I've got a new post over on the SFARI (Simons Foundation Autism Research Initiative) blog discussing the use of control groups in autism research.Control groups are an essential part of autism research, providing a benchmark against which to assess those with autism. Finding, for instance, that participants with autism score an average of 68 percent on a test is meaningless if you don’t know how people who don’t have autism do on the same test. A control group can also be used to try and rule out alternative and perhaps uninteresting explanations for group differences. The logic is simple: If two groups are matched on one measure, such as intelligence or age, then this can’t explain differences on another measure, such as performance on an emotion recognition test, that is under investigation. Despite its widespread use, there are many issues to consider when designing an experiment with matched controls or when reading and attempting to evaluate such a study. Who should be in the control group? On what measures should they be matched? And how do we decide if the groups are truly matched? The post focuses on this last question and a recent paper by Sara Kover and Amy Atwood, which I think makes some pretty sensible recommendations.Be warned, it involves statistics and made-up data.Reference:sKover ST, & Atwoo AK (2013). Establishing equivalence: methodological progress in group-matching design and analysis. American journal on intellectual and developmental disabilities, 118 (1), 3-15 PMID: 23301899Mervis, C. B., & Klein-Tasman, B. (2004). Methodological Issues in Group-Matching Designs: α Levels for Control Variable Comparisons and Measurement Characteristics of Control and Target Variables. Journal of Autism and Developmental Disorders, 24, 7-17. PDF... Read more »
Kover ST, & Atwoo AK. (2013) Establishing equivalence: methodological progress in group-matching design and analysis. American journal on intellectual and developmental disabilities, 118(1), 3-15. PMID: 23301899
Last weekend, the American Psychiatry Association announced that the DSM-5 has been ratified and that changes to the diagnosis of various “mental disorders” would go ahead as planned early next year. The implications for autism are still unclear. Most of the published studies suggest a reduction in autism rates. However, the argument from the DSM-5 working group responsible for autism has always been that everyone should wait for the official field trials to be published before jumping to any conclusions.The field trials were finally released a couple of weeks ago, with methods and results spread across three articles in the American Journal of Psychiatry, the APA’s own journal. If ever there were a case for scientific documents to be made freely available, this would be it. Instead, they’re hidden behind a paywall, costing $35 to “rent” each article for 24 hours. It’s taken me a couple of weeks to get my head around them (luckily my university library has paid for unlimited access to the journal). For those who are interested but don’t want to spend the kids’ Christmas present money on journal articles, I’ve done my best to summarise the study – at least those parts that are relevant to autism.The field trials were spread across 11 sites, all in North America. But only two – Baystate and Stanford - were involved in trialling autism diagnoses. Both centres were also assessing a number of other diagnoses, so autistic kids only made up a small proportion of those assessed. At Baystate, 23% of kids in their sample of 569 met DSM-IV criteria for an ASD (i.e., they already had a diagnosis of autistic disorder, Asperger syndrome, or PDD-NOS). At Stanford, the figure was 26% from a sample of 463. The main aim of the field trials was to assess reliability – whether two different clinicians would give the same person the same diagnosis. This meant that, to be included in the final sample, each child had to be assessed twice and, as a result, only a small fraction of the children initially screened into the study were included in the final analysis. At Baystate, 146 of 569 kids made it all the way through. At Stanford the figure was 149 out of 463.In an effort to make sure they had enough kids for each of the diagnoses, the kids were assigned to various “strata” corresponding to the DSM-5 diagnoses under investigation. To be in the ASD stratum, a child had to have a DSM-IV diagnosis of ASD. However, some of the kids with ASD also met criteria for other strata and, in an effort to even up the numbers, were assigned to those rather than ASD. At Baystate, this happened to 20 of the 132 ASD kids. At Stanford, 21 of 119 ASD kids were reassigned. However, there is no indication of which strata they were assigned to.First 6 columns taken from Paper I (Clarke et al.). Columns 7 and 8 calculated by me. Column 9 take from Paper II (Regier et al)Because of this biased sampling, the authors employed a complicated formula to estimate DSM-5 prevalence. This essential piece of information can be found in Footnote E of Table 1 of the second paper. Calculating the ASD prevalence was effectively done as follows:The authors calculated the proportion of kids in each stratum who met DSM-5 criteria for ASD (kids who were diagnosed with ASD by one clinician but not the other were effectively treated as half an ASD kid). They then multiplied this by a weighting factor, which corresponds to the proportion of kids in the original sample who were assigned to that stratum. Having done this for all the strata, they added up all the values to get the prevalence. At Baystate, estimated prevalence was 24%, a slight increase from DSM-IV (23%). At Stanford, estimated prevalence went down fairly dramatically to 19%, compared with 26% under DSM-IV.One of the slightly weird consequences of this approach is that different kids would have ended up contributing different amounts to the ASD prevalence depending on which stratum they’d been assigned to. For example, at Baystate, a child assigned to the non-suicidal self-injury strata would have been “worth” double a child assigned to the bipolar disorder strata. Having scratched my head about this for a while, I think it does make sense – but only if sampling really was deliberate. If it wasn’t, then there’s no reason to weight a child with ASD differently just because they weren’t assigned to the ASD stratum.My suspicions here are raised by the non-suicidal self-injury stratum. Despite being one of the rarest target diagnoses, this was wildly under-sampled. This suggests that sampling was really a process of assessing as many kids as possible before time ran out. In that case, the formula could seriously distort the true DSM-5 prevalence.A much simpler and more transparent way to compare DSM-IV and DSM-5 rates would have been to look only at the kids in the final sample and ask how many had an ASD diagnosis under each diagnostic scheme. The authors told me that, across both centres, 79 of the kids in the final group met DSM-IV criteria for ASD (note that only 64 of these were in the ASD stratum, which is why most reports have mistakenly said there were 64 ASD kids in total). However, the authors haven't responded to my question about numbers in DSM-5.One thing they did tell me is that another paper is being prepared that looks in more detail at the autism results. In addition to the actual numbers of kids diagnosed under DSM-IV and DSM-5, we’re also currently missing information about the make-up of the autism groups at both centres. If anyone is going to miss out on ASD diagnosis, it’s likely to be the less clear-cut cases – those who'd meet criteria for Asperger’s or PDD-NOS under DSM-IV.The authors cheerfully conclude that children missing out on an ASD diagnosis will be better served by the new Social Communication Disorder (SCD) diagnosis:"A careful review of data from both sites showed that the decrease at the Stanford site was offset by movement into a new DSM-5 diagnosis called social (or pragmatic) communication disorder (data not shown). Since autism spectrum disorder requires both deficits in social communication and fixated interests/repetitive movement, the more specific deficit assessments in DSM-5 should facilitate more focused treatments for those with social communication deficits only."Given that most autism interventions are targeted at social and communication difficulties, it's not clear what "more focused treatments" the authors have in mind. And, as many people have noted, there is currently no mandated provision for SCD anywhere in the real world. Indeed, the press release announcing the ratification of DSM-5 made no mention of SCD - despite the fact that it would be a major change to diagnostic practice. As with so many other aspects of DSM-5, your guess is as good as mine.References:Clarke DE, Narrow WE, Regier DA, Kuramoto SJ, Kupfer DJ, Kuhl EA, Greiner L, & Kraemer HC (2012). DSM-5 Field Trials in the United States and Canada, Part I: Study Design, Sampling Strategy, Implementation, and Analytic Approaches. The American journal of psychiatry PMID: 23111546Regier DA, Narrow WE, Clarke DE, Kraemer HC, Kuramoto SJ, Kuhl EA, & Kupfer DJ (2012). DSM-5 Field Trials in the United States and Canada, Part II: Test-Retest Reliability of Selected Categorical Diagnoses. ... Read more »
