Sunday, June 26, 2016

Autism In Adults: A Survey - Part 2: Executive Dysfunction

This is a continuation of my series on a survey I did in 2012, and am only now analyzing the results. To get background information on the subjects and methods of this survey, look at part 1.

In this section, I will discuss questionnaire results regarding one area of common coexisting difficulties - executive dysfunction.

Executive dysfunction was assessed using the Frontal Systems Behavior Scale (FrSBe), a 46-item assessment commonly used for adults with neurodegenerative disorders and brain injuries. The questionnaire is normally informant report, but was adapted to self-report by changing the wording of items to first person.

Scores on the FrSBe ranged from 24-93, with a mean of 67.11+/-15.610 (higher scores mean more impairment). Although I couldn't find a cut-off score for this scale, this mean is slightly higher than both pre-Huntington's (mean 59.6) and normal control (mean 54.8) participants in this study, but much lower than participants with a variety of neurological conditions (means from 98.63-140.9) in this study. On balance, this comparison would seem to suggest that the typical participant in my study had mild impairment on this scale, although some were well in the normal range and some had scores typical of individuals with early-stage Alzheimer's disease.

The FrSBe items are divided into three subscales - apathy, disinhibition and executive dysfunction. Apathy items reflect a failure to do activities and a general lack of motivation; disinhibition reflects unusual, socially inappropriate and impulsive behavior; and executive dysfunction reflects disorganization, poor planning and poor self-monitoring.

On the 14-item apathy scale, participants' scores ranged from 11-33, with a mean of 22.66+/-5.439. Compared to the same studies as before, this score is substantially higher than pre-Huntington's (mean 13.7) and normal control (mean 11.7) participants, but lower than the mixed neurological conditions group (means 35.6-48.6). This suggests that many of my participants had mild struggles with apathy and lack of motivation.

On the 15-item disinhibition scale, participants' scores ranged from 6-36, with a mean of 17.85+/-7.313. This score is similar to the pre-Huntington's (mean 13.7) and normal control groups (mean 18.8), and far lower than the mixed neurological group (means 27.4-34.5). This suggests that most of my participants did not have significant difficulty with disinhibition, although the highest scorers had scores similar to neurologically impaired adults.

Lastly, on the 17-item executive dysfunction scale, participants' scores ranged from 6-40, with a mean of 25.8+/-7.115. This score is substantially higher than both pre-Huntington's (mean 19.3) and normal control (mean 24.3) participants, but lower than the mixed neurological conditions group (means 41.3-57.8). This suggests that many of my participants had mild struggles with organization and self-monitoring.

I assessed intercorrelations between the FrSBe subscales and found that executive dysfunction scale was significantly correlated with the other two subscales, which were not correlated with each other. In addition, the FrSBe apathy scale was negatively correlated with age (r = -.376, p = .018), although an ANOVA by three age categories was nonsignificant (p = .068).

There were no significant gender differences, but FrSBe total and executive dysfunction scores were significantly higher in the non-white participants (total p = .018, executive dysfunction p = .006), with FrSBe total means of 65.00+/-15.186 vs 84.25+/-5.500 and executive dysfunction means of 24.81+/-6.675 vs 34.75+/-4.425. However, since there were only 4 non-white participants, this should be replicated with a larger sample size.

Next, I assessed the correlation between AQ and FrSBe. The AQ total score was not significantly correlated with FrSBe total or any of the FrSBe subscales. However, FrSBe total was positively correlated with AQ attention to detail (r = .424, p = .011) and communication (r = .469, p = .004).

AQ communication was also significantly positively correlated with FrSBe disinhibition (r = .339, p = .035) and executive dysfunction (r = .411, p = .010), Meanwhile, AQ attention switching was positively correlated with FrSBe apathy (r = .357, p = .025) and AQ attention to detail was positively correlated with FrSBe executive dysfunction (r = .474, p = .003).

