Does Size Matter?

Written by: Manu Schuetze  

Do people with Autism* have larger heads? Do they have larger brains? Or are some parts of the brain larger?

Leo Kanner first gave a clinical description of Autism in 1943. His main observations of atypical communication and interaction have been confirmed in several studies since then and now make up part of the diagnostic criteria for Autism. However, Kanner had also described something for which no clear evidence has been found in studies since then: “larger-than-normal” head circumference in people with Autism. Indeed, in 2014 a research group at the University of Alberta studied over 400 children and came to the conclusion that head size cannot be used as a marker for Autism (1). 

Besides simply looking at the size of a child’s head, brain imaging studies using MRI have tried to understand whether the size of the brain differs between children and adolescents with Autism and their typically developing peers. Since we know that certain parts of the brain are particularly involved in behavioural symptoms of Autism, these studies not only look at the whole brain, but zoom in to test for size and shape differences in these specific parts.

The basal ganglia play a particularly important role in Autism and consist of several brain regions that are located very deep in the centre of the brain. We call these regions subcortical, because they are located underneath the cortex. So far, research studies that tried to understand if and how these subcortical regions differ in size in people with Autism showed inconsistent findings. For example, some studies found that the caudate nucleus (one of the subcortical regions) is larger in people with Autism, while others have found that it is in fact smaller - others found no difference at all.

Why do research studies come to different conclusions?

Let’s take a look at a supposedly simple research question from above: Do people with Autism have larger heads than people who don’t have Autism? All we need to do is to gather a group of people with Autism, measure their heads, gather a group of people who don’t have Autism, measure again and compare the average size from both groups. Sounds easy, doesn’t it? But here are some questions to consider: How do you measure head size exactly? Along which line do you take the circumference? What do you do when people have thick and curly hair? And even if you figure out a way to measure everyone in your group consistently, maybe someone else asked the same question, but measured their groups different (even though they, too, are consistent with their own measure). Then it’s possible that you and that other person end up with different results. In this case, the reason for inconsistent results has to do with the methods that you and the other person chose.

Now, let’s take a look at something else that is often the reason for inconsistent results: the group of people that are studied. Think about your 10 closest friends, or 10 of your colleagues at work of whom you know don’t have Autism. At first it seems as if they would make a good control group because they don’t have Autism. But here are again some questions: Do you include both men and women (men tend to have bigger heads than women)? How do you go about age (children have smaller heads than adults)? And do you include particularly large people who also might have larger heads? Ideally, you need a control group of people who differ from your study group (the people with Autism) in only thing: the fact that they don’t have Autism. How many people do you know who only differ in one thing? While some things probably don’t matter too much for your question (whether or not they like dogs or cats probably doesn’t have to concern you when you look at their head size), other things might be very important to control for. If your group of people with Autism also has other conditions such as ADHD or anxiety (which is often the case) and you do find larger head size in these people, then how do you know whether it’s because of their Autism or ADHD or a combination of both? Again, if someone else asked the same question and measured head size in exactly the same way as you did, but looked at different people (maybe they only included people with Autism who don’t have ADHD, while you only controlled for anxiety), they could come to a different conclusion than you.

With this in mind, we can go back to the subcortical regions that have been linked to behaviours in Autism and think about possible ways to answer the question.

Does the size of subcortical regions differ between people with and without Autism?

First of all, we need to not only look at your 10 closest friends, but at a very large group of people to make sure that they don’t differ in systematic ways. Unfortunately, brain imaging with MRI is very time-intense and expensive and not many research groups usually look at more than 20-40 participants. However, in the Autism Brain Imaging Data Exchange (ABIDE) project, researchers from all over the world can contribute their neuroimaging data to an online database. This database allowed us to analyze neuroimaging data from over 700 boys and men with and without Autism between the age of 7 and 35 (see our other post below on why often more men than women are studied in Autism research). Secondly, we want to make sure that all data contributed from different research groups all over the world are analyzed with the same method. This type of analysis is very time- and labour- intense because it is usually done manually: many images are taken to create one 3D model of one person’s brain and a researcher would have to manually trace each region on each of these images. Doing this for over 700 people would have taken way too long. To overcome this problem, our graduate student went to Montreal and collaborated with a research group at McGill University who had developed a computer algorithm that could do this time- and labour- intense analysis automatically. Our analysis showed that subcortical regions did not differ in size between our groups of people with and without Autism. Given that we used the same method on a large group of people, this is the most compelling evidence to date.

What about shape?

Now, imagine that the size of a region is similar between two people, but the shape developed completely different. Instead of a bulky front part and an elongated end, it could have developed in an opposite pattern with an elongated front part and a bulky end. While the reality is probably much more subtle, looking at the shape of subcortical regions is an exciting and relatively new way of looking at brain structure. If a region develops in a different shape, then it could end up connecting to other brain areas very differently and affect behaviour. Remember that subcortical regions are located very deep in the centre of the brain, making them perfect relay points for information highways in the brain. However, different parts of subcortical regions connect with different areas. For example, the back of the putamen (but not the front) connects to areas that are involved in motor behaviours. If the shape in the back of the putamen is more concave in the group with Autism (either because there might be less neurons or less connections between neurons) then that could suggest a link to impaired motor skills in Autism. And indeed, this is what we found in our study (2). 

Overall, our study resolved inconsistent findings about the size of subcortical regions which are important relays in brain networks related to symptoms in Autism. We were also able to show subtle changes in the shape of these regions which can help us understand in which way networks are working differently in Autism and how that contributes to symptoms in Autism.


*Since 2013, Autism Spectrum Disorder is the correct way to refer to all subtypes of Autism. However, we’re only using the word Autism here for better readability.



(1) Zwaigenbaum, Lonnie, et al. "Early head growth in infants at risk of autism: A baby siblings research consortium study." Journal of the American Academy of Child & Adolescent Psychiatry 53.10 (2014): 1053-1062.

(2) Schuetze, M., Park, M.T., Cho, I.Y., MacMaster, F.P., Chakravarty, M.M., & Bray, S.L. (2016). Morphological Alterations in the Thalamus, Striatum, and Pallidum in Autism Spectrum Disorder. Neuropsychopharmacology.