Autism – Closing In On the Causes

Autism is a poorly-understood condition characterized by problems with social interaction and communication. It is clearly a complex neurological issue, and its symptoms range from quite mild to very severe. As a result, neurologists tend to use the term “autism spectrum disorders” (ASDs), as they suspect autism is made up of a group of disorders with similar features.

I have a good friend with Asperger Syndrome, which is an autism spectrum disorder. While he seems mostly like any other person, he has some obsessive, repetitive rituals, and he sometimes experiences great difficulty in communicating with people, especially those who are unfamiliar with his personality. On the other side of the spectrum, a couple I know fairly well has a son with severe autism. It is difficult for them to communicate with him. It is as if he lives in his own little world. Additionally, he often experiences “meltdowns” in which he slams himself against the ground or the wall and screams at the top of his lungs. His behavior is not the result of “bad parenting.” It is the result of a serious neurological disorder.

What is frustrating for both health-care providers and parents is that so far, medical science has little to offer in terms of explaining what causes autism. In addition, while there are behavioral therapies that have helped many people with ASDs, it is difficult to prescribe a specific therapy for a specific individual. This, of course, leaves doctors and parents rather frustrated.

While there is a lot we don’t know about ASDs, there are things we do know. We know that they are on the rise. Even though there are many different ways to define ASDs, which leads to many different specific numbers, a good overview can be found here. Based on their numbers for the U.S. and outlying areas, for example, ASDs among people age 6-22 have increased 18-fold since 1992!

What are the causes of ASDs? The answer is that we don’t know. However, medical scientists are at least closing in on them.

For a while, it was popular to blame the rise of ASDs on vaccinations. After all, the number of vaccines given to children is rising, and the rate of ASDs is rising. Also, ASDs are generally diagnosed when a child is very young, which is also when a child gets a lot of vaccinations. Over the years, however, lots of studies have been done, and those studies clearly show that vaccination is not related to autism. Indeed, some of the authors on these studies are the same ones who demonstrated a serious side effect of the old rotavirus vaccine. That side effect was bad enough that the vaccine was taken off the market. If scientists who have a proven track record of finding problems with a vaccine couldn’t find a link between vaccination and autism, then it simply doesn’t exist.

More importantly, some very good studies have shown the ASDs are mostly (but not exclusively) inherited. For example, there was an excellent study published in 1995 that analyzed ASDs among British twins.1 It found that among identical twin sets in which one child had autism, 60% of the time the other twin had the same basic kind of autism. However, among fraternal twin sets in which one had autism, there was not a single example of the other fraternal twin having autism. If they expanded the study to include all ASDs, 92% of identical twins shared ASDs, and only 10% of fraternal twins did. Since identical twins have the same DNA and fraternal twins have significantly different DNA, this study shows that genetics is the underlying cause of most ASDs.

But wait a minute. If the underlying cause of most ASDs is genetic, why are ASDs rising so quickly in the United States? Is is really possible that human genetics is changing so rapidly that a genetic disease has increased 18-fold in less than 20 years? Surprisingly, the answer to that question is “Yes.”

People tend to view genetics as relatively stable, but that is just not right. Instead, the human genome changes rather rapidly, in many different ways. For one thing, genes tend to get copied in the human genome, and the number of those copies can be dramatically different from person to person. For example, people in the U.S. have, on average, seven copies of the gene for amylase, a salivary enzyme that helps to break down starch. However, a given individual in the U.S. can have as few as two copies of the gene or as many as fifteen! It turns out that this actually tracks with diet – the average number of amylase copies in a population rises with the amount of starch in the population’s diet.2

The human genome also changes as the average age of a population’s parents changes. Genetic mutations in the germ line (the reproductive cells) accumulate as we age, especially among men.3 As the average age of parents increase, then, they produce children with more mutations.