Sunday, June 19, 2016

Autism In Adults: A Survey - Part 1: Background and AQ

In 2012, I posted a self-report autism survey to the Wrongplanet forums. I got a good turnout, but ran into some issues with a couple of the measures I used, and ended up running out of spoons and abandoning the survey. Just recently, I decided to get back to the survey and finish analysing the results.

I got 45 respondents in total. Their ages ranged from 17 to 60 years, and 88.9% reported their ethnicity as white. The gender ratio was different from most autism samples, since 53.3% identified as female, 35.6% as male and 11.1% as other or transgender. In my experience, autistic women are more likely to frequent online forums related to autism than autistic men are.

Overall, 73.3% reported at least one autism spectrum diagnosis, with 60% reporting a diagnosis of Asperger Syndrome, 6.7% PDD NOS, 8.9% autism, 2.2% semantic pragmatic language disorder and 2.2% nonverbal learning disability.

In addition, 20% of the sample reported a diagnosis of ADHD, 6.7% sensory processing/integration disorder, 6.7% obsessive compulsive disorder, 4.4% dyspraxia and 2.2% other learning disabilities. In the 'other' textbox, 11.1% mentioned an additional diagnosis, with 6.7% each mentioning anxiety, depression and selective mutism.

On average, individuals reported having 1.51+/-1.392 diagnoses, ranging from 0-8. Seven individuals (15.6%) reported no diagnosis at all, 46.7% had one diagnosis, 24.4% had two diagnoses and 13.3% had more than two diagnoses.

The first questionnaire they filled out was the Autism Quotient (AQ). Their scores on the AQ ranged from 25 to 48, with a mean of 38.66 and a standard deviation of 5.851. The recommended cut-off for the AQ is 32, and 84.4% scored above that cutoff. This is actually higher than the AS/HFA reference sample in the original study of the AQ (p = .05), who had scores ranging from 18 to 48, a mean of 35.8 and 79.3% scoring above cutoff. Therefore, it's quite likely that our sample consists entirely of autistic individuals, even though 26.7% did not report an autism spectrum diagnosis. In a previous study of Wrongplanet members, I found that self-diagnosed and suspected autistics did not differ significantly from diagnosed autistics on the AQ. Similarly, in this sample, AQ scores did not differ by age, gender or diagnostic status.

The AQ has been divided into several subscales. On each subscale, my sample also scored significantly higher than the reference AS/HFA sample.

  • AQ social skill mean 8.4+/-1.502, range 5-10
  • AQ attention switching mean 9.05+/-1.154, range 6 to 10
  • AQ attention to detail mean 6.95+/-2.342, range 3 to 10
  • AQ communication mean 8.12+/-1.549, range 5 to 10
  • AQ imagination mean 6.21+/-2.030, range 2 to 10
Like the total AQ, most of the subscales did not differ by age, gender or diagnostic status. However, AQ communication was significantly negatively correlated with age (r = -.381, p = .015), suggesting that social communication may improve throughout adulthood. However, a three-way ANOVA between 17-24 year olds, 25-35 year olds and 36-60 year olds was not significant, probably due to small sample size.

In addition, AQ attention switching differed by gender, with men scoring significantly higher (p = .048) and having a smaller range of scores (p = .042) than women. This suggests that a subset of autistic women have lower repetitive behaviour than other autistic individuals. However, it may also be a bias with the AQ, as many of the items in this subscale reflect interests that are stereotypically male.

Tuesday, June 14, 2016

The Dangers of Hand-Over-Hand

Hand-over-hand is nearly ubiquitous in the teaching of developmentally disabled children. But there are important reasons to be concerned about it. Indeed, hand-over-hand could even be harmful, for several reasons. (Note: In this discussion, I'm using 'teachers' in the general sense as 'people who teach', regardless of their job description.)

