Autism: Reading the Brain in the Face

The human embryo develops three basic germ layers, which each give rise to various and occasionally overlapping structures. This is known as the Germ Layer Theory, originally proposed by Karl von Baer.

Until more recently, it was believed that the mesodermal layer (see image below) was the only layer from which skeletal and connective tissues arose. It wasn’t until the latter part of the 20th century that it was realized that most of the bones of the skull, including the face, were actually derived from the neural crest, a portion of tissue located in the ectoderm that also gives rise to the nervous system, melanocytes (pigment cells), and some skin cells.


We have known for some time that tight links exist between the development of the brain and the formation of the face and skull. Not only are these two systems derived from similar groups of neural crest progenitor cells, but they also appear to be in constant crosstalk with one another during early development, likely sharing and responding to many of the same gene products due to their shared parentage. Therefore, it’s interesting though unsurprising to find many individuals with various neurodevelopmental conditions who share craniofacial dysmorphism (malformation of the skull/face) as a related feature.

Autism is no exception. Of course, there are many genetic syndromes that have autism as a primary or secondary feature that exhibit facial malformations. Individuals Fragile X, for instance, often have characteristic facial features such as macrocephaly, a large forehead, long face, prominent jaw, and large ears.


A boy with Fragile X Syndrome.

Individuals with Smith-Magenis Syndrome, another condition that often has autism as a secondary feature, likewise have particular facial characteristics, including a broad face, underdevelopment of the midface giving the cheeks and eyes a sunken flat appearance, a broad nasal bridge, and a downturned mouth.


A young man with Smith-Magenis Syndrome.

As a final autism-related example, people with Pitt-Hopkins Syndrome tend to have multiple facial anomalies, including a square forehead, a protruding lower face, broad nasal bridge, flaring nostrils, and a wide open mouth.


Two individuals with Pitt-Hopkins Syndrome, from youth to adulthood.

Each of these are somewhat extreme examples of craniofacial dysmorphism, most likely due to the fact that they’re caused by rare mutations that have serious consequences for tissue development, including the brain, when they do occur. Most of these individuals tend to have varying degrees of intellectual disability alongside their other features.

But what about autism that isn’t due to highly detrimental rare mutations? Do they show facial dysmorphisms as well?

According to Aldridge et al. (2011), some subgroups in idiopathic autism (as it’s currently known) do actually show distinctive facial characteristics that vary from control groups, though the features may be comparatively more subtle requiring the assistance of technology. While a large portion of kids in the Alridge et al. study didn’t appear to differ from controls, two autism subgroups did emerge, which together comprised about 1/3rd of the larger group.

Screen Shot 2015-09-06 at 6.03.06 PM

The first autism subgroup (Subgroup 1) shown above, which made up about 19% of the group, exhibited increased distances (in black) between numerous landmarks between the eyes, nose, and mouth, with a decreased distance between landmarks along the mouth itself ranging down to the chin (white). These boys tended to be more severely affected, with fewer instances of Asperger’s Syndrome and macrocephaly in the group.

Meanwhile, boys in Subgroup 2 (~9%) tended to show decreased inter-landmark distances surrounding the orbital and nasal regions and increases in the length of the upper lip. In contrast to the first subgroup, this second subgroup had lower severity scores, more instances of Asperger’s Syndrome, and more occurrences of macrocephaly.

This study suggests that, while not all individuals with autism exhibit measurable facial dysmorphism, significant subgroups do– subgroups which tend to occupy the two extremes of functioning. Since all the boys participating in this study had to be capable of a certain level of self-restraint in order to sit still long enough to have their pictures taken, one can assume that the lower functioning ranges of autism were not well-represented in this study.

However, the fact that two separate subgroups emerged, reflecting not only a divergence in facial features but also levels of functioning, it may suggest different but overlapping causes to their respective autisms, in terms of genetic and environmental factors. These two subgroups may in fact be two different types of autism, etiologically.

