A Cognitive Neuroscience Approach to the Early Identification of Autism
In Plain English:
A scientist investigates the patterns of neural wiring in infants whose older siblings have autism
What it covered:
Dr. Charles “Chuck” Nelson is one of the best known (and judging from the way he was introduced and addressed at this colloquium, he’s also one of the best-liked and most-respected) researchers in the field of neurological development. Before coming to Boston Children’s Hospital, he made a name for himself by working on face recognition in infants.
He stopped by the MIT Simons Center for the Social Brain colloquium to tell other researchers about his team’s latest findings in neurological development in autistic infants (and their siblings).
He prefaced his talk by saying that he was really torn about whether this talk should focus on the more mechanistic aspects of his work (“Which neurons are firing?” & “What neurotransmitters are making them do that?” type questions) or the more descriptive aspects (questions about overall statistical trends in “at-risk” populations) of his work.
Dr. Nelson carefully pointed out out that autism (as defined by the DSM) is a set of behaviors; it’s very possible that there are several different physical causes for autistic behavior. For instance, many people with genetic disorders like Fragile X Syndrome, Down Syndrome, 16-P, and Duchenne’s muscular dystrophy are also diagnosed with autism, but it’s possible that different disorders are causing the same behaviors via different routes through the brain.
Dr. Nelson heads up a large multi-institutional study (with 18+ sites including Boston Children’s and BU) that tracks the neurological development of younger siblings of autistic children. Infants who have at least one older sibling with autism are considered to be a “high risk” group for autism, because they’re statistically more likely to develop autism than infants in the general population.
Even though we tend to spend a lot of time talking about the differences between autistic and “normal” children, we don’t have very good definitions of what makes these children different at a neurological level.
Another problem with our current understanding of autism is that we can’t diagnose it until the children are old enough to exhibit severe social impairments (usually 2 years old at the earliest), and by that point, many aspects of the child’s neurological wiring are pretty fixed.
The goal of Dr. Nelson’s study was to track infants’ neurological wiring through the first three years of life and see if there were any neurological variations that distinguished the children who develop autism from those who don’t.
His teams performed brain scans on the same set of infants at intervals of three months through the first year of life (3, 6, 9, and 12 months old) and at 18, 24, 36 months old. At the end of the study, children were given an official diagnosis of autistic or not autistic. Then Nelson and his colleagues ran stats comparing three different populations: the “low risk” kids (the control group of children whose older siblings were allistic or not autistic), the “high risk” kids who developed autism, and the “high risk” kids who didn’t develop autism.
And some of his results were pretty mind-boggling.
Nelson’s teams used EEG to measure repetitive patterns in the babies’ electrical synapses. These patterns are called neural oscillations (or “brain waves”), and by analyzing them, scientists can figure out which groups of neurons are firing together. And then by looking at which groups of neurons fire when the babies look at specific stimuli scientists can figure out which parts of the brain are in charge of which tasks.
For example, Dr. Nelson and his colleagues showed the babies pictures of both family members’ and strangers’ faces so that they could see whether the neurons responsible for facial recognition are in the same place in autistic and non-autistic brains.
The clusters of neurons that recognize faces are distributed across several different regions the brain, but in most people, the face-recognizing neurons tend to live on the right side of the brain. (To be clear: everyone has some face-recognizing neurons that live in the left hemisphere, but if you’re not autistic, it’s pretty safe bet that you have more face-recognizing neurons in the right half of your brain than the left.) Dr. Nelson found that autistic brains skewed the other way; they had more face-recognizing neurons in the left half of their brains.
But here’s the part that really blew my mind: The kids with autistic siblings also tended to keep their face-recognizing neurons on the left side of their brain, even if they didn’t develop autism. In the “high-risk” kids who didn’t develop autism, the leftward skew was milder (more like a 50-50 distribution), but it was still there. And you could very easily see that there were three different wiring patterns in play in the graphs Dr. Nelson showed.
This pattern held true for Dr. Nelson’s analyses of brain waves as well.
Broadly speaking, brain waves are divided into different types based on how often they repeat (measured by the EEG in hertz), and they tend to regulate different aspects of neural function. In this presentation, Dr. Nelson focused on gamma waves, which are known to be down-regulated in autistic brains and mu waves, which are suppressed in mirror neurons (the neurons that allow you to observe another person’s body language; many people believe that autistic brains have some kind of problem in the mirror neurons).