Clarke DE, Narrow WE, Regier DA, Kuramoto SJ, Kupfer DJ, Kuhl EA, Greiner L, & Kraemer HC. (2012) DSM-5 Field Trials in the United States and Canada, Part I: Study Design, Sampling Strategy, Implementation, and Analytic Approaches. The American journal of psychiatry. PMID: 23111546
Regier DA, Narrow WE, Clarke DE, Kraemer HC, Kuramoto SJ, Kuhl EA, & Kupfer DJ. (2012) DSM-5 Field Trials in the United States and Canada, Part II: Test-Retest Reliability of Selected Categorical Diagnoses. The American journal of psychiatry. PMID: 23111466
One of the big concerns about forthcoming changes to autism diagnosis in DSM-5 is that the new rules may miss out a sizeable chunk of the autism population, particularly those who, under current diagnostic guidelines, would be considered to have Asperger’s Disorder or PDD-NOS. A number of studies published over the last year or so seemed to confirm those fears, although each of those studies had its limitations.Earlier this month, a study in the American Journal of Psychiatry by Marisela Huerta and colleagues appeared to show quite conclusively that almost every autistic person, including people in the PDD-NOS bracket, would qualify for an Autism Spectrum Disorder diagnosis in DSM-5.Emily Singer has a nice summary on the SFARI blog, complete with interviews with various researchers all agreeing that the study really seems to have settled the issue.However, I'm not so sure. Here's my attempt to explain why.Huerta et al.'s approach was to map items from the ADI-R and ADOS diagnostic tests onto "symptoms" in the DSM-IV and DSM-5 criteria. They then looked to see whether, for each individual assessed, the combination of symptoms qualified them for a diagnosis under DSM-IV or DSM-5. It's important to note that they didn't use the official algorithms for translating item scores into diagnoses. Instead, they asked whether, for each symptom, there was evidence from at least one item in ADI-R or ADOS.Next, they compared the diagnostic outcome of the DSM-IV and DSM-5 recodings to what they call "clinical best estimate diagnosis". I was a little unsure what this meant, so I emailed Dr Huerta, who confirmed that this was essentially the opinion of experienced clinicians at the time of the original assessment. The diagnosis was guided by DSM-IV and was based partly on outcomes of the ADOS and ADI-R, as well as developmental history and performance on some standardized tests.Have a look at Table 2 below. This is actually just the top third of the table as the data Huerta et al used comes from three different sources. For now, we're just looking at data supplied by the Collaborative Programs of Excellence in Autism.Table 2 from Huerta et al (2012). Click to enlarge.We're interested here in the kids with best estimate diagnoses of PDD-NOS or Asperger syndrome. I've highlighted in red the row entitled "Either parent or clinical measure". This means that a symptom box could be ticked if it was present in either the parent measure (ADI-R) or clinical measure (ADOS).The first column of figures shows the sensitivity and specificity according to the proposed DSM-5 criteria (columns to the right of this show outcomes for potential tweaks to the DSM-5 criteria). In this sample, 94% of kids with best estimate diagnoses of Asperger’s or PDD-NOS met the DSM-5 criteria for Autism Spectrum Disorder (see blue oval). Across the three datasets, the value was 96%.But now look at the column headed "Autistic Disorder". This shows the proportion of kids meeting DSM-IV criteria for Autistic Disorder according to the recoding. In this case, the figure is 84% (green oval).This is distinctly odd. According to DSM-IV, PDD-NOS and Asperger's disorder should only be diagnosed if Autistic Disorder has been ruled out. But the authors are reporting that, of the kids with a best estimate diagnosis of PDD-NOS or Asperger's disorder, 84% actually met criteria for Autistic Disorder under the recoding. For the other two datasets, it's even more extreme. The corresponding figures are 90% and 97%. Overall, it works out that 94% of kids diagnosed with PDD-NOS or Asperger's actually meet criteria for Autistic Disorder.Now look at the specificity, highlighted in orange. Across the three datasets, specificity for DSM-5 was 33% (purple oval). In other words, two thirds of the non-PDD kids met DSM-5 criteria for Autism Spectrum Disorder. We're not talking about autism risk here. This relates to actual diagnosis. 33% specificity would be terrible.For DSM-IV it's even worse. Specificity is 10%. That means that 90% of the non-PDD kids met criteria for PDD-NOS (in fact two thirds of them met criteria for Autistic Disorder). Something is not right. 94% of kids who were considered by clinicians to have PDD-NOS or Asperger's ended up meeting criteria for Autistic Disorder under the recoding. And 90% of kids who weren't considered autistic at all ended up meeting criteria for PDD-NOS.One possibility is that the clinical best estimates are wrong. If this were the case, it would completely undermine the claim that most kids with PDD-NOS or Asperger's will meet DSM-5 criteria - because the study didn't actually include any such kids. Or at least, if it did, they were all in the non-PDD group!More likely I think is that the coding scheme used in this study is far more liberal than best estimate clinical diagnosis. Getting a diagnosis from a real life clinician requires more than just a single example of each relevant behaviour. If this is the case then it again undermines the study's conclusions. If the recoding is too liberal, if boxes are being ticked that would not ever be ticked by a clinician conducting a real best estimate diagnosis, then we can't trust the DSM-IV coding. And by the same token, we can't trust the DSM-5 coding either.Emily Singer's article has an interesting quote from Cathy Lord, who's a member of the workgroup responsible the DSM-5 changes to autism diagnosis, as well as being a co-author on the Huerta et al. paper."It's still not the same thing as taking the new criteria and testing them out, which is why we didn't do this analysis before," says Lord. "But clearly people have been analyzing much more restricted datasets, so we thought we better get in here and do it."I think she's absolutely right to worry about the limitations of recoded data. But we need to be skeptical whether or not the results support DSM-5.Reference:Huerta M, Bishop SL, Duncan A, Hus V, & Lord C (2012). Application of DSM-5 Criteria for Autism Spectrum Disorder to Three Samples of Children With DSM-IV Diagnoses of Pervasive Developmental Disorders. The American journal of psychiatry, 169 (10), 1056-64 PMID: 23032385... Read more »
Huerta M, Bishop SL, Duncan A, Hus V, & Lord C. (2012) Application of DSM-5 Criteria for Autism Spectrum Disorder to Three Samples of Children With DSM-IV Diagnoses of Pervasive Developmental Disorders. The American journal of psychiatry, 169(10), 1056-64. PMID: 23032385
This week's big autism story was a genetic test able to predict with 70% accuracy  whether or not a child had autism. Rather than looking for a specific gene that might differentiate autistic from non-autistic people, Stan Skafidis and colleagues developed the test by combining information about many different genetic variations. Critically, having developed the test based on one set of genetic data, they then tested the test on genetic data from a completely new set of people.I don't want to take anything away from the basic science. And it would, of course, be incredibly useful to know early on whether or not a child is likely to develop autism. But the headlines are misleading. The unfortunate truth is that we're still a long way from a genetic test for autism. A screening measure with 70% accuracy would only be slightly better than completely useless.Here's why.Say you had 1000 kids and you ran the genetic test to see which ones would become autistic. If we assume that the rate of autism in the population is around 1% then we'd expect 10 of the 1000 kids to be autistic. Given 70% accuracy, we'd expect 7 to show up on the genetic test as autistic.The problem is the other 990 who aren't autistic. 70% accuracy means that 30% would be incorrectly diagnosed as autistic. 30% of 990 is 297.Putting those together, our genetic test thinks that 304 of the 1000 children are autistic. Of those, only 7 really are autistic, the other 297 have been mis-diagnosed.In other words, if your child took the test and the test came out positive, there would still be a 98% chance that your child was not autistic.This screener's fallacy is a well known problem. Dorothy Bishop covered it when discussing screening for autism based on speech recordings and MRI scans. Ben Goldacre talked about it in relation to screening for terrorists. It's a problem any time you're looking for something that's relatively rare. Even a really good test gets torpedoed by a disasterous rate of false positives.To be fair, the researchers did suggest that the test would be most useful for parents who already had one autistic child. The latest research suggests that, if you have an autistic child, the chances of a second autistic child may be as high as 19%. Even in that situation, you'd still expect roughly two thirds of the kids picked up by the test to actually be non-autistic.Notes "Positive and negative predictive accuracies were 70.8 and 71.8% respectively"ReferenceSkafidas E, Testa R, Zantomio D, Chana G, Everall IP, & Pantelis C (2012). Predicting the diagnosis of autism spectrum disorder using gene pathway analysis. Molecular psychiatry PMID: 22965006 Full TextRelated postCan MRI scans be used to diagnose autism?... Read more »