20 responses to “Autism: Reading the Brain in the Face

  1. Facial dysmorphism is seen in genetic syndromes and also in environmental based syndromes such as congenital rubella syndrome, valproate acid syndrome and fetal alcohol syndrome. In none of the genetic and environmentally induced syndromes is autism 100% penetrant. The majority do not meet diagnostic criteria for autism. There are other factors that discriminate the group with autism from those without autism. Gazuidan published a small study which showed that in Down syndrome and autism there was a family loading of broad autism phenotype traits that was not seen in Down syndrome and no autism. None of the first degree relatives of children and autism had either Down syndrome or autism. The structural anomolies in the brain may induce BAP traits at the extreme resulting in a trajectory that leads to an autism diagnosis in children positive for a family history of BAP traits in first degree relatives:

    • Hard to say, since many studies are prone to remove regressive autism from their patient pool, which is a real shame. Why not just study them as another subgroup, I say.

      From my own current work, regressive autism likely lies on a continuum with various forms of autism without regression, both of which may involve components of cellular stress as a part of their phenotype. Support for this comes from the fact that there are many instances of multiplex autism families in which one sibling had a regressive form of autism meanwhile the other had a more “congenital” form.

      In regression, there is typically a stimulus (onset of seizures, an immune challenge, etc.); however, it is likely the straw that broke the inevitable camel’s back. However, I’m currently starting to investigate the possibility of medicinal intervention to alleviate the severity of the regression and hopefully ultimately ameliorate the severity of the ensuing autism symptoms. At present I’m reviewing some potential medications that could fit this bill. The biggest challenge will be to identify the children experiencing regression soon enough in order to treat them and make a real difference, since it’s usually the loss of skills that alerts parents to the fact that something’s wrong in the first place. The logistics may actually be more of a challenge than the effectiveness of the treatment itself.

  2. My daughter has a genetic form of autism that causes regression, in many cases, after immunizations. Makes. my. head. spin. She did not regress, nor are autoimmune issues common in my family. I was not worried about immunizations despite others I know being careful. I watched carefully, she was fine. Others, who have a family history of immunization reaction, had a similar autism outcome but certain symptoms were worsened after seizures. Fun to see this here, Judy described their study and how Temple Grandin went up to the pic and said, “That one is me,” at the Thompson Center. Since this was done at the TC and they were vocal about regressions being an important part of autism, doubt they changed the pool. You’d have to look at the study or email one of them…

    • Hi, twinkle. That’s very interesting. May I ask, does your daughter have Dravet’s Syndrome? (That’s the primary one that comes to mind when talking about autism, seizures, and potential for regression.) Hopefully the fact that your daughter hasn’t regressed means she will have a lighter variant of the condition. I hope so. 🙂

      From my review of the Dravet literature, I remember there was a series of studies that investigated whether vaccination could cause Dravet’s Syndrome. The researchers determined that, for those individuals who had an immediate reaction following DPT vaccine, the individual outcomes were the same as those who had onset following illness or no apparent stimulus at all. However, the vaccination did cause a somewhat earlier onset of illness.

      It would be interesting to study individuals with a regressive genetic syndrome who do not regress. Perhaps there are genetic/environmental protective factors worth studying.

      Thanks for sharing!

      • Not Dravet, many similar aspects, but Dup15q (IDIC15q to be more specific, you can peek at my blog, but there isn’t much there). Much more uniform diagnosis as it is a specific range of duplicated genes, as opposed to a phenotypic diagnosis under which several genetic mutations have been aggregated– am I remembering Dravet right? Seems like there was a GABA gene that affected dravet … yup. The Dup15q 11.2-13.1 region has three overexpressed GABA genes, among others, and the SUDEP risk is similar to Dravet when Benzo’s are used. We use Keppra and Piracetam in my daughter to inhibit GABA overtransmission and prevent overload of chloride in the cell (my theory based on Yehezkiel Ben-Ari). IDIC– Has a seizure profile more like Tuberous Sclerosis, and a similar brain structure to TS if I remember right. Music/animal lovers, need high levels of proprioceptive input, high risk of coeliac disease, chronic thirst, underdeveloped palate/philtrum, etc. The few cases where I have specifically heard of regression were cases which resulted in mitochondrial dysfunction and where folate cycles were severely messed up in the family, especially one friend who had chronic miscarriage. The few cases which have not been vaccinated do seem to have fewer autism diagnoses. Many of the Dup15q kids have autistic features but are comfortable with human interaction. The one thing lacking seems to be eye contact, socially. Communication is almost always severely delayed. Kid who had regressions after vaccinations tend to start symptoms early, and in this disorder, that makes a huge difference for lifelong seizure risk and treatability of seizures. I have asked several times… but it would be hard to research as many publishers are against that type of work, and the political environment surrounding those hypotheses is.. um… strenuous at best? 😀 I do not have much mitochondrial symptomatology outside what I believe is a slightly higher requirement for Carnitine (To correct what I think is acidosis resulting from overuse of muscles by my dopa-responsive-dystonia body. Give me levodopa, carnitine, duloxetine, and b’s/folates, and I am my most focused. No blood tests yet, but it does prevent weakness.)

