One in every 59 children is diagnosed with autism spectrum disorder, according to a 2018 report by the Centers for Disease Control and Prevention. Those with autism have difficulty with cognitive empathy, the ability to read emotions in other people. They may have a harder time with social interactions. Testosterone has long been targeted as a possible cause because males are four times more likely to be diagnosed with autism than females. But new research challenges a decade-old theory about that link. The results of the largest study of its kind show no evidence that testosterone reduces cognitive empathy in men. Gideon Nave, Wharton professor of marketing, and Amos Nadler, visiting professor of economics at the University of Toronto, co-authored the study along with Colin F. Camerer of the California Institute of Technology, David Zava of ZRT Laboratory, Triana L. Ortiz and Justin M. Carré of Nipissing University, and Neil V. Watson of Simon Fraser University. Nave and Nadler spoke with Knowledge at Wharton about their paper, which appears in the journal Proceedings of the Royal Society B. (Listen to the podcast at the top of this page.)
An edited transcript of the conversation follows.
Knowledge at Wharton: One theory about autism is tied to something called “extreme male brain.” Can you explain what that is?
Amos Nadler: This was a proposal that testosterone shapes neurodevelopment in the womb. The hypothesis is that there is an empathizing brain, which is considered more female, and there’s a systemizing brain, which is considered more male. If you take the extreme version of systemizing, you end up with a brain that has no empathy at all and is purely systemizing, and that is considered essentially to be autism. The belief is that testosterone underlies that neurodevelopmental trajectory, and we end up with someone who is autistic.
Knowledge at Wharton: What is the importance of your research?
Gideon Nave: I think the main strength of our research is that we’ve built upon previous studies, and we used exactly the same methods that they did. I think it’s important to know that we don’t necessarily agree with everything that was done before. The claim that autistic people do not have empathy is very controversial, and I would definitely not make it. We’ve gotten some emails from autistic people or their parents telling us that they are very happy that we have disproved, or maybe caused some doubt about, this theory because they don’t agree with it.
The strength of this research is that it tests what was done before — very high-impact studies that were published in high-impact journals. We did it on a much larger scale with superior methods, and we showed that this research did not replicate in the larger sample — which will hopefully direct resources in more fruitful directions.
“It’s very important to base our beliefs on solid evidence and, once we have an intriguing research direction, to verify it before we build an entire house of cards on top of it.” –Gideon Nave
Knowledge at Wharton: In the previous studies you mentioned, the subjects numbered in the teens, whereas your study looked at hundreds of subjects.
Nave: Yes, one previous study used females — not males, as ours did — which is a bit problematic because when you give testosterone to females, it almost immediately converts into estrogen. That study used only 16 females … and the authors claimed that giving females testosterone impaired their performance in the experimental task.
Of course, using 16 people gives your study very low statistical power. You cannot correctly estimate the true effects. What’s interesting is that several other studies that were also small basically built upon this study and tried to illuminate the mechanism at work, but nobody noticed that these [subsequent] studies failed to reproduce the original effect. We ended up building a kind of house of cards — just building upon the original result while failing to replicate it.
I think the main issue here is that when you run studies that are very small, you are very vulnerable to influences that are not related to the manipulation that you’re doing. So, we decided to run a very large investigation and maybe solve the puzzle once and for all. And that’s what we did.
Knowledge at Wharton: How important is this study in gaining insights into autism that can lead to better treatment?
Nadler: I think our study makes two major contributions. One is showing that if we believe as a scientific community that there is an effect — that testosterone has a direct link to autism — it’s going to prime us to think of treatments or interventions that might somehow save a kid from becoming autistic. [But] if it’s not true, believing that could be quite dangerous…. There are a bunch of non-obvious consequences of believing in this relationship.
We’re not saying that testosterone has no effect on autism. Small samples can’t show it either way, but we are demonstrating that giving people testosterone does not affect cognitive empathy.
[Second,] if we have a small study that has a splashy result and is published in a high-impact journal, it tends to shift public opinion and research agendas and money from the National Institutes of Health towards certain pathways. What we have shown is that if we do it right the first time, we may not go down a path that’s misleading…. The methodology used [in previous studies] was not particularly strong and may have created some mistakes and misbeliefs that are just unhelpful.
Nave: Yes, we can think of the study that was published in The Lancet about eight years ago about vaccines. There was a study claiming that there is a link between some vaccines and autism, which was disproven. But still to this day, some people are against vaccines and keep citing that study.
“It’s not important to us to support a theory for someone’s name or to get citations. We’re seeking the truth at whatever cost.” –Amos Nadler
I think it’s very important to base our beliefs on solid evidence and, once we have an intriguing research direction, to verify it before we build an entire house of cards on top of it. And this is what we did in this case…. I am very happy that there is a shift in science now and there is an acknowledgement that the large study that shows no effect may be just as important as a small study showing a flashy effect that is not likely to be true.
Knowledge at Wharton: Given the number of children diagnosed with autism, it’s no surprise that your research is getting attention.