Skafidas E, Testa R, Zantomio D, Chana G, Everall IP, & Pantelis C. (2012) Predicting the diagnosis of autism spectrum disorder using gene pathway analysis. Molecular psychiatry. PMID: 22965006
Last Monday, the Australian current affairs program Four Corners featured a Canadian documentary entitled The Autism Enigma (it’s still currently available on iView in Australia). The main thrust of the program was that autism is caused by harmful bacteria in the gut. Not surprisingly, the program caused quite a stir, prompting responses from various members of the Australian autism research community. Critics argued that it over-simplified the problem of autism, ignored alternative explanations, promoted interventions that weren’t evidence-based, and offered false hope to parents.Not knowing very much about this area of research, I’ve since been reading around some of the studies that were mentioned in the program.Trailer for the Canadian showing of The Autism EnigmaThe vancomycin trialOne of the key studies featured in The Autism Enigma was a pilot study of antibiotic treatment in 11 autistic children, conducted by Dr Richard Sadler and colleagues, and published in the Journal of Child Neurology in 2000.The “index case” at the beginning of the paper describes Andy Bolte, one of the boys featured in the program. Andy appeared to be developing normally but regressed severely at 18 months. Andy’s mother Ellen suspected that the regression was caused by antibiotic treatment for an ear infection. This, she hypothesized, had wiped out many of the bacteria in his gut, allowing harmful Clostridia bacteria to thrive. Andy was given Vancomycin, a more powerful antibiotic that can target Clostridia, and his symptoms improved temporarily. But when he finished the course of Vancomycin, he regressed again.The Sadler study set out to determine whether other autistic children would also respond to Vancomycin. The researchers deliberately recruited kids with a similar developmental history to Andy (see Table 1 of the paper). This makes sense but already means that we’re talking about a subgroup of autistic kids - not autistic kids in general. The children all had regressive autism and in each case the onset of autism symptoms followed treatment with antibiotics and diarrhoea. While it’s tempting to assume that the antibiotics must have caused the regressions, it’s also important to remember that many kids around that age will be given antibiotics, and that diarrhoea is a common side-effect of antibiotics. Professor Sydney Finegold was a co-author of the study and one of the central characters in The Autism Enigma. As he stated in the program, the results indicated that “80% of the children improved”. However, it’s not quite that straightforward. The 80% figure comes from a video analysis in which 10 of the 11 children were taped before and during treatment. The tapes were given to a child psychologist, who was asked “Does the child appear better overall in one tape over the other?” 8 of the 10 children were rated as showing better behaviour during treatment than before.To reduce any bias, the psychologist rating the tapes was not told which ones were made before or during treatment. However, bias might have crept in elsewhere, for example in choosing when to tape the kids’ behaviour. It’s also worth bearing in mind that autistic kids are often pretty anxious in a new environment around unfamiliar people so, all else being equal, you’d expect them to show better behaviour on the second visit. Finally, the question the psychologists were asked doesn’t tell us in what sense the kids were better. It certainly doesn’t allow us to conclude that they were less autistic.So the study results were promising, but far from compelling or conclusive. No study is perfect, but these are the kinds of concerns that you’d expect to be addressed in a follow-up study. As far as I can tell, there has been no such follow up – Finegold’s review certainly doesn’t mention one. The program hinted at ethical concerns about the use of Vancomycin (particularly as the benefits weren’t permanent), but in the absence of replication and with the limitations of the original study, we have to be very cautious.Number TwosIn 2002, Finegold and colleagues published an analysis of stool (poo) samples from 11 autistic children (presumably these are the same children tested in the Sadler et al study but that’s not entirely clear). Their main finding was an increased level of clostridia in the samples, compared to non-autistic children. According to The Autism Enigma, “several researchers have replicated [this] finding”. However, I could only find one independent replication - a 2006 study by Parracho et al. [PDF].Other studies haven’t replicated the findings. In a more recent paper, Finegold et al 2010 reported that clostridia actually accounted for significantly less of the gut bacteria in severely autistic kids, compared with non-autistic control children. Instead, they found an increase in Desulfovibrio bacteria. Finegold now argues that this is a more fruitful line of enquiry. In fact, there was a point in The Autism Enigma when Finegold, discussing the impact of antibiotics said “…the organisms that tend to persist are Clostriddia…” and then the audio was cut mid-sentence. My guess is that he went on to mention other bacteria but the program-makers didn’t want to complicate their simple narrative.I’m not in a position to comment on the technical aspects of the studies. However, even to a non-expert, an obvious limitation of these studies is that they have only 8 or 10 control children. This makes it very difficult to be sure what is “normal”, particularly given the emphasis on bacteria being found in autistic children but never in controls. One recent Australian study [PDF] took a different approach, comparing 28 autistic children to a much larger sample collected as part of other research studies. They found relatively little evidence of bacterial abnormalities - and only 1 out of the 28 children with autism had clostridium counts that were outside the normal range. Similarly, several studies have pointed to “abnormalities” of gut bacteria in the stool samples of non-autistic siblings of autistic children. This could be interpreted in many different ways, but clearly complicates any story linking bacteria to autism.Finally, we shouldn’t confuse correlation with causation. Even if there really are abnormalities of gut bacteria in kids with autism, they could simply be a consequence of their often altered diet. As The Autism Enigma points out, many autistic kids will only eat a restricted set of foods - and many will also eat things they’re not supposed to (Andy Bolte ate ashes and paint). Some parents administer probiotics or will enforce gluten or casein free diets. In their 2002 paper, Finegold et al. acknowledge that “we are not aware of any studies that have indicated whether such a diet influences the makeup of the bowel flora” (admittedly there may be recent studies that do indicate). Similarly, Parracho et al. (2006) noted “an association between high clostridial counts and individuals consuming probiotics”. It’s difficult to know what’s cause and effect, but The Autism Enigma only considered one possibility - the bacteria came first.Rats on acid Setting those concerns aside, the theory put forward in The Autism Enigma was that autism is caused by propionic acid, a common food preservative that also happens to be a bi-product of clostridia. According to the narrator, “Once in the brain it changes brain cells to become like those of autistic people”.The program featured a study by Dr Derrick MacFab... Read more »
