can generalize about why older students are enrolling.

The inductive approach begins with a research question, moves to data collection, then a tentative hypothesis, and ends with new theory or ideas based on this data.

The deductive approach has a cartoon text bubble which says: “Look. We know some stuff already. Let’s come up with a theory-based hypothesis and then test it.”

The abbreviated example of the deductive research is: We know that older students bring more life experiences, and we know a lot of other things about them. Would partnering older undergrads with younger graduate students improve academic performance for either partner? Let’s set this up, test our hypothesis with data, and make a conclusion.”

The deductive approach begins with review of theory, moves to a theory-driven hypothesis, which is then tested with data collection, and conclusions are made (which might refine theory).

All of the above text appears in the image.

If you’re autistic and considering research, one of those approaches is probably going to sound a lot more appealing to you than the other!

Want to discuss this topic?

Go for it! There is this thread on facebook and this post on Bluesky and this post on Mastodon.

References

Tarasi, Luca, Jelena Trajkovic, Stefano Diciotti, Giuseppe di Pellegrino, Francesca Ferri, Mauro Ursino and Vincenzo Romei. 2022. “Predictive Waves in the Autism-Schizophrenia Continuum: A Novel Biobehavioral Model.” Neuroscience & Biobehavioral Reviews 132:1-22. doi: https://doi.org/10.1016/j.neubiorev.2021.11.006.

Ursino, Mauro, Michele Serra, Luca Tarasi, Giulia Ricci, Elisa Magosso and Vincenzo Romei. 2022. “Bottom-up Vs. Top-Down Connectivity Imbalance in Individuals with High-Autistic Traits: An Electroencephalographic Study.” Frontiers in Systems Neuroscience 16. doi: https://doi.org/10.3389/fnsys.2022.932128.