It turns out that both of these effects play at least some role in autism. For example, several studies have shown that older parents are more likely to have autistic children than are younger parents. One study, for example, analyzed more than 130,000 Jewish people. They found that fathers who were over 40 were almost 6 times more likely to have children with ASDs than were fathers under 30.4 Other studies show that the mother’s age also plays a role in the likelihood of a child having ASD.5 So at least one of the reasons ASDs are increasing in the United States is that the average age of parents is increasing.

More recently, a major genetic study on ASDs was published, and the results indicate that the number of copies of certain genes in a person’s genome dramatically influences the likelihood of a child having ASD.6 The study looked at the genomes of more than 1,000 children with ASDs and compared them to the genomes of almost 2,000 children who did not have ASD but were otherwise similar to the children with ASDs. They found that compared to the non-ASD children, these children had significant variation in the number of copies of certain neurological genes. Interestingly enough, sometimes the number of copies were inherited from a parent. However, some of the variations in number of copies (about 6% of the total) were not inherited from either parent – they were brand new, appearing only in the child.

So there is a lot of evidence that the causes of ASDs are genetic. While it is good to know that scientists are closing in on the causes, it is also a bit disheartening, because a lot of genes were implicated in this study. As a result, there is probably no easy answer in terms of treatment or prevention. Certainly the medical community now knows more than it did before, but there is clearly a lot left to learn.


1. Bailey A, et al., “Autism as a strongly genetic disorder: evidence from a British twin study,” Psychol Med., 25:63-77, 1995.
Return to Text

2. George H. Perry, et al., “Diet and the evolution of human amylase gene copy number variation,” Nature Genetics, 39:1256-1260, 2007.
Return to Text

3. A. J. Wyrobek, et al., “Advancing age has differential effects on DNA damage, chromatin integrity, gene mutations, and aneuploidies in sperm,” Proc Natl Acad Sci U S A., 103:9601–9606, 2006.
Return to Text

4. Abraham Reichenberg, et al., “Advancing Paternal Age and Autism,” Arch Gen Psychiatry, 63:1026–1032, 2006. (available online)
Return to Text

5. Lisa A. Croen, et al., “Maternal and Paternal Age and Risk of Autism Spectrum Disorders,” Arch Pediatr Adolesc Med, 161:334-340, 2007. (available online)
Return to Text

6. Dalila Pinto, et al., “Functional impact of global rare copy number variation in autism spectrum disorders,” Nature, doi:10.1038/nature09146, 2010. (available online)
Return to Text


  1. Ben Michael Fournier July 15, 2010 12:09 am

    I’ve also been diagnosed with Asperger’s syndrome. For me, the traits that qualified me were my tendency to fixate on topics, poor executive skills, and introversion.

    • jlwile July 15, 2010 3:26 pm

      I had no idea, Ben. Do you think that the condition limits you in any way, or is it more just a part of your personality? I ask because my other friend with Asperger’s syndrome recognizes that he is a bit different from the average person, but he considers it a plus.

  2. Ben Michael Fournier July 16, 2010 2:39 am

    I tend to have difficulty at job interviews. I tend to talk too much and not smile enough, which are the opposite of what they want to hire. That tends to stink, but otherwise I don’t have a problem with having Asperger’s. On a forum called Wrong Planet it gives me a commonality. Many of the people who have been newly diagnosed are often of the mood of hating “NTs” or neurotypicals, but most people who have been diagnosed for a while tend to recognize that people with AS are pretty much the same as NTs with the exception of a communication difficulty. After a while they realize that there are just as many bullies with Asperger’s as there are who are neurotypicals, especially if they attempt debate with another Aspie. Debating with other Aspies, especially if it is against their current interest, usually evokes resentment at the least and a lot of name calling and insinuation. At least that has been my experience with other Aspies on WP in the PPR subforum. There are others who don’t really like all the drama and just want to avoid all of that. But for me, I tend to think I’ve only been limited in being able to be hired at places, but otherwise I think its a benefit to be able to learn subjects rapidly, although depending upon interest in said subjects.