This week, a guest post from the wonderful Ellie Wilson. Ellie's now at the Institute of Psychiatry in London, working with Declan Murphy. But she has the honourable distinction of being my first ever PhD student. The post is about the two studies that book-ended her thesis on face recognition in autism. Take it away, Ellie...Faces are essentially very similar: two eyes above a nose and a mouth. Yet most people are really good at noticing subtle differences between faces, and interpreting accurately. This helps enormously with social interaction: we can tell who they are, if we know them, we can also tell if they are male or female, roughly what age they are, and what that person might be feeling.In autism, deficits processing facial expressions are widely acknowledged, but there is an increasing amount of evidence for impaired facial identity recognition from scientific studies as well as personal anecdotes.Several years ago I worked as an ABA therapist for a little girl, Clare (not her real name). She was profoundly autistic and her quirky ways and bounding energy made her popular with her classmates. Despite her popularity Clare was always getting the names of the other children mixed up. She was unconcerned by her mistakes and paid little attention to repeated corrections. But we were a little worried, figuring that after a while the other kids might be offended that she still couldn’t identify them. So, in an attempt to protect her social reputation, her mother took a photograph of each child and we Clare and I played various ‘who’s this?’ games. She did get better at naming the photographs. But I’m not sure she ever actually got better at naming the kids in real life.A year later I found myself reviewing literature on face recognition in autism for my PhD. ‘Face recognition’ AND ‘autism’ returns over 600 hits on Web of Science. The experimental methods used in these studies varied hugely. There were memory tests, matching tests, spot-the-difference tasks. Faces were turned upside down, blurred, or shown with all the hair removed. Features were moved around or switched or presented without the face they came from. .. and the results? Some studies reported impairments, some did not.I was confused.Image matchingThen it occurred to me that perhaps some tasks allowed participants to do well even though they were not actually very good at facial identity recognition. At this point I got some helpful advice (and pictures of faces) from Prof Mike Burton, who at the time was with the Glasgow Face Recognition Group. The group had done some really interesting work showing that whilst people can easily match the same image of a person, they have a lot more trouble matching different images of the same person when that person is unfamiliar. So – we thought – when autistic participants do well on the face recognition tasks, are they just ‘image matching’?Examples of two trials. On the left, one of the images on the bottom is the same as the top person. On the right, the bottom left image is the same person but different photoMy first experiment tested this hypothesis. We gave kids two different versions of a simple face-matching task. They’d see one face on a computer screen. Then they’d see two more faces and have to decide which of these was the same person as the first face. In one condition, the correct face would be exactly the same image as the first face. In the other condition, the correct face would be a different image of the same person. Obviously, the second condition is more difficult, but we predicted that the effect of changing the image would be even larger than normal for autistic children because it would prevent them using an ‘image matching’ strategy.Performance on the test was much better foridentical images than different images, butthe effect was identical for both groups.Armed with laptops, touch screen monitors, IQ tests, and a generous supply of highly motivational stickers, I visited schools all over Sydney, testing children who varied widely in age (4 – 16years), IQ (60 - 130) and enthusiasm (totally uninterested – wildly keen). The results from over 70 children revealed that the effect of changing the image was almost exactly the same for children with autism as it was for typically developing control kids.Hypothesis rejected.Each symbol here represents a single child.Scores above -1.64 are considered to be age-appropriate. Roughly half of the autistic kidswere in this "normal" range.Individual variationBut then we looked again at the results for individual children. Unsurprisingly, older children were better at the task than younger children, so we calculated age-standardized scores to show how well each child performed relative to their age. Results showed that ability level within the autism group varied enormously. Some kids were impaired, some were not and this was not accounted for by differences in intelligence.The question for the rest of my PhD became: why are some autistic children bad at face recognition? Eye-gazeProbably the most interesting results came from the last study I did where I used an eyetracker to monitor where kids were looking as they did a face recognition test. One idea I tested was that performance on face recognition tasks would be associated with the amount of time participants spent gazing at the eye-region of the faces they were trying to learn. Avoiding eye-contact is a common symptom of ASD and several studies have shown that at least some autistic kids avoid looking at the eyes of people even in movies. This could be detrimental to face recognition because the eye region is thought to be particularly useful for recognising identity [PDF].However, in our study, we found no correlation between gaze-time on the eyes and performance on the test. Another hypothesis rejected.Dynamic scanningSo then we looked at distribution of attention across the face. Although most adults focus more on the eyes than other facial features, attention is distributed between core features of the face (eyes, nose, mouth). This is thought to build up a unified percept of the face containing information about individual features as well as their spatial relationships. A failure to distribute attention could impede successful recognition.... Read more »
Wilson CE, Palermo R, & Brock J. (2012) Visual scan paths and recognition of facial identity in autism spectrum disorder and typical development. PloS one, 7(5). PMID: 22666378
Alan Turing (Kings College, Cambridge)It’s no exaggeration to say that Alan Turing was one of the most influential figures of the 20th century. Regarded as the father of computer science and artificial intelligence, he also made ground-breaking contributions to the fields of mathematics, chemistry, and biology. Most famously, during World War II, he played a crucial role in cracking the Nazi's Enigma code.He was also, it's argued, a person with Asperger syndrome.There's something of a cottage industry in "outing" historical figures with autism or Asperger syndrome. Candidates include Mozart, Einstein, Isaac Newton, Ludwig Wittgenstein, Thomas Jefferson, Andy Warhol. The list goes on. In many cases, it seems, being brilliant at something and having a reputation for social awkwardness is all that it takes for a "diagnosis".In Turing's case, there is at least some more concrete evidence to go on. In a 2003 paper, Henry O’Connell and Michael Fitzgerald trawled through Turing's biography, looking for anecdotes and descriptions of Turing that would support a diagnosis of Asperger syndrome.The authors used the Gillberg criteria for Asperger syndrome - a set of six "symptoms" that must all be present for a diagnosis to be conferred. Turing, they concluded, met all six criteria:Severe impairment in reciprocal social interactionSchool report described him as "antisocial"Only one friend at schoolUnable to control younger boys at school or manage co-workersNo attempt to socialise with academic superiorsAll-absorbing narrow interestInterests in science, mathematics, chemistry, codes and ciphers, natureImpositions of routines and interests (on self or other)Always ate an apple before bedHouse was cluttered with whatever he was interested in at the timeAlways put the cork back in the wine bottle at the end of a mealOften worked through the nightWrote about his work to people with no scientific backgroundNonverbal communication problemsStiff gaze in photographsLack of eye contactAwkward appearanceCharacteristic response to presentation of new ideas (stabbed fingers and said "I see, I see")Speech and language problemsHigh pitched voiceMisunderstood enrolment form for Home GuardOver-analysed colleagues' approachesMotor clumsinessPoor handwritingAlways got ink on his collar at schoolCertainly, a case can be made for Turing meeting each of the six criteria. But some of the observations, such as a high pitched voice, or working late at night don't really constitute evidence. And can we really say that he had narrow interests when he influenced so many distinct fields? Was having only one friend at school a reflection of social impairment or of having few peers who shared his interests?The difficulty with making historical diagnoses is that there's no opportunity to ask further, more targeted questions. What happened if Turing didn't get his nightly apple? Did it bother him, or did he eat some other fruit?A proper diagnostic interview might uncover further evidence that would provide a more compelling and watertight case for diagnosis. Even so, Turing's case highlights the subjective nature of diagnosis. This is particularly true around the edges of the autism spectrum where, as Lorna Wing put it, autism "shades into eccentric normality".Attempts to diagnose Turing arguably reveal more about our current fuzzy concepts of autism than they do about Turing the man. And they make plain why we're still a long way from understanding the enigma that is autism.ReferenceO'Connell, H., & Fitzgerald, M. (2003). Did Alan Turing have Asperger's syndrome? Irish Journal of Psychological Medicine, 20 (1), 28-31Further reading:23rd June was the centenary of Turing's birth. Below, a selection of the many articles commemorating his life.The rich legacy of Alan TuringThe spirit of Alan TuringThe highly productive habits of Alan TuringAlan Turing: Society failed the genius, we must learn from his lossAlan Turing's cryptographic legacyAlan Turing and the bullying of Britain's geeksGary Kasparov versus Alan Turing's 1950 chess program... Read more »