Nadler: It has been a rising issue over time, and part of the reason that I think we’re having more diagnoses is because there’s a heightened sensitivity to it. Individuals who may have not been diagnosed previously are now being diagnosed. For example, I have a second cousin who has been diagnosed as having Asperger’s syndrome, but that was not a diagnosis in the era he grew up in. He was finally diagnosed in his 40s or 50s, after going through a lifetime of just being “different” or feeling different, and I’m sure feeling very frustrated.
To say that there is a rise in the incidence of autism could be a mistake. There’s just more sensitivity and more diagnosis of it. There are other conditions that have the same pattern. Unfortunately, this paper doesn’t tell us what does cause it. We wish we could pinpoint the cause, but that’s not the main contribution here.
Nave: Yes, I think one thing we have to embrace is our ignorance on the topic. The fact that we found that giving testosterone does not impair cognitive empathy does not mean that there is no association at all between testosterone and autism. But I think that before we go ahead and make claims of such association, we should see evidence. At the moment, I don’t see evidence. This does not mean that evidence won’t come in the future.
Knowledge at Wharton: Would you say this research is just one small piece of the autism puzzle?
Nadler: Yes, absolutely. The challenge is doing human research, which is to say that you can’t ethically manipulate testosterone in utero in a human. If someone said, “I absolutely believe that testosterone is causing autism, and I’m going to test this,” they are not going to do it by randomized controlled trials giving pregnant women testosterone. That’s completely unethical, and there are no grounds for that. There are animal models, but the manifestation of autism in animals is not an easy analog to humans. Figuring this out is not an easy nut to crack, unfortunately.
“Many human traits have a biological basis, but this biological basis is very complex.” –Gideon Nave
…What we’re trying to do is show the uncertainty completely nakedly. To be frank, neither of us has a dog in the fight. We just care about what’s true, and it’s not important to us to support a theory for someone’s name or to get citations. We’re seeking the truth at whatever cost.
Knowledge at Wharton: What was the original driver for doing this research?
Nave: We know that autism starts very early in life, and that males are disproportionately affected by it. It’s a logical theory. Of course, there are also sex differences in height. There are sex differences in schizophrenia. There are sex differences in many things. It doesn’t mean that all of them are biologically caused by prenatal exposure to testosterone. But it’s a very good candidate to test in any place where there is a sex difference. Testosterone may play a role. It’s clearly a trait where men and women are differing.
I think that what we’ve learned over the past 10 years with larger datasets, especially the proliferation of genetic datasets, is that many human traits have a biological basis, but this biological basis is very complex. There are many contributions by many small things — little influences of many genetic variants. It may be a prenatal environment, maybe development, lots of things. And all of them have small effects maybe eventually drive the risk to show this trait. I think that this is where we are probably going in autism research, and I’m very hopeful that we’re going to find more and more as time progresses.
Maybe we are a bit too harsh judging this original theory [about testosterone], because at the time that the paper came out, this theory might have been the only way to start looking for something. Think of a drunk guy who is looking for his keys under the streetlights. You ask him, “Why are you looking for them under the streetlights?” He says, “Well, that’s where the light is.” You’re not going to find it anywhere else, so that makes sense.
Nadler: There’s a joke in economics and other fields that goes, “Science advances one funeral at a time,” which means to say that the power in a field tends to hold until the senior people pass along. But I think that’s changing, which is very encouraging to everybody because the rate of change is much faster. I think we’re seeing senior researchers more pliable, more open to other approaches, to other theories. It makes me optimistic that the field would have an appetite for null results. I think that’s a positive direction in science overall.
“I think we should celebrate uncertainty. We should communicate our lack of confidence, and we should embrace it. That’s why science exists, because we don’t know everything.” –Gideon Nave
Just look at medical history and some of the methods that were used that may have had intuitive appeal. Leeches were used because there was some logic. Literally poking a hole and having people bleed out to improve their conditions, or removing parts of the brain had some sort of logical basis. But the execution really hurt a lot of people. I’m not likening the extreme male brain hypothesis to puncturing people’s brains for psychosis, but I’m saying that there’s a pattern where humans, at our base nature, don’t like ignorance. We don’t like not knowing. If we have something that possibly explains something, I think we feel better about it than having nothing to explain something. I think there is something very natural and human about science in itself, and this is maybe an example of the sociology of science and how it plays out.
Nave: I think my main takeaway from all of this is more about a ‘history of science’ lesson. I think we should celebrate uncertainty. We should communicate our lack of confidence, and we should embrace it. That’s why science exists, because we don’t know everything. If we knew everything, there wouldn’t be science. Let’s just embrace our ignorance, celebrate it and study it better.
Knowledge at Wharton: What are the next steps for this research?
Nadler: The next steps would be preserving the large samples. Gideon has another large project that he’s working on, and there are some things that I’m working on as well. What I’m taking forward would be to maintain the rigor, to vet the experiment. We pilot the heck out of these things before we run them. We do power analysis on the sample size to make sure that we’re set up for success — and success isn’t that we show a “significant” result.
In terms of autism, specifically, I can’t say that’s right down my professional alleyway or something that I want to continue to work on, because there are other people who know a lot more about the specific complexities. But I think there’s room for more experiments and to rethink the way that we arrive at knowledge creation. And that part, I think, is very, very exciting.