  3. Eric July 17, 2010 7:14 pm

    Dr. Wile, been reading for a while, but this item prompted me to respond. My 6-yr old son is autistic. Thank you for stating vaccines don’t cause autism. This myth is somewhat accepted by some of the homeschoolers I’ve talked to. And most of your other points are good, too. But I’d suggest that autism has not been rapidly increasing as you suggest. The link you’ve provided is to numbers produced by Thoughtful House, part of the bio-med treatment movement, that was founded and headed until recently by Andrew Wakefield, of MMR causes autism infamy. A more accurate take is here:

    I don’t understand how you think the genetic code changes so fast (the last 18 years). Finally, the Pinto study is also discussed in depth here:

    thanks for writing this blog.

    • jlwile July 18, 2010 12:44 pm

      Hi Eric,

      Thanks for the comment! While I agree that Thoughtful House has some less-than-reputable characters involved in it, their sources for the data are quite reliable. Indeed, the graph presented at your source is in fairly good agreement with the ones to which I linked. While the data are roughly the same, the conclusion of your source is (in my opinion) not at all correct. What the author seems to be missing is that in 1997, a new category was added to the classification system used by IDEA. It is called “developmental delay.” All those cases used to be lumped into “mental retardation.” Thus, cases were not transferred from mental retardation to autism. Indeed, the diagnosis of mental retardation has nearly no overlap with the diagnosis of autism, so it is not even reasonable to assume such a transfer would be made. The transfer was from mental retardation to developmental delay.

      In answer to your question, the science of genetics shows that the human genome does change very rapidly, especially on the margins. Remember, autism affects a small number of children. Even modest genetic changes can produce a shift that will increase incidence, and if incidence is small to begin with, modest changes show up dramatically. Think about it. The study I mentioned in the article indicates that once the father is over 40, the rate of autism rises by a factor of 6. So if even a small shift in the average parental age occurs, suddenly a lot more children develop autism, and since the number was initially quite small, those new cases cause the rate to rise rapidly.

      Your second link is mostly a good one. It is more detailed than I think most people want to read, but I love how he takes the quacks to task. Of course, he also makes the incorrect statement that autism only appears to be rising due to diagnostic substitution. That is just not correct. As Psychology today says:

      Can the rise in the prevalence of autism from 1 in 10,000 infants to 1 in 150 in a few decades really be due simply to our greater awareness of and readiness to diagnose the disease? Boy, that’s as hard to swallow as the thimerosal linkage. Maybe greater sensitivity to – and a broadening of the definition and criteria for – autism account for some of the increase since 1992, but all of the 1500% increase? Autism often causes a major – sometimes near-total – failure at functioning (as it did with Noah Greenfeld). Did we really miss all of these until recently? Not likely.

  4. Eric July 19, 2010 6:36 pm

    You are welcome. Here is one more autism prevalence link:

    What is “on the margins” of genetics? Also, age of father isn’t a genetic effect, right?

    • jlwile July 20, 2010 3:41 pm

      Thanks for the additional link. I think it is better than the previous one. It recognizes that increased surveillance alone can’t be responsible for the increase, and unlike the previous link, it recognizes that diagnosis issues cannot be the sole cause of the increase. It then raises the possibility that both of these effects together could account for the increase. I supposed that’s possible, but until someone can actually present some evidence that it is the case, I think the most reasonable alternative is to go with the obvious conclusion – that autism rates really are rising.

      In this context, the phrase “on the margins” refers to genetics that are rare in a population. When a genetic trait is rare in the population, it can be selected out of the population entirely, or it can rise dramatically on a per capita basis, depending on all sorts of reproductive and environmental issues. Once a genetic trait is more common within a population, it is difficult for it to be selected out of the population completely, and while it may not stop rising, the rate of per capita rise typically falls off, unless selection pressures are very large.