O'Connell, H., & Fitzgerald, M. (2003) Did Alan Turing have Asperger's syndrome?. Irish Journal of Psychological Medicine, 20(1), 28-31. info:/
Sometimes, it's good to get away. In February, we spent two weeks cruising around the North Island of New Zealand in a campervan, quickly christened Campo by my four-year-old. We saw the giant Kauri trees of Waipoua and the giant sand dunes on 90 mile beach; we went sailing on the Bay of Islands and bathing in the volcanic springs of Hotwater Beach. And it only rained twice. I learnt a little of the art of campervan maintenance . And, while I was under strict instructions not to do any work, the winding roads of Northland and Coromandel gave plenty of opportunity for idle philosophising. For holiday reading, I picked up a copy of Richard Dawkins' The Greatest Show on Earth, in a lovely little second hand bookshop in Devonport. Dawkins, it must be said, is on top form, laying out the evidence for evolution and laying into creationists and flat-earthers at every turn. The highlight for me was his joyful description of Lenski’s experiments in bacterial evolution. Even for someone far removed from this line of research, it’s inspirational stuff, demonstrating the elegance and power of science done well. The dead hand of Plato In Chapter 2, Dawkins ventures on an interesting tangent, asking why it took us all so long to figure out evolution. After considering some of the more obvious explanations (religious objections, the unimaginable timespan of evolution), he ultimately concludes by laying the blame at the feet of the ancient Greek philosopher, Plato . Plato’s idea was that all classes of things have an essence – a set of defining properties. Members of that category may vary in other respects, but they all share that essential nature. Chairs, for example, can vary in size, shape, colour, comfort, and so on, but they are still all essentially chairs; they all have the essence of chairness; they are variations on an ideal chair.As the evolutionary biologist Ernst Mayr pointed out, this essentialist way of thinking about things becomes problematic when trying to understand evolution. Plato would consider any natural variation amongst rabbits as "flawed deviation from the ideal essence of rabbit". In Mayr's view, Darwin succeeded by breaking away from this Platonic mindset and realising that it is this variation, coupled with non-random selection, that is the driving force behind evolution. For Darwin, there was no essential quality of rabbitness; no ideal rabbit; and, crucially, no guiding hand directing historical proto-rabbits to become more rabbit-like. Giant Sandunes, NorthlandDawkins concurs with Mayr, and spends the rest of The Greatest Show on Earth piling on the evidence for evolution from every conceivable angle. The relevant point here, however, is that natural selection is counterintuitive. Our default mode of thinking in terms of essential, idealised qualities of different species proved an obstacle to scientific progress. This probably explains why Darwin was so late on the scientific scene; why he wasn’t beaten to the punch by a scientist centuries earlier. It probably also goes some way to explaining why so many people today still deny the possibility of evolution, in spite of the overwhelming evidence. It’s an illustration that our intuitions are often wrong or misleading and that the whole point of science is that it can and frequently does defy those intuitions .The essence of autismI was trying hard not to think about work. But as I read about Plato, Mayr, and the essence of bunny rabbits, it struck me that there might be important lessons for autism research. It's not that there are any obvious analogies that I can think of. But, like Darwin's contemporaries, I wonder whether we, as people interested in autism, may be stuck in a similar essentialist rut. Cape Reinga - where the Tasmanmeets the South PacificThe following, from a recent post on the SFARI blog, expresses a familiar sentiment: "Autism is a complex, heterogeneous disorder. But the core phenotype, which can be recognized to some degree in any individual on the autism spectrum, nonetheless suggests that there must be some common underpinnings.” We acknowledge the heterogeneity within autism, but our intuitions still drive us to seek a common essence of autism. It’s easy to see where this intuition comes from. Essentialism underlies our definitions of autism. Diagnostic criteria are aimed squarely at defining the “core” (essential) characteristics of the disorder. We can even think of an "ideal" autistic person as someone whom Kanner would have identified as autistic – someone with "classic autism". Diagnostic boundaries indicate how much variation away from this ideal can be tolerated before the individual is deemed to be not autistic. Common symptoms that are not part of the diagnostic criteria, such as language delay, intellectual disability, attention deficit, and so on are considered to be non-essential “co-morbidities”, separate and on top of the autism. The essentialist view of autism goes hand in hand with the way autism research is conducted and reported. Most studies involve taking a group of individuals with autism and comparing them to a control group. The assumption is that the group average is what matters. Individual differences within the autism group are considered to be non-essential variation. Sandspit estuaryStudies are then reported as showing, for example, that people with autism are good or bad at a particular test, that their brains are over- or under-activated in response to a particular stimulus, or that they do or do not respond to a particular intervention. This kind of generalization, from the handful of individuals taking part in the study to "people with autism", is only licensed if people with an autism diagnosis are interchangeable - if they are essentially the same. Yet one only has to meet a few to realise that this is not the case.Universals and specifics Perhaps most clearly and explicitly, the Platonic mindset is revealed by the widespread view that theories of autism must be evaluated according to their universality and specificity; the theory should apply to everyone with an autism diagnosis but nobody without autism and failure on either of these criteria is grounds for rejecting a theor... Read more »
Brock J. (2011) Commentary: Complementary approaches to the developmental cognitive neuroscience of autism--reflections on Pelphrey et al. (2011). Journal of child psychology and psychiatry, and allied disciplines, 52(6), 645-6. PMID: 21574994
Pelphrey, K., Shultz, S., Hudac, C., & Vander Wyk, B. (2011) Research Review: Constraining heterogeneity: the social brain and its development in autism spectrum disorder. Journal of Child Psychology and Psychiatry, 52(6), 631-644. DOI: 10.1111/j.1469-7610.2010.02349.x
Earlier this week, I wrote about the recently published study by McPartland, Reichow and Volkmar, looking at the potential impact of the proposed DSM 5 criteria for Autism Spectrum Disorder. There were caveats-a-plenty but the data suggested that autism rates may go down considerably as a result of the renegotiation of diagnostic borders. In fact, a new diagnosis, Social Communication Disorder, has been created to capture all the people who won't now make the ASD cut-off.In the comments, an important question was raised by Harold Doherty, author of the Autism in New Brunswick blog: What will happen to autistic people with intellectual disability in DSM 5?Although I didn’t mention it in the original post, this was actually something that McPartland et al looked at in their study.Harold's concern, which he's been raising for some time, is that the wording of the new criteria will actually (and, in his view, deliberately) exclude people with an intellectual disability.Under the draft version of DSM 5, in order to receive an ASD diagnosis, a person would have to show:"persistent deficits in social communication and social interaction across contexts, not accounted for by general developmental delays." It's that last part that is the issue for Harold.The idea, as I interpret it, is to make sure that people getting an ASD diagnosis really are autistic and not merely