Tactile Defensiveness
Many autistic people, and some people with other developmental disabilities, have a characteristic known as tactile defensiveness. Tactile defensiveness is a sensory processing issue in which the person finds certain ordinary tactile sensations overloading or possibly even painful. People with tactile defensiveness have compared it to 'rubbing sandpaper on their skin', getting an 'electric shock', and similar descriptions.

For a person with tactile defensiveness, hand-over-hand could be quite distressing. And of course, when you are feeling overloaded or in pain, that's not a state conducive to learning. People with tactile defensiveness often report finding a tactile sensation easier to tolerate if they can predict and/or control the sensation, but this is usually not the case for a child receiving hand-over-hand. Teachers generally don't ask permission (especially with students who can't communicate verbally), and they may not even warn the child first.

It's important to note, also, that not everyone with tactile defensiveness will pull away. Sensory overload can sometimes trigger a 'freezing' response, which can be difficult to overcome. (I've experienced this in response to loud noises, such as a fire alarm.) In addition, some people may have been trained out of resisting by prior experience with teachers who didn't let them pull away.

Abuse Triggers
Unfortunately, children with developmental disabilities are at an elevated risk of abuse. One study found that, in a sample of DD children with an average age of 9 years, 61% had experienced abuse or neglect at the hands of a caregiver. That's two thirds of children! When I hear these high numbers, I want to dismiss them, I want to ignore them, but the data is clear. Study after study have found similar high rates of risk.

Individuals with a history of abuse often see touch as potentially dangerous. Physically abused individuals may be observed flinching, expecting a blow, when others make sudden movements. Sexually abused individuals often feel a strong need for control over how and by who they are touched. Both groups may find that touch - especially unexpected touch and touch without permission - triggers feelings associated with the abuse. In severe cases, they may even experience sensory re-experiencing, literally hallucinating sensations linked to the abuse.

You may not know if an individual you are working with has been abused. Judging from the statistics, at least 2 out of 3 of them probably are. Even individuals with good communication skills may not be able to talk about abuse - if the person also struggles with communication, it may be completely impossible for them to report any abuse.

Abuse Prevention
It's also important to remember the possibility of future abuse. Even if a particular individual has not been abused yet, they still remain at risk. Sexual abuse, in particular, is often perpetrated by non-caregivers, or by caregivers that the child has only recently met. And while it is never the child's fault that they were abused, research has suggested that there are things children can do to resist sexual abuse - a child who strongly protests and resists is less likely to be victimized.

Among non-disabled children, one major factor in abuse risk is how well the child can refuse an unwanted touch from an adult. This is a common features in abuse prevention programs, which have been found to reduce actual abuse risk. Children who can firmly refuse an adult's touch will sometimes discourage an abuser from taking the abuse further, and are more likely to report the abuse. Obviously, the benefits on reporting abuse might not apply to a minimally verbal child. But a child who resists could discourage an abuser, or elicit attention from others which discourages the abuser.

What impact would hand-over-hand have on a child's ability to resist an abuser? Unless you have asked the child permission and gotten verbal or nonverbal consent, by using hand-over-hand, you are communicating that you can touch them whether or not they want it. And if the child resists and you don't let go, you are teaching them that resisting your touch is futile. And what happens if they apply those lessons to an encounter with an abuser?

Alternatives
So, if hand-over-hand has so many problems, what else can you do?

First, you can use non-touch ways of teaching. The PECS program leaps to hand-over-hand without any attempts at any other form of prompting or modeling. However, modeling use of an AAC system by using it yourself has been found to be an effective teaching method for many children. Similarly, video or picture modeling of sequential activities is a proven learning technique for autistic children.

Shaping is another non-touch way of teaching. This is how a typical baby's 'mamamam' morphs into 'mama' - mother reacts to anything that sounds like 'mama', teaching the child that his or her babbling has meaning. All sorts of behavior can be shaped in a similar fashion. As the behavior becomes established, move the goalposts to a closer approximation of the desired behavior.