      The age of the father is definitely a genetic effect, as is the age of the mother. In the case of the father, sperm are being produced throughout the man’s lifetime. As a result, mutations can occur in the process of sperm production. In addition, and probably more importantly in the case of autism, genes can be copied during the process of sperm production, which gives rise to copy number variations in the child. The older the man, the more likely this is to happen, as the genetic “proofreading” process deteriorates with age. Thus, the genetics of the sperm in a young man are different from the genetics in the sperm of that same man when he is older.

      With the mother, the effect exists, but it is less pronounced, because the eggs are all made during fetal development and then “frozen” in an early stage of a process called “meiosis I,” which is part of what takes an immature egg and turns it into a mature egg. This means that the most error-prone activity related to genetics and reproduction (the copying of DNA) is accomplished before the mother is even born. Nevertheless, in order for the egg to be ready to be fertilized, it must complete “meiosis I” as well as another process called “meiosis II.” While this does not require any additional DNA replication, it does require a lot of chemistry, which can in turn cause changes in the DNA.

      Thus, a changing age for parents means a changing genome for children. The older the parents, the more likely the children will have mutations and copy number variations.

  5. Lydia T July 20, 2010 8:44 am

    Hey Dr. Wile,

    This was really interesting to read about.. I have a brother (adopted) who has severe autism like the couple’s son that you know. I never could understand why, but using all your information in this post up against my brother’s birth family and stuff, it makes a lot of sense.. especially the part about parents having more copies of certain genes.

    • jlwile July 20, 2010 3:44 pm

      Hi Lydia,

      Thanks for the comment! I am glad that the article was helpful.

  6. Eric July 22, 2010 10:51 am

    “When a genetic trait is rare in the population, it can be selected out of the population entirely, or it can rise dramatically on a per capita basis, depending on all sorts of reproductive and environmental issues.”

    But isn’t this true only for multiple generations? Natural selection doesn’t work within one generation, does it?

    • jlwile July 22, 2010 3:24 pm

      Eric, you are thinking a bit too strictly for a population of people. For convenience, we split populations into generations, but in reality, changes occur continuously, as do births and deaths. Thus, the genetic traits of a population change continuously. A genetic trait that is on the margins, therefore, can disappear or grow quite quickly. In the context of our discussion, natural selection hardly applies, since developed countries insulate their individuals from natural selection.

      So imagine the scenario in which the average age of the parents is the most important genetic factor, and imagine further that there is a significant increase in copy number variation or mutation after the male reaches 40. If the average age of the parents increases by only a month each year, then in 12 years, the average age of the parents will have increased by a full year. If the average age of the parents was 39 before and turns to 40 over the course of 12 years, suddenly, the AVERAGE person is significantly at risk for the copy number variation or the mutation.

      Now that’s not what’s happening, of course. Average parental age is well under 30 (I expect), and I don’t know how quickly it is increasing. However, the illustration shows that in less than one generation, a dramatic change in the genetic traits of the newly born can happen, if those trait start out as rare in the population.

      I agree that changes in the AVERAGE genetics of a population take several generations to realize. However, significant changes of genetic traits that are on the margin can occur very rapidly, as long as there is a continuous trend that increases the likelihood of that trait appearing.

  7. The Black Sheep July 22, 2010 2:39 pm

    I like it when you write about genetics. Way more interesting than that age of the Earth stuff!

    It seems both genetic and psychological problems of children always get blamed on the mother. Even as a female with no kids (and no intentions of having them… biologically, at least) it’s nice to hear for once that the “effect … is less pronounced” in the mother. It’s not the girl’s fault for once! WOOHOO!!!!!

    • jlwile July 22, 2010 3:30 pm

      Black Sheep, that’s why I wish you would comment more frequently! Indeed, at one time, doctors used to look at the mother’s age as significantly more important. Given the fact that the mother’s body nurtures and houses the baby for its first nine months of life, they assumed that the older the mother got, the worse it would be for the child. In turn, since the man “just” donates sperm, and since the sperm is constantly “refreshed,” it was assumed that the age of the father had almost nothing to do with the health of the child. What we have found over time is that the opposite seems more true. The constant “refreshing” of sperm makes them more susceptible to genetic issues the older the man gets, and the fact that the eggs are “frozen” in an early stage of meiosis I makes the woman’s age much less important when it comes to the genetic health of the child.