people who have social difficulties that are a result of developmental delay.An analogy would be with people who are born deaf or blind; they inevitably have social interaction difficulties but we wouldn't necessarily call them all autistic. In the same way, someone with intellectual disability might struggle in certain social situations because of their poor language and memory skills, but again they wouldn't necessarily be considered to have autism.Autism is more thanjust being a little behind in terms of social development. It’s qualitativelydifferent. The challenge is to define those qualities.This interpretation of the new criteria is entirely consistent with the quotation Harold provides from Cathy Lord, a member of the DSM-5 Neurodevelopmental Work Group:Catherine Lord… said that the goal was to ensure that autism was not used as a “fallback diagnosis” for children whose primary trait might be, for instance, an intellectual disability or aggression. New York Times, 20 Jan, 2012 Clearly, the intent is to stop non-autistic people getting an ASD diagnosis. That isn’t the same thing as saying that, if you’re intellectually disabled, you can’t also be autistic.So much for the intentions. What about the guidelines in practice?In one of their analyses, McPartland et al. divided people up according to their IQ. Those with IQs below 70 were considered to have “low cognitive ability”, which I think for our purposes we can treat as being synonymous with intellectual disability.As in the previous post, I’ve plotted the results with kids achieving a DSM-5 ASD diagnosis in blue (ASD+) and those missing out on a diagnosis in red (ASD-).As we saw last time, a good chunk of the kids failed to meet DSM-5 criteria for ASD. However, kids with intellectual disability were much more likely to be diagnosed with ASD than were those with IQs above 70. This is the opposite of what we’d expect if DSM-5 really was going to selectively exclude the intellectually disabled.As I discussed before, we need to treat these findings with a good deal of caution. But I can't see any intention to exclude people with intellectual disability, and there's certainly no evidence at this time that this will happen in practice.Reference:McPartland, J., Reichow, B., & Volkmar, F. (2012). Sensitivity and Specificity of Proposed DSM-5 Diagnostic Criteria for Autism Spectrum Disorder Journal of the American Academy of Child & Adolescent Psychiatry, 51 (4), 368-383 DOI: 10.1016/j.jaac.2012.01.007... Read more »
McPartland, J., Reichow, B., & Volkmar, F. (2012) Sensitivity and Specificity of Proposed DSM-5 Diagnostic Criteria for Autism Spectrum Disorder. Journal of the American Academy of Child , 51(4), 368-383. DOI: 10.1016/j.jaac.2012.01.007
In January, at a meeting of the Icelandic Medical Association, Yale researcher, Dr Fred Volkmar gave a presentation of data from a study looking at the implications of changes to autism diagnostic criteria in DSM 5. His conclusion was that many people who are currently diagnosed with autism, Asperger's, or PDD-NOS would not meet the new proposed criteria for autism spectrum disorder in DSM 5.Volkmar's remarks were picked up by the New York Times, who ran with the lede:"Proposed changes in the definition of autism would sharply reduce the skyrocketing rate at which the disorder is diagnosed and might make it harder for many people who would no longer meet the criteria to get health, educational and social services"Not surprisingly, the article caused much consternation in autism circles. But because the study itself hadn't been published, members of the DSM 5 Neurodevelopmental Work Group, charged with implementing these changes, were unable to pass comment.After all the brouhaha, the Yale study, with Dr James McPartland as first author, was finally published online two weeks ago - to suprisingly little fanfare. It was also accompanied by a commentary from Susan Swedo and members of the DSM 5 group, who had finally got hold of the paper.DSM 5 and autismFirst, a quick recap. There are essentially two big changes on the cards in DSM 5 as far as autism is concerned. First, rather than there being separate diagnoses of autistic disorder, Asperger's syndrome, and PDD-NOS, there will be a single category of "Autistic Spectrum Disorder".Second, rather than organising the criteria into three themes, DSM 5 will collapse that down to two.But there's also a more subtle and potentially more important change. DSM IV took a kind of pick'n'mix approach, with over 2000 combinations of 12 "symptoms" that would get you a diagnosis.In DSM 5, this is simplified. The "symptoms" themselves are each somewhat broader and perhaps more flexibly interpreted, but there is much less scope for picking and mixing.McPartland et al.'s studyWhether or not you agree with the changes in principle, what really matters is how the new criteria translate into the real world. There are currently ongoing "field studies" that are testing out the new criteria to see how they work in practice. McPartland and Volkmar essentially took a short cut by reanalysing the data from a similar field trial conducted 20 years ago when DSM IV was being developed.To do this, McPartland et al had to translate the questions that were asked in the original study into something approximating an item in DSM 5. They then reanalysed the data, applying DSM 5 rules to "re-diagnose" the kids in the original study.This is how things worked out. Kids in blue (ASD+) were those who met criteria for ASD in DSM 5. Those in red (ASD-) failed to meet the criteria.As you can see, roughly three quarters of kids originally diagnosed with autistic disorder met DSM 5 criteria for ASD. However, only a quarter of those with a diagnosis of Asperger syndrome or PDD-NOS made the grade, so to speak.These figures seem to back up Volkmar's claim that many people, currently diagnosed, would not meet the new criteria. However, with publication, the limitations of the study are now clear, as noted in the commentary by Swedo and colleagues, and by the authors themselves.LimitationsThe most obvious problem is that the original study was not designed for these purposes. While some of the DSM 5 criteria mapped neatly onto items in the original study, others didn't. For example, DSM 5 includes this pretty vague statement:Symptoms must be present in early childhood (but may not become fully manifest until social demands exceed limited capacities).Translated into English, this I think means that parents should be able to look back with hindsight and recognise early signs of ASD, even if they missed them at the time. Quite how this will be ascertained without asking leading questions is not clear. But the important point is that there was nothing in the DSM IV questions that really corresponded to this question, so McPartland had to go with the nearest thing, which was "onset before 36 months".As Swedo et al caution, it's probably best to wait until the results of the actual field studies are in. However, given that the DSM 5 criteria are much more prescriptive than those for DSM IV, we have to expect results along these lines. Further comment comes from David Skuse. Writing in his editorial capacity, he highlights the pressing issue of the people who don't meet the DSM 5 criteria. As mentioned in a previous post, DSM 5 is set to introduce a shiny new diagnostic category of "Social Communication Disorder" which would capture many of these people. The clinical and scientific merits of this new diagnosis are, however, yet to be established.For me, this all highlights the inherent circularity and subjectivity of the DSM 5 debate. McPartland et al. effectively use DSM IV diagnoses as the standard against which to judge DSM 5. In the original trial 20 years ago, these would have been judged against DSM III. Inevitably the outcomes are slightly different, but who is to say which is better or worse? Swedo et al. make precisely this point, arguing that the true "gold standard" is expert clinical opinion. Or, to put it another way, we know it when we see it. An admission, perhaps, that autism defies definition.Reference:McPartland, J., Reichow, B., & Volkmar, F. (2012). Sensitivity and Specificity of Proposed DSM-5 Diagnostic Criteria for Autism Spectrum Disorder Journal of the American Academy of Child & Adolescent Psychiatry, 51 (4), 368-383 DOI: 10.1016/j.jaac.2012.01.007Related posts:Exactly how many ways are there to get an autism diagnosis?What is PDD-NOS?Social Communication Disorder - A New Category in DSM 5... Read more »
McPartland, J., Reichow, B., & Volkmar, F. (2012) Sensitivity and Specificity of Proposed DSM-5 Diagnostic Criteria for Autism Spectrum Disorder. Journal of the American Academy of Child , 51(4), 368-383. DOI: 10.1016/j.jaac.2012.01.007