If you must use touch, take some lessons from the recommendations for visually impaired individuals. The VI field has done a lot of thinking about how to create predictable and respectful touch. Warning about touch before touching them is a big thing. Even if you're not sure they understand, warn anyway.

Better yet, ask permission. Even nonverbal individuals may be able to communicate whether or not they are willing to be touched. For example, if you ask a child 'Can I help you do X?' and reach towards their hand, they might pull away and make a protesting sound or they might soften their hand and offer it to you. Those behaviors are nonverbal signals of whether the child is willing to be touched. And certainly if the child refuses, don't pressure them. (With other forms of touch, such as changing their diaper or stopping them from doing something dangerous, you may have to ignore a child's refusal. But providing a lesson on how to perform a motor activity is not worth overriding a child's refusal.)

VI therapists also recommend a practice called hand-under-hand. This practice is basically the opposite of hand-over-hand - instead of placing your hands on the child's hands, you place your hands under the child's hands, so they can feel what you are doing but are free to pull away if they wish. Again, it's still important to give warning and try to get consent, but even if you overstep things, it's easier for the child to stop the touch at any time they choose. Hand-under-hand is not just for VI children - any child who doesn't learn well through visual demonstration can benefit.

In general, be careful when you touch a vulnerable individual. Touch can be hurtful if it's not used carefully.

Tuesday, May 17, 2016

We Can't All Get Along

I came across this article recently.

Unfortunately, comments were closed on it, explicitly because it 'could turn bad and hurt people' and the owner of the community 'wants all people who come here to feel safe to express something'. A strange justification for shutting down conversation about a serious issue, in my opinion. But in any case, that's why I'm posting a blog entry about it instead of just having my say in the comments.

Essentially, the article imagines a hypothetical disability conference that brings together everyone in the 'disability community' to speak our piece with one rule - no one can say they're offended by something. If you get offended, and show it, you'll be booed out of the hall. This is painted out as a wonderful thing, but it sounds absolutely terrible to most self-advocates.

What the 'let's all get along camp' don't seem to get is that there really are irreconcilable differences. There are people in the 'disability community' who think people like me are better off dead. This is not hyperbole, this is serious. There are also people in the 'disability community' who openly endorse practices that I have nightmares about, practices that deeply wounded me as a child. (They weren't practiced under the same name, because I was undiagnosed, but the damaging aspects are the same.)

And I've had it relatively good, among autistic self-advocates. There are many people who have been hurt far worse than me. Think about it - this 'inclusive conference' would invite both Issy Stapleton and her mother - who, if you've followed the news story, you'll know has been convicted of trying to murder Issy. This conference would place a survivor of an attempted murderer in the position of being asked to listen to her murderer defend her actions and not take offense at that. And if she had the reaction that most victims of such a serious crime would have, she'd be 'booed out of the hall'.

There are also people who've gotten seriously hurt over less serious and more widely accepted practices in the disability community. Practices such as restraints, ABA, genital examinations, and others have been reported by self-advocates to have caused significant trauma and PTSD symptoms. It's not reasonable to expect someone not to get offended when you're describing their traumatic experiences and saying it's a good idea to do these things. It would be an exceptional person who could stay calm under those circumstances.

If you're a supporter of LGTB rights, imagine a conference like this one, but about LGTB issues. And we invite everyone - not just LGTB people and PFLAG members and supportive doctors and therapists, but also ex-gay ministries, church leaders, parents who abuse their LGTB kids, therapists like Kenneth Zucker (who teaches trans kids to act their birth gender), everyone who has an interest in LGTB people in any form.

And then you say that even if someone is endorsing practices that drove you to past suicide attempts or caused years of misery in your life, you're not allowed to get offended. If you act offended, your voice will be shut out.

Does that really sound like a good thing?

Thursday, April 28, 2016

Are We Getting Smarter?

I have a book somewhere called Are We Getting Smarter?. It's written by James R Flynn, the man for whom the Flynn effect is named. For those of you who haven't heard of it before, essentially, every decade the average IQ increases. It's estimated that using an IQ test normed in 1997, the average IQ in 1932 would have been 80 (it should be 100).