All of the “high risk” kids (both the ones that went on to develop ASD and the ones that didn’t) showed low levels of gamma in their brain scans at 3 and 6 months old. The levels of “gamma” for the “high risk” kids rose steadily until 9 months when they caught up with the “low risk” kids (who didn’t have autistic siblings). In contrast, the “low risk” kids’ gamma levels started high and slowly fell until the 9-month-old mark.
Dr. Nelson didn’t separate the autistic and the allistic members of the “high risk” group in this slide (even though he did mention that there was a slight difference between the two outcome groups). I’m guessing that either the difference in gamma between the autistic “HR” kids and the allistic “HR” kids wasn’t significant or that Dr. Nelson was trying to emphasize the point that even the allistic “HR” kids weren’t exhibiting completely normal brain patterns.
In his graph for the mu waves, Dr. Nelson separated the kids into their three groups. Both the “HR-ASD” kids and the allistic “HR” kids showed higher levels of mu than the “LR” kids, which may suggest that they have trouble suppressing mu in their mirror neurons (leading to dyfunctional mirroring). The “HR-ASD” kids also exhibited a very distinctive spike around 7-12 hz. No one really understands what this blip means in terms of practical brain function yet, but several studies have seen the same blip, so it may be one of the first reliable physical markers of autism spectrum disorders in the brain.
But once again, the allistic “HR” kids exhibited a milder version of the same atypical brain wave pattern. It was less extreme than the atypicalities observed in the “HR-ASD” group, but it was still clearly not “normal”.
And yet, when it came time to do the official behavioral diagnosis, these atypical “high-risk” kids were not considered autism spectrum at all. Dr. Nelson was careful to point out that some of these kids showed other neuroatypicalities (He didn’t specify, but it’s probably safe to assume that ADHD and dyslexia were among “the other things that they get”). But they didn’t have Asperger’s or any other form of “mild autism”.
These results suggest that even though these children have a very similar phenotype to their autistic siblings, they exhibit a different behavioral pattern or endophenotype. Which kind of throws a wrench in the arguments of people who talk about autism like it’s mainly a genetic disorder.
My Personal Take:
This was a tough one for me. Partly because my academic background is in general biology and I’m not used to have to keep track of electrical AND chemical synapses as they move through three dimensional spaces (A lot of times, when you’re looking at cells in a petri dish, they’re more or less in a two-dimensional plane. Human brains are not like that).
And partly because talking about autism hits me literally hits me where I live.
I have a neurodevelopmental disorder (ADHD), and I have an autistic brother. We both have had our fair share of problems as a result of our disorders, but I can’t imagine who either of us would be if our brains were wired “normally”.
I never really know what’s going in my brother’s head (and he has a really hard time telling us what’s going on), but a lot of times I feel like my weird brain “gets” him better than most other people do.
A lot of times I’m the only person in the room who understands that my brother isn’t just being “grouchy” or “antisocial”; he actually has a hard time understanding why people are asking him certain questions and expecting certain reaction. So I end up acting like an interpreter, telling him things like, “Well, when someone asks you who your favorite band is, it’s not because they’re trying to force you to pick THE ONE favorite. It’s usually more because they don’t know you, they don’t know what you like to talk about, and a lot of people like to talk about bands. And they’re hoping that you’ll like a band they also like so they can talk to you about it. You can tell them that you don’t like to pick favorites or mention more than one band. But you should probably say SOMETHING other than ‘I don’t know’.”
My mild social-development delay sucked when I was a kid/teenager, but it made me a lot more empathetic to my brother who has more severe trouble with social stuff.
When Dr. Nelson started talking about how the allistic siblings of autistic children also exhibited a milder version of the same leftward skew in their facial recognition neurons, my heart stopped. All I could think was, I could very easily be one of those kids.
I know I shouldn’t jump to conclusions or try to diagnose myself, but it’s hard to shake a thought like that.
My brother and I have distinctly different behavior patterns, but there are some similarities. We both spend a lot of time learning about topics that other people think are “weird”, and neither of us makes eye contact “correctly”. (Honestly, I think the neurotypical obsession with socially appropriate eye contact is actually kind of ridiculous. I have to bite my tongue to avoid saying things like: If I can think more clearly about the idea I want to tell you while I’m not looking at your face, why would you force me to look at your face while I’m trying to talk to you? Is your ego really so fragile that you need me to constantly reassure you with my face? And if I’m the one who’s easily distracted, why are you the one complaining that my hand-flapping and pacing are “disruptive” and “weird”?)