  8. The Black Sheep July 22, 2010 7:16 pm

    So wait, you wish I would comment more so that you can tell me I’m wrong?

    • jlwile July 22, 2010 9:30 pm

      Black Sheep, I wish you would comment more because your comments are good.

  9. Eric July 24, 2010 12:22 am

    Dr. Wile, I’m sorry, but I’m terribly confused. First, I said natural selection because of the quote from you I posted had “selected” so I thought it must be natural, are there other kinds? (besides sexual I guess)

    “the genetic traits of a population change continuously” this is confusing. lets just stick with individuals. I suppose the genetic traits in a man’s sperm can change continuously (but all that much or fast? this was my original question), but these traits can only be passed down through children. So only changes before conception can be passed down. Can you explain this more?

    “since developed countries insulate their individuals from natural selection” This seems wrong. Are you saying autistic individuals are less likely to reproduce but developed countries somehow counter that less likelihood?

    “However, significant changes of genetic traits that are on the margin can occur very rapidly, as long as there is a continuous trend that increases the likelihood of that trait appearing.” What is the trend that increases the likelihood of autism appearing? From the studies, I got that most, if not all, of the increase in autism prevalence can be explained by the factors we know about – increased awareness, diagnostic substitution, wider definition, etc. But what are the genetic or environmental factors that are increasing autism? I’m afraid I can’t believe the age of the father is that important.

    • jlwile July 24, 2010 11:40 am

      Eric, sorry to confuse you. In the human population, artificial selection is the dominant mechanism, not natural selection. So, for example, if people start decided to have children later in their life, that is a form of artificial selection, as they are selecting (without knowing it) children with more mutations and more copy number variations.

      What I mean by “the genetic traits of a population change continuously” is that we don’t all have our kids at the same time and then wait for them to all have kids at the same time. People are having children all of the time. Thus, EVERY YEAR is a new generation for some segment of the population. The parents who had kids 18 years ago, for example, just had grandchildren this year. The parents who had kids 17 years ago will have grandchildren next year. Thus, for the former, this year represents a new generation. For the latter, next year represents a new generation. So to say that the genetics of a population can change over many generations really just means the genetics of a population can change over just a few years.

      Yes, the genetics in a man’s sperm can change very quickly, especially as the man gets older. The older he gets, the more susceptible his sperm are to mutations and copy number variations, so the older he gets, the more his sperm changes.

      When I said, “developed countries insulate their individuals from natural selection,” I mean that we do not allow the weak to die out. This is not specific to autism – it is common to all genetic “abnormalities.” In a less-developed country, a child who is less fit to survive will be more likely to die. In a developed country, a child with the same genetic condition will be more likely to survive to reproduce, given that there is better medical care, better treatment options, etc.

      The continuous trend that we know affects autism is the average age of the parents. We know that the older the parents are, the more likely the child is to have autism. Whether or not you believe the age of the father is important doesn’t matter. The studies clearly show that it is. When a father is over 40, his children are SIX TIMES more likely to be autistic. You can choose to deny this if you want, but you are doing so against the data. Since we also know that the average age of parents is increasing in the U.S., we now know there is a continuous change that is producing more autism.

      As I have said before, you can assert that the rise in autism is due to factors like increased awareness, diagnostic substitution, wider definition, etc. However, there hasn’t been any evidence to show that’s the case. In addition, as the quote I shared earlier indicates, it strains the limits of credulity to think that these factors are all that’s behind the meteoric rise in autism cases. Thus, the most reasonable conclusion is that autism rates are, indeed, rising. I choose to believe the most reasonable conclusion.