Having spent much of the past week struggling to make sense of my data, it’s good to come home, pour a glass of wine, put on some Sharon Jones, and, er… play with somebody else’s data!Recently, I’ve discovered DataThief - an application that allows you to scan in a graph from a paper and extract the data points. Sometimes, this provides insights that really aren’t obvious from the original paper.The other week, for example, I came across an intriguing neuroimaging study reported on the SFARI website. In the study, Judith Verhoeven and colleagues used diffusion tensor MRI to examine the superior longitudinal fasciculus, a bundle of nerve fibres that is assumed (although see this paper) to connect two brain regions involved in language production and comprehension - Broca’s area (left front-ish) with Wernicke’s area (left and back a bit).Verhoeven et al. reported that integrity of the superior longitudinal fasciculus was compromised in kids with specific language impairment (SLI) – that is, kids who have language difficulties for no obvious reason. However, the same was not true of kids with autism, even though they had poorer language skills than those with SLI.Taken at face value, this is a pretty major blow to the idea that autism and SLI have anything more than a superficial resemblance [pdf].DataThief, however, suggests otherwise.The figure below is a scatterplot with each coloured shape representing a single child. On the x-axis is performance on a language test. On the y-axis is fractional anisotropy (FA) – the imaging measure used to assess the integrity of the left superior longitudinal fasciculus.Figure 3a from Verhoeven et al 2011, showing integrity of the left superior longitudinal fasciculus plotted against the child's language scores (z-scores). Children with SLI in red, autistic kids are the blue squares. Control children are the green and blue circles. The purpose of the graph was to show the significant correlation between these two measures in the SLI group. But if we can read off the y-coordinates of each shape, we can show the distribution of fractional anisotropy scores for all three groups.Cue DataThief.It’s really just a case of clicking on three reference points for which you know the coordinates and then clicking on each of the data points in turn. Then you simply export the coordinates of the data points as a text file. The only thing I had to remember was to do the three groups separately so I knew which point belonged in which group.Here’s the fractional anisotropy data replotted to show the distribution for each group. What we can now see is that there is a small subgroup of control kids who have really high FAs. There is also one autistic kid and one kid (arguably two) with SLI who have low FAs. Everyone else is pretty much in the middle.Verhoeven et al.'s data replotted to show the distribution of fractional anisotropy for each groupOn average, kids with SLI have lower than ‘normal’ fractional anisotropy , but looking at the spread of scores, you’d be hard pressed to conclude that this was a characteristic of SLI. Likewise, the overlap between the distributions of the autism and SLI groups is almost complete. Hardly evidence for fundamentally different neural mechanisms in the two disorders. At the risk of sounding like a broken record, this once again highlights the importance of looking at individual variation within diagnostic groups such as autism and SLI, rather than (or as well as) looking at group averages.But it also emphasizes a more general point (and this I have to stress is no criticism of the authors of this particular paper).The data reported in a journal article are really just a snapshot of the actual data recorded, filtered through the authors’ preconceptions about what questions are interesting to ask and how to go about doing that. There’s an imperative to present the data in a neat, sanitized package, with all the rough edges and anomalies smoothed out; to tell a coherent story that will convince reviewers and editors that it’s worthy of publication in a reputable journal. Years of work and terabytes of data may be compressed into just two or three pages.DataThief only takes us so far. It allows us to extract the information presented visually in the published article, but no further.Most of the past week has been spent convincing myself that it doesn’t really matter how I analyse my data because the results come out the same regardless. This is reassuring for me, but it doesn’t mean that somebody else, looking at my data with fresh eyes and a different perspective, would not come to an entirely different set of conclusions.In an ideal world, when a paper is published, researchers should also be able (and encouraged) to publish the data on which the paper is based, as well as the script showing exactly how those data were analysed.There are, of course, many obstacles in the way and questions to be answered before this becomes standard practice. Who would host and maintain the data? Just how raw should the raw data be? What if the authors are writing multiple papers based on the same data set? Who gets credit for reanalyses of the data set? What happens if a reanalysis shows up an error in the original paper? If the research involves human participants, how do we reassure them that their anonymity will be maintained?Undoubtedly, there are many more problems that I haven't thought of. But, as scientists, we need to work through these issues and find ways to set our data free.Footnotes: The analyses involved an ANOVA with left and right hemisphere as a within-subjects factor. This showed a main effect of group, but no group by hemisphere interaction.Reference:Verhoeven, J., Rommel, N., Prodi, E., Leemans, A., Zink, I., Vandew... Read more »
Verhoeven, J., Rommel, N., Prodi, E., Leemans, A., Zink, I., Vandewalle, E., Noens, I., Wagemans, J., Steyaert, J., Boets, B.... (2011) Is There a Common Neuroanatomical Substrate of Language Deficit between Autism Spectrum Disorder and Specific Language Impairment?. Cerebral Cortex. DOI: 10.1093/cercor/bhr292
Like most things in autism research, the idea that people with autism have big brains goes back to an observation in Leo Kanner’s original autism paper, where he noted that some of the kids in his group had larger than normal heads. Over the years, there have been dozens of studies looking directly or indirectly at the issue of brain size in autism. In 2005, Martha Herbert provided a comprehensive review [pdf] of 25 such studies, describing the tendency towards large brains as "the most replicated finding in autism neuroanatomy".Redcay & Courchesne 2005Also in 2005, Elizabeth Redcay and Eric Courchesne published a meta-analysis, in which they ingeniously plotted how much bigger or smaller than average the autism brains in different studies were as a function of the mean age of the participants in the study. They concluded that there is an early period of brain ‘overgrowth’, with autistic brains being on average 10% larger than normal. But then growth slows down and typically developing kids eventually catch up.Redcay & Courchesne's meta analysis. Red circles (my annotation) indicate studies in which brain volume was inferred from head circumference measures.Redcay and Courchesne's paper has been extremely influential. But their analysis rests on a number of assumptions that are worth highlighting.First, the data are cross-sectional. Different people are being measured at each of the different ages. This is inevitable because brain imagining technology hasn't been around long enough for a proper longitudinal study be conducted, following individuals over the first decades of their life. But if we think of the curve as being the actual growth trajectory of a person with autism (as Redcay and Courchesne want us to) then we are essentially assuming that a 30-year-old autistic adult in one study is what a 5-year-old in another study will be like a quarter century from now. This is a pretty big assumption.Second, data for the youngest age groups actually came from measurements of head circumference taken during regular infant check-ups, rather than actual brain scans. Head circumference is correlated with brain size in infants, and realistically it's the only way to study brain size in autism pre-diagnosis. But in order to plot the data on the same graph, Redcay and Courchesne had to make quite a few assumptions about the relationship between head size and brain volume .Courchesne et al 2011More recently, Courchesne and colleagues published an update, pulling together data from all of their MRI studies. The data were still largely cross-sectional but, this time, they fitted growth curves to the data from autistic and non-autistic individuals.This gives a clearer sense of the variation within each group, which, even for typically developing children (blue circles), is huge. Having a brain that is 10% bigger than average (as Redcay and Courchesne's analysis suggests) isn't actually all that abnormal. Nonetheless, Courchesne et al.'s curve-fitting led them to again conclude that autism is associated with “early brain overgrowth”.Individual brain volume data as a function of age (Courchesne et al., 2011). Red squares are boys with autism. Blue circles are typically developing boys.Comparing the two curves suggests that the difference between autistic and non-autistic brains is largest in the period between around 20 and 32 months (which I've conveniently highlighted). However, if we look at the actual data in this period, ignoring the curves, then we see that none of the autistic kids had brains that were unusually large for their age.The curve for the typically developing boys is strongly curved because it has to fit the data from the youngest kids (12 to 18 months). The autism data don’t start until around 20 months, so the curve is inevitably less curvy. This gives the impression of a big difference in brain size at around 2 years of age, which I don't think really exists in the data.... Read more »