When he was talking about interventions, Dr. Nelson said that most of the interventions that will help autistic people probably won’t hurt allistic people. I can buy that logic when we’re talking about behavioral strategies like getting the infants to spend more time around people and more time looking at faces. But what about further down the road when we’re looking at drugs that will counteract the “connectopathy” and get the autistic brain to wire itself more normally?
If the kids who think and act like me aren’t distinguishable from our autistic siblings at the genetic level, a lot of those early interventions for autism aren’t just going to rewire autistic brains; they’re going to rewire a lot of other peoples’ brains, too.
What if forcing the brain to wire itself for social interaction damages other processing areas? What if some of these drugs “over-correct” the autistic “miswiring” and we end up with people whose brains are extremely well-suited to social interaction but are impaired when it comes to noticing non-human things in their environment? How can we be comfortable with allowing parents to decide what sorts of things their children’s brains should be wired to pay attention to?
It’s easy to say that non-verbal autism is a detriment, but once you start to talk about correcting neural wiring, it’s hard to know where lawyers and parents would draw the line between pathology and personality. What if we apply the anti-autism drug to a kid who would have turned out to be hyperlexic and in doing so, rob them of a talent for writing in iambic pentameter? What if an extroverted parent wanted to rewire their children’s brains so that their children would also be extroverted and therefore more compatible with the existing family structure?
I want neuroscientists to find out more about neurological wiring, but I also think that as a society, we need to have a serious, all-inclusive conversation about the way we treat people with neurological disorders and what types of interventions we’re willing to legalize.
It’s going to be a hard conversation. It’s going to be stressful and heart-wrenching, and we’re probably not going to find a policy that will satisfy everybody. But we still need to have that conversation, preferably sooner rather than later.
I don’t know whether my facial recognition neurons are sitting in the left half or the right half of my brain, and I don’t know whether I share the same gene variants that predisposed my brother to autism. But if there’s a subset of the non-autistic population that is genetically indistinguishable from their autistic siblings, that fact needs to be common knowledge well before anyone tries to push a prenatal autism screening test onto the market.
As a member of the scientific community, I think that the work Dr. Nelson presented was rigorous, logical, and excellent in terms of adherence to scientific method. It could also lead to the identification of biologically-grounded markers for early stage autistic development.
But I visibly cringed when he said that his work was funded by Autism Speaks, as if that association were a source of pride. (Many people in the autistic community protest that autistic people can speak for themselves and that as an organization, Autism Speaks is more concerned with making life easier for parents than actually helping its autistic constituents. Virtually everyone agrees that it is a politically biased group.) No scientist who truly hopes to make the lives of autistic people better should be proud to be associated with an organization that marginalizes the voices of people with autism.
I would like to see researchers in Dr. Nelson’s discipline continue investigating the way our brains are put together, but I would also like to see them back off a little bit when it comes to using terms that equate autism with being a disease or a danger that we should fear.
I’m scared of what might happen if people who don’t see the potential of the autistic brain take control of the research and try to “cure” or “prevent” autism rather than learning how to help autistic people learn how to communicate and how to make the most out of their off-kilter brains without succumbing to the stress of being “too different”. That fear is always with me. I feel it every single time I hear the word “autism”.
Dr. Nelson seems very sincere, and his particular genre of neuroscience research is going to be very important in shaping mental health policy in the next couple of decades. I learned a lot from his talk, and I want to learn more.
I hope that other people will try to learn more about autism too before demanding that neuroscientists come up with “cures” or asking for a prenatal screening.
Biggest Misconception to Avoid:
When psychiatrists and neuroscientists talk about autism, they are not talking about a physical problem. They are talking about a set of behaviors and impairments that often occur in the same individuals. When doctors diagnose autism, they are looking at how the child acts and not what’s actually happening in the child’s brain. It’s entirely possible that there are several different neural “mis-wirings” that could result in the same behavior. So autism could, in fact, have many different physical causes. (The same goes for other neurodevelopmental disorders like ADHD and bipolar disorder.)
The goal of Dr. Nelson’s research is to figure out what is physically going on in autistic brains.
Without that knowledge, psychiatrists are kind of stuck with throwing random drugs at the symptoms, and parents & educators are stuck with trying to eliminate disruptive behavior. Understanding the differences in neurological wiring that underly autistic brains will help us understand why autistic people percieve/react to the world the way they do.
Neuroscientists like Dr. Nelson tend to focus on infants, because when you’re trying to understand a long term developmental problem, it makes sense to start at the beginning. That way they can see when and where the autistic wiring starts to differ from neurotypical wiring. Investigators who are looking at older autistic children and adults have a much more difficult job, because they’re basically starting their research stories in media res with no data from the past to explain how these ASD brains are set up the way they are.