Flynn's argument is that this change does not reflect actual intelligence. Instead, it reflects changes in testability. Essentially, the increased exposure to tests, due to higher education rates, higher rates of testing in education, and so forth, makes people better at taking tests, and this increases their IQ.

This is certainly possible. But I doubt it accounts for all of the increase. There are a lot of reasons why we should actually be getting smarter, and not just better at taking tests.

First, nutrition. Severe malnutrition in infancy can lower a child's IQ by around 20 points compared to better fed children. In regions where many children are malnourished, a significant correlation between height and IQ is generally found, because both are reduced by malnutrition.

There's an increase in the rates of obesity and obesity-related diseases throughout most developed countries. This is a concern, certainly, but the bright side of this change has been a decline in the rate of severe malnutrition in all age groups, including children. This is also shown by the historical increase in average height in these countries. Further, we've also seen the discovery of vitamins in 1912 and the first vitamin supplements in the 1930s.

In addition, nutritional status for infants and pregnant women has been subject to particular changes. Like all adults, pregnant women have had an increase in overall nutrition. However, they have also been marketed far more nutritional supplements than the general population, and are more likely to take these supplements. Birth control has also improved the health and nutritional status of mothers. Larger families cost more to feed, and repeated pregnancies put more strain on the mother's body - especially if she's also breastfeeding. With birth control, women have more control over their rate of birth, and typically choose to have a small number of children who are well-spaced apart.

And speaking of breastfeeding, the ratio of bottle to breastfeeding has gone through two distinct transitions. Most of us today are taught that 'breast is best', but when formula was first invented, this was not necessarily the case. A malnourished mother will generally produce poorer-quality breastmilk, to the point where if she has access to clean drinking water and enough quality infant formula, her baby could very well be more healthy if formula fed. (In modern times, unfortunately, many malnourished mothers lack access to safe water and/or can't afford to buy enough formula to actually meet their child's needs. If the formula is watered down to make it last, or if the water used to mix the formula contains pathogens, breastfeeding is definitely better regardless of maternal nutritional status. However, due to the lack of birth control in that time period, many mothers from otherwise affluent backgrounds were malnourished purely because of back-to-back pregnancies.)

By the 1970s, however, maternal nutrition had improved substantially. It was around this time that the shift back to breastfeeding began, in part spurred on by evidence that breastfed infants appeared to be doing better than formula fed infants. Soon after, the first studies linking breastfeeding to higher IQ were performed.

Lastly, we've seen the adoption of public vaccination programs. The intended impact, of course, was to reduce the rate of serious complications from viral illnesses, such as congenital rubella syndrome and measles encephalitis (both of which can cause severe cognitive disability). Those complications are rare even when infection rates are high, and therefore would have only a minimal impact on average IQ. But illnesses also interact with malnutrition. Fighting illness takes resources, and a child who is frequently sick will need to eat more when healthy to compensate for the work of fighting off infections. A child who is both malnourished and frequently sick will be in a worse nutritional state than a child with the same diet who has been vaccinated for those illnesses.

So changes in infant and childhood nutrition and health certainly can account for some increase in IQ in the past hundred years. But this isn't the only factor that we've changed.


From ancient times, lead has been used to make many different things, such as pipes for drinking water. The Romans used lead pipes despite at least some awareness of the danger of lead poisoning. Medieval Europeans, on the other hand, seem to have forgotten that there were any dangers to lead, and this continued into the industrial and modern period, when the uses of lead extended from pipes to include paint, gasoline, and other things.

The clinical symptoms of acute lead poisoning are severe and obvious, but such poisoning has always been fairly rare. However, chronic low-level lead exposure in the first five years of life has been shown to lower IQ slightly in children without any clinical symptoms of lead poisoning. Such exposure would have been nearly ubiquitous before we knew the dangers of lead, and has been declining steadily since we removed lead from gasoline and paint and began gradually removing existing lead sources from our lives.