Nordahl, C., Lange, N., Li, D., Barnett, L., Lee, A., Buonocore, M., Simon, T., Rogers, S., Ozonoff, S., & Amaral, D. (2011) Brain enlargement is associated with regression in preschool-age boys with autism spectrum disorders. Proceedings of the National Academy of Sciences, 108(50), 20195-20200. DOI: 10.1073/pnas.1107560108
Redcay, E., & Courchesne, E. (2005) When Is the Brain Enlarged in Autism? A Meta-Analysis of All Brain Size Reports. Biological Psychiatry, 58(1), 1-9. DOI: 10.1016/j.biopsych.2005.03.026
Anger he smiles, towering in shiny metallic purple armour.Queen Jealousy, envy waits behind him, her fiery green gown sneers at the grassy ground.Blue are the life giving waters taken for granted, they quietly understand.Once happy, turquoise armies lay opposite ready, but wonder why the fight is on.My red is so confident that he flashes trophies of war and ribbons of euphoria.Orange is young, full of daring but very unsteady for the first go round.My yellow in this case is not so mellow, in fact I'm trying to say it's frightened like me.And all of these emotions of mine keeps holdin’ me from givin’ my life, to a rainbow like you.Hendrix fans amongst you will recognise the above as lyrics from Bold as Love, the title track of Jimi's second album. According to the sleevenotes (from a long-lost cassette, so you'll have to trust my memory on this), the concept for the song was the idea of using emotions to describe colours to a blind person. The more obvious interpretation is that Hendrix was using colours as a metaphorical device to describe his own conflicting emotions.Earlier this week, I was reminded of Bold as Love when I came across an intriguing study reported in the journal, Neurocase . The authors, VS Ramachandran and colleagues, described the case of TK, a young man with Asperger syndrome who was encouraged to use colours to help him understand emotions:"Around the age of 10 his mother suggested that he attempt to label the feeling of each emotion (presumably based on context, social situation, and facial expressions) with a specific color, in an attempt to relay the appropriate emotions to his father and her. For example, while experiencing what he considered happiness he would tell his parents that he was feeling ‘green’." "Further, by comparing the color elicited by another person with the emotion that would be associated with the same color in his own mind, TK was able to ‘read’ the other individuals’ emotions more accurately.""At about the same time that he began associating colors with emotions, he also began seeing colored halos around individuals. The color of these halos corresponds to TK’s emotional stance toward that particular person, and when a new individual is encountered a blue halo emerges de-nouveau and the color evolves progressively with repeated exposure."Purple hazeTo objectively measure the halo perception, Ramachandran et al asked TK to identify letters that were projected onto a white screen. An unfamiliar person, identified as having a blue halo, stood in front of the screen.If the letters were blue and projected close to the person (i.e., within the halo) then TK was unable to identify the letters above chance levels - presumably because the letters and background appeared to be in the same colour. When the colour was changed or the letters moved to outside the halo, TK's performance was flawless.It's a shame that the authors weren't able to test TK with a second person standing in front of the screen, whom he perceived as having a different coloured halo. Nonetheless, these results appear to provide some objective confirmation of TK's unusual subjective reports .Love or confusionIn a second experiment, Ramachandran et al tested TK and 15 control subjects on a Stroop interference test.Participants were given words printed in colour and had to say the colour of each word, ignoring what the word itself said. In the classic version of the test, the words are all themselves colour names.In the congruent condition, the word matches the colour in which it's printed: RED BLUE YELLOW GREENIn the incongruent condition, the word and its colour are mismatched RED BLUE YELLOW GREENPeople are generally faster to name the ink colours when the word matches the colour. Even though they're supposed to be ignoring what the word says, they can't help but read it, and this affects their response to the actual colour. As you can see from the graph below, TK was no exception. Like the control group, his reaction times were longer for the incongruent condition than for the congruent condition.TK showed a similar effect when the words were emotions, being quicker to name the colour if the emotion word matched the colour he associated with that emotion (e.g., PRIDE and AGGRESSION) than when they were incongruent. When the control participants were given the same stimuli, they showed no such effect .Ramachandran et al. interpret these findings as evidence of "emotion-colour synaesthesia" - the implication being that TK actually perceives emotions as colours. This is certainly one possibility, but it's worth noting that similar effects are commonly observed in typical adults using non-emotion words that have associations with colours. For example, it's easier and quicker to name the colours in FIRE GRASS LEMON SKY than it is in FIRE GRASS LEMON SKY.Seeing the word SKY makes us think of the colour blue, which then affects our ability to name colours; but there's no suggestion that we actually perceive the colour blue every time we read SKY. By the same token, TK associates PRIDE with blue and this affects his colour naming, but the data from the Stroop task don't show that he experiences pride as the colour blue.Crossbrain trafficIt's undoubtedly a fascinating case study. Ramachandran et al.'s data indicate that TK perceives a blue halo around certain people and it's safe to say that he has strong cognitive associations between colours and emotions. However, we are still relying on TK's subjective reports that the halos vary from person to person and that emotions are actually experienced as colours. This is not to cast doubt on TK's reports, merely to note that the objective evidence is not perhaps as strong as the authors claim.Ramachandran et al. speculate that TK's experiences derive from increased connectivity between brain regions involve in vision (V4), face processing (FFA), and emotion (insula and amygdala). It would certainly be interesting to know whether this is supported by brain imaging.I'd also like to know more about TK's As... Read more »
Ramachandran VS, Miller L, Livingstone MS, & Brang D. (2011) Colored halos around faces and emotion-evoked colors: A new form of synesthesia. Neurocase. PMID: 22115465
Evidence and Artifacts: 1 in 110A few years ago, before I got into autism research, I worked on a couple of projects looking at Down syndrome and Williams syndrome. Down syndrome, I assume, is familiar to most readers. Williams syndrome is much rarer and less well known, but is of considerable interest to researchers, not least because the extremely sociable personalities of many people with Williams syndrome provide an interesting (although complex) contrast with autism.What these two syndromes have in common is that both are defined in terms of their genetics. Down syndrome is the result of having an extra copy of chromosome 21. Williams involves a deletion of a small sequence of genes on chromosome 7. But even without genetic testing, people with Down's and Williams are pretty easy to identify because each syndrome has its own characteristic facial structure.Doing research on these two syndromes wasn't easy, but at least we could be confident that the people in our study all had the same condition - and that our participants were coming from the same population as those in other studies by other research groups.Girls with Williams syndrome (left) and Down syndrome (right)The same, unfortunately, can't be said about autism. While it's certainly a genetic disorder, in the sense that it is highly heritable, we're still some way from fully understanding what those genetic mechanisms might be. And so, at least for now, autism is defined in terms of behaviour. Show enough atypical behaviours, tick enough boxes, and you get an autism diagnosis.What this means in practice is that people with autism are an extremely varied bunch. In a 2007 paper [pdf], Daniel Geschwind and Pat Levitt argued that