So if a neuroscientist is looking at the early development of autistic brains and talking about “early interventions” and “protective factors”, it doesn’t necessarily mean they agree with the curists. It means they’re trying to find concrete things that explain why some people develop autism (e.g. exhibiting autistic behavior) and others don’t. Finding these concrete things will probably help people learn to communicate with ASD children more effectively and may lead to the development of medication that relieves some of the stress autistic people experience as a result of sensory overload (among other things that the autistic community would find helpful.)
But there’s no way to generate that kind of specific knowledge about neural wiring without finding possible “cure” targets as a by-product of this type of research.
Dr. Nelson (on getting funding and carrying out research): “Getting people to play in the same sandbox for the greater good isn’t always easy.”
Best Audience Question:
(I’m paraphasing because it was a long question with quite bit of technical jargon, and the answer was even longer, but I did my best to get it all down in my notes.)
Audience member: Thank you for this talk. I think your work is absolutely brilliant. I’ve been doing research on 5-8-year-olds on the autism spectrum, and we’re seeing the same blip in Mu at 7-12 hz. Does that blip/spike suggest that something more complex is going on than a simple failure to suppress Mu?
Dr. Nelson: Yes. One possibility is that some of these atypical brainwave patterns are compensatory changes (e.g. If one part of its neural circuitry isn’t working, an autistic brain may start using Mu circuitry more. But this is just a conjecture at this point; we still don’t know very much about how synaptic pathways assemble themselves.)
You see lots of anecdotal stories where a researcher or a parent says, “I’m really worried about this kid. It looks like they’re really disengaged.” But then some of those kids improve and never develop autism.
To really know what’s going on there, you would have to look at individual development trajectories rather than statistical averages for a sample group (which is what we did in our study), and there isn’t a good way to do that type of analysis yet. It’s also possible that since all of the kids in our study have older siblings (and in the high-risk group, all of them have an older sibling with autism), the parents are interacting with the second child differently. So that could have an effect.
It’s also important to remember that autistic kids do recognize faces as being family or being strangers. We have to be careful about looking at the EEG and saying “They don’t know who their own mothers are.” They recognize their mother; they’re just using a different set of structures to do that recognition.
Just because they’re processing that information in the other hemisphere doesn’t mean that it’s an impairment. There’s a subset of left-handed people who are left-dominant for language processing but they don’t have a language problem. And there are people who are left-dominant for facial recognition who don’t exhibit the impairments associated with autism.
Question Most Likely to Yield a Post-doc Thesis:
Audience member: Is the atypical lateralization the result of hyperactivation in the non-dominant hemisphere or of hypoactivation in the dominant hemisphere?
(Translation: Do autistic people’s facial recognition neurons skew left because the left side of their brain is over-active or is it because the right hemisphere isn’t doing its job, leaving the left hemisphere to pick up the slack?)
Dr. Nelson: …(decides to tell a joke) See, this is why I don’t like coming to MIT; you always get these really hard questions. (But seriously-) I don’t know. That’s a really good question.
- Phenotype = The expression of an organism’s genes as influenced by the environment. Two organisms with identical genes can exhibit different phenotypes if they’re exposed to different environmental factors.
- Allistic = Not autistic. Includes both neurotypicals and people with other neurodevelopmental disorders.
- Endophenotype = A term used by epidemiologists who work on genetic disorders. They use it to differentiate between stable phenotypes, where the connection between genes & gene expression is well-understood and situations where the relationship appears to be more complicated. Dr. Nelson seemed to think we need to think about endophenotypes with regard to autism more.
- Neural oscillations (or “brain waves”) = Electrical impulses that occur between neurons and recur at predictable intervals.
- Neuronal lateralization = brains tend to delegate tasks unequally between brain hemispheres. Some tasks tend to be carried out by neurons in the right half; and some tend to be carried out by neurons in the left half. Dr. Nelson’s research suggests that autistic brains delegate tasks differently than non-autistic brains along the left-right axis.
tl;dr: A really smart neuroscientist compared babies with autistic older siblings to babies whose older siblings weren’t autistic. Some of the kids with older autistic siblings developed ASD; some didn’t. But the non-autistic siblings of autistic kids had different brain wiring patterns than the kids without any history of autism in the family. This suggests that you can express the genes that predispose your siblings to autism without necessarily being autistic yourself…which is kind of crazy to think about.