And lead is not the only environmental toxin we have reduced our exposure to, though it's the best documented. While there are probably some new toxins in our environment that we don't yet know the risks of, overall, we have gotten much more careful about what we have in our food, drinking water and the air we breathe - especially for young children. Though we don't know the impact of many of these other toxins on IQ, it's likely that at least some of them can decrease IQ in children.

Lastly, there are also nonbiological environmental effects on IQ. During the first 2 years, the brain is actually pruning unneeded neurons to make room for the ones we will need. And one big determination of whether a neuron makes the cut is how it's being used - for example, a 6 month old baby can distinguish all the possible phonemes in all human language, but by 9 months, babies can only distinguish phonemes that are important in the languages they've heard regularly spoken around them. (For example, a 9 month old exposed to only Japanese will have lost the 'l' and 'r' distinction.)

This becomes particularly crucial when we consider the most disadvantaged children in society. At the beginning of the last century, if a child was orphaned or their parents couldn't care for them, society's answer was to put them in an institutional setting. A child who spends the first few years of life institutionalized will frequently end up with an IQ in the borderline to mild cognitive disability range, as well as suffering a wide range of behavioural and emotional problems. Fortunately, with greater awareness of the impact of orphanages on child development, most developed countries have eliminated or greatly reduced their use - replacing them with foster care, which, while still problematic emotionally, does not have a noticeable effect on the child's IQ score.

But it's not just orphanages that can result in pruning important neurons because of insufficient stimulation. Among children living in family environments, children who are victims of parental neglect typically show a decrease in IQ compared to adequately cared for children. The most dramatic examples are children like Genie, who spent the first 13 years of her life in a single room, chained to a potty chair for most of the time. Although it's uncertain whether Genie's IQ was normal to begin with, her early pre-isolation development was definitely not consistent with the severe cognitive and language impairments she showed in her teens and adulthood.

Of course, cases like Genie are extremely rare, but many more children are exposed to subtle neglect. For example, a parent who regularly leaves her baby to be babysat by a five year old sibling not only places the physical safety of both children at risk - this also results in the baby being exposed to less adult conversation and less competent scaffolding of early interaction and play. Even a parent who is suffering from serious depression tends to interact less with their infant, resulting in poorer language and social skills.

The good news is that exposure to child maltreatment is also decreasing. It used to be that children were not apprehended at all, even for the most severe abuse - only children who were orphaned or willingly relinquished wound up in state care. Since then, more and more children are being removed against their parent's will, including neglected children like the hypothetical five year old babysitter and infant sibling described above.

Even when children are not removed, their care is improving. Psychiatric treatments and parenting skills programs are more readily available and have become more effective. For example, before the invention of antidepressants, the only treatments for depression were hospitalization or expensive psychotherapy, neither of which were as effective. Several decades later, CBT was also invented, and has since become the front-line psychotherapy for depression. Both antidepressants and CBT are as effective in treating depressed parents as they are in treating anyone else, and are certainly reducing the rate of infants exposed to chronic parental depression. Parent-specific programs, such as parental sensitivity training, have also increased tremendously.

In addition, we have also seen an increase in programs aimed specifically at children. Head Start, a program that provides education and support to toddlers and preschoolers from low income families, was first implemented in the 60s. Participation in Head Start appears to improve cognitive ability in children. The same effect may also be seen among low-income children attending preschools in general - and certainly preschool attendance has increased tremendously.

Wednesday, April 13, 2016

Parenting Impact on Autistic Kids: You Can't Have Your Cake and Eat It Too

In the 1960s, autism (and childhood schizophrenia, which included many kids who'd now be diagnosed with autism) was thought to be caused by bad mothers. The theory referred to 'refrigerator mothers' - mothers so cold and distant that the child turned to autism as a coping strategy.