  3. Yeah, I think it’s really unfortunate how the sociopolitical environment surrounding vaccine research in relation to anything neurodevelopmental is exceedingly biased. I would really love the opportunity to study the background of the kids who have an identifiable reaction to vaccination and/or illness. I think we could learn a lot that is currently ignored, except within certain small circles, such as mitochondrial disorders. But a scientist, the few who may actually be interested in this topic, would hardly be able to get grant funding for something of that nature. It’s gotten too much bad press, and there’s incentive from within the medical community to keep people vaccinated. So no federal funding agency will pay to have the pot stirred further. Sad really. (It’s a shame to think that, in spite of what I would consider some reasonable concerns by parents for the safety of their kids, the ways in which that concern was communicated via spokepeople such as Jenny McCarthy has in fact led to the exact opposite of what they wanted to achieve: more safety studies, improved vaccine safety, etc. But now instead, no respectable scientist would touch the topic with a ten foot pole.)

    It seems you’ve done quite a bit of footwork on your child’s condition, as well as your own unique makeup. That’s excellent. PS: I also really love Ben-Ari’s work. Last year I attended a small congress he holds biannually at his home in La Ciotat, France. I didn’t get to speak with him too long, as he was busy coordinating. But he is certainly quintessentially French, hehe. In fact, he was one of the reviewers on my last manuscript, for which I was happy. But his work makes a lot of sense. I’ve even thought that in at-risk kids, a temporary application of an NKCC1 antagonist just prior to and during vaccination (very short-term use, mind you) could be helpful in preventing regression. I’m considering other medications as well though that aren’t GABA-related.

    I also agree with you that, even if a vaccination or illness causes an earlier onset of a condition, it is worthwhile to consider an alternative vaccination schedule or even forego vaccination altogether, depending on the condition. (For instance, my step-grandson has a Congenital Disorder of Deglycosylation and he therefore can’t be vaccinated. It triggers massive seizure activity.) Which is why I found it shocking that the researchers in the Dravet’s Syndrome work recommended no change in vaccine schedule. I found that pretty cold. I think they were just attempting to rock the boat as little as possible. I believe originally they were hoping to DISPELL concerns, whereas in fact they ended up reporting solid evidence to the contrary.

  4. In Dup15q, the earlier onset changes the type of seizure– kids who get seizures in infancy tend to end up with Lennox-Gestaut Syndrome (spelling…?) and those who get them after puberty get easier-to-control absences and tend to have conversations as opposed to minimal/no speech. That is a broad overgeneralization but gives you an idea of how the bell curves lie… ;-D Huuuuuuge range of outcomes. I prefer to be as conservative as possible.

    • I think there’s a similar bell curve regarding seizures in general, though I’m definitely no expert. But infants, toddlers, and children tend to have different seizure types and patterns. It’s fascinating, from the scientific perspective, in the sense that I suspect there’s something neurologically about the toddler’s brain that is potentially vulnerable to regression and perhaps even overall seizure propensity. For instance, febrile seizures are rare in infancy, comparatively common in the toddler years, and then decrease substantially in childhood.

  5. Not sure if you heard of the United States CDC researcher who said his team destroyed data to avoid rocking the vaccine boat? Reported it to Congress to have them investigate. Blew me away. I believe most of the controversy is emotional- very protective doctors who fear losing patients vs. parents with ASD kids. But also, there is a disconnect between the way the medical community communicates, and the way the families communicate the same information. Scientists harp on definition and completely lose the meaning of what parents say because they have lost the ability to read and comprehend idiomatic speech. That’s also among the things that stun me…

    • Yeah, pretty scary. Although I suspect most doctors and scientists truly believe in what they are saying. In most instances it’s a bias in hearing what you want to hear and then being content in looking no further. In this case, there’s also incentive to try and protect herd immunity and prevent the resurgence of some pretty nasty and deadly diseases. The epidemiological studies on vaccination and autism gave a resounding “no”– amplified by media coverage– and most people are happy with that. Until I read the Dravet literature, I myself was relatively happy with that. While obviously autism doesn’t equal Dravet’s syndrome (or Dup15q), it’s an example to the contrary and could be used to understand the mechanisms underlying an adverse neurological reaction following vaccination. For one, even if autism numbers are fairly stable, it hasn’t been addressed whether autistic kids have more severe reactions to inoculation– or whether an adverse reaction increases the severity of a condition that was already present, in which case the numbers themselves wouldn’t technically rise. There are many many questions to be answered, and few who are willing and know enough to ask the right ones.

      From what I’ve read in the literature and what I understand of the underpinnings of autism, I would literally be shocked if there were no connection in a subset of cases. That just wouldn’t make sense to me. You have a child with a propensity for cellular stress, be that a pure seizure disorder related to changes in sodium channels, glutamate/GABA receptors, etc., or a secondary feature due to metabolic stress on various organelles (ER, mitochondria, vesicular trafficking, etc.), and you cause an immunological reaction that subsequently triggers molecules, e.g., IL-1, that have known excitatory features, most likely through the neural activation of inflammatory pathways. These may subsequently trigger neuronal hyperexcitation, triggering a cell stress response and a whole cascade of neurodegenerative-like effects, ensuing regression, potential increase in autistic features, seizures, etc. While it’s a hypothesis, it’s one that is based strongly on evidence and gives reason for concern as well as areas of further study. I don’t know if these will be answered while we’re in this current social climate. In addition, any evidence supporting a claim against vaccination would need to be disseminated very carefully, with means already in place for prevention so that, what with herd mentality, people wouldn’t freak out more and even more people forego vaccination. Just a hundred years ago, you could expect about half of all children 8 years old and under to die, often of something communicable. Whereas today, we literally have little concept of those stats. We didn’t live during those times. But we do live in an age of autism, so in the human mind that makes it far scarier. So before such information should be spread to the public, we would need a plan in place to altering vaccination schedule based on patient vulnerability (the new era of personalized medicine), plus perhaps some sort of short-term medicinal intervention to protect the brain from regressing if the child does have an adverse reaction.

      • I have looked for 15 years for someone in science to look outside the dogma. There is a reason why parents hang on so tenaciously to the idea…it’s because so many of their children have had reactions. Dravet, now epilepsy…slowly the screw turns…Eventually science will have to admit that reactions may be indicators of innate neurological differences, and not “epidemiologically unrelated”.

      • I agree, though it may be a somewhat long and stubborn process now that public and professional opinions tend to be so biased against the idea. I think it was there all along, they just weren’t asking the right questions. And I suspect a number of scientists were guilty of seeking the answers they wanted to hear– not that they may have been consciously aware of that. But then I strongly believe Wakefield was guilty (and probably more consciously so) of the same. I wouldn’t trust that guy as far as one could throw him. Bad scientist.

  6. Bad scientist, potentially with good motives. They say his material was all fabricated, although for a while there were several on his team saying the allegations were false. I am not sure how much politics had a role in how that was all portrayed. I could put together his study for real, I only have to ask a few friends to pull in a few favors to give the same kinds of examples. That’s why I find it hard to believe the study was exaggerated…

    But the sad thing is that the political climate drives coverage of one scientist, Wakefield, instead of the CDC.

    Have enjoyed this little convo. First for a refreshing break from a community where hardly anyone has studies past high school. Secondly, a true scientist even getting that inflammation affects autism symptoms.

    I think the climate will change. Parents are getting better at understanding and defining things. We will look at immunity and inflammation differently in a few years- especially with U. V’s study this year of immune structures in the brain. Not having any part of the brain connected to the immune structures themselves was a hit against immune theories of autism- we know better now. This is the strongest evidence I have to be careful with my daughter’s environment.

  7. Pingback: Autism: Reading the Brain in the Face | No really, I WANTED three boys!·

  8. imo autism is linked to recently discovered circulation in connective tissues. Face is big part of intake/exhaust part of that system and it kind of work like VTEC engine where intake power is adjusted to how engine (brain) works. Imo in kids with autism it may work at high speed all the time.

Leave a Reply to Robert Jensen Cancel reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s