Now, of course, we know this is nonsense, and very hurtful to the mothers who were so unjustly blamed. Parents of autistic kids don't consistently differ from parents of non-autistic kids in their parenting skills. But many people in the autism community* go too far in the opposite direction - denying any impact of parents on their autistic kids.

What they don't seem to realize is that if parenting styles don't affect autistic kids, then a lot of autism therapies would also be useless, because these therapies involve deliberately and systematically doing things that some parents do on their own.

The easiest example is relational therapies such as DIR and Floortime. The interactional style that the therapists take in these therapies is pretty much the same as the parenting dimension known as 'parental sensitivity' - a very well studied parenting dimension that has a lot of important implications for child development in both typical and disabled children. Parents high in parental sensitivity tend to have children who are more securely attached, have fewer behaviour problems, have better social-cognitive skills, and even have slightly better language skills (especially if they have a disability affecting language development, such as deafness).

Based on this, we would predict that an intervention mimicking sensitive parenting behaviour should reduce behaviour problems, improve social skills, and improve language development, as well as improving attachment security. At least two of those effects has been documented as a result of relational therapies, with this study, this study, this study and this study all finding improvements in social skills and this study finding improvements in expressive language in autistic children receiving relational therapies. But in order for this treatment to work, autistic kids must also be affected by naturally-occurring differences in parental sensitivity (such as differences due to the parent's own attachment style or marital conflict).

ABA is less easily equated to parenting styles, because there are two distinct aspects to ABA - direct teaching and prompting of skills, and consistent rewards and punishments to modify behaviour. In parenting styles research, those two components split up into separate domains of parenting behaviour.

The impact of consistent rewards and punishments has been very extensively studied under the dimension of consistent discipline. Children who get consistent discipline tend to show lower behavior problems and better attention and impulse control. So it stands to reason that ABA, which includes consistent discipline, would reduce behaviour problems, and the research supports this. But similarly, naturally occurring variations in how consistently parents discipline their children (such as differences due to parental depression) must also affect autistic children's risk of behaviour problems.

Parental teaching has been studied far less than consistent discipline. However, parental direct teaching is associated with improved emergent literacy skillsmathematics skills and earlier toilet training (most parents directly teach toilet training, but the parents who start earlier tend to have kids who are fully toilet trained earlier). So, at the very least, this suggests that ABA's direct teaching should improve academic skills and self-care skills. This study and many others have found that ABA improves self-care, but I couldn't find any data on ABA's impact on academic skills. Once again, if ABA can directly teach skills to autistic kids, individual differences in parents' tendency to directly teach skills must also affect their kids.

Almost all of the practices commonly used in autism therapy are also things that a subset of parents do on their own. So it's intellectually dishonest to simultaneously claim that these therapies can affect autistic kids' development and at the same time insist that naturally occurring differences in their parents' behaviour can't also affect their children. And if you claim those treatments can cure autism (a claim that isn't supported by the data, by the way), then parents who do the same things spontaneously should logically also be able to cure autism (or prevent it - a really early cure is indistinguishable from prevention).

If parents can't cause autism, they also can't cure it. And neither can a therapist who only does things that many parents do anyway.

* Note - I'm using 'autism community' to refer to the community of mostly parents and professionals, while 'autistic community' refers to the community of mostly autistic adults.

Tuesday, March 01, 2016

Are Most Autistic People Low Functioning? The Answer is NO!

You often see this claim in the comment section of news articles about high functioning autism. People say 'Well, they're the lucky few. Most people with autism are low functioning, like my child." And then they go on to describe their child is the most negative light possible.

Which made me wonder - what's the truth? I always figured that high functioning is more common, just like mild cognitive disability is more common than severe cognitive disability, because we're the tail end of a normal spectrum. But does the data back me up on that?

Now, it's important to keep in mind that low functioning autism tends to be diagnosed earlier and more easily than high functioning. So if we look only at already-diagnosed autistic people, low functioning will be over-represented. I will be looking at autism screening studies on a random selection from the general population, finding ones that a) provide some information relevant to functioning level, and b) used a design capable of detecting both extremes of functioning (eg screened children older than toddler-aged, did not select based on a sign of good functioning such as attending mainstream schools). In addition, I will only be looking at studies published in 1990 or later, and they must have found at least 15 autistic people.

Next question is how to define functioning level. I've seen a number of definitions - IQ score, language level, adaptive functioning, even presence or absence of certain behavior problems. It gets complicated. In this analysis, I'll be focusing on IQ score, language level and adaptive functioning, using the following definitions:

High functioning autism (all three of these):
* normal or above average baseline verbal functioning (though may have nonverbal episodes due to overload or other issues)
* normal or above average IQ
* adaptive functioning is mildly impaired or better (note - studies usually find a gap in IQ and adaptive functioning among autistic people with normal IQ)

Low functioning autism (at least two of these):
* baseline minimally verbal or nonverbal (though may use AAC devices)
* IQ score less than 50
* adaptive functioning is moderately to profoundly impaired

Anyone who doesn't meet criteria for either group is medium functioning.

If data is given on only one of these metrics, I'll base my judgments of functioning level on that metric alone, and make it clear that I'm doing so. Unfortunately, I only found three relevant studies, and all reported solely on IQ.

The first study was performed in a South Korean community in 2011. They screened both a random population sample of 7-12 year old children and a high risk sample, but I'll only discuss the general population sample here. They found a prevalence of 2.64% autistic kids in the general population sample, and ascertained 201 children.

In this study, functioning level data was based on IQ. The autistic kids from the general population sample had an average IQ of 98, which is clearly in the normal range. Indeed, an IQ of 98 is not significantly different from the general population average of 100. Only 16% of the children had an IQ less than 70, with the percentage less than 50 not being reported (note - in the general population, 3.5% have an IQ less than 70). Indeed, 12% had IQs over 120, in the high-average to gifted range. Therefore, the proportion of autistic kids in this sample who are high functioning is estimated at 84%.


The next study screened children in two UK communities in 2001. They screened 2.5 to 6.5 year old children. They found a prevalence of .6% and ascertained 97 children.

In this study, functioning level data was based on IQ. They do not report the average IQ, but 25.8% of the children had an IQ (or DQ, for the younger children) less than 70, with the prevalence of IQ less than 50 not being reported. Therefore, the proportion of autistic kids in this sample who are high functioning is estimated at 74.2%.

The last study was performed in Toyota, Japan in 2008. All children were screened for autism at 18 month and 36 month check-ups. They found a prevalence of 1.81% and ascertained 228 autistic children.

In this study, functioning level data was based on IQ. They found that 66.4% had an IQ of 70 or higher, and 16.1% had an IQ of 50 or less. Therefore, the proportion of autistic kids in this sample who are high functioning is estimated at 66.4% and the proportion who are low functioning is estimated at 16.1%.

The three studies all found a very high proportion of high functioning children, with 66-84% of the autistic kids having an IQ over 70. While not all of these children will be high functioning according to my criteria, most probably are. In addition, the older the sample of children, the higher the proportion of high functioning children, suggesting that HFA may be more difficult to diagnose in children under age 4, or that some kids may move from medium-low functioning to high functioning during this period. However, even the Japanese study, which performed its' second screening at 3 years of age, found a majority of children with average IQs.

Only the Japanese study provided data on how many autistic children had an IQ of less than 50, finding 16.1%. However, the other two studies almost certainly had even lower prevalence - particularly the South Korean study, which found 16% with IQs below 70.

Clearly, those commenters are wrong. Descriptions of high functioning autism are actually a far better representation of the majority of autistic people than descriptions of low functioning autism. Despite the scare tactics used by many 'awareness' campaigns, most autistic people have an average IQ. The severe, low functioning end is actually a minority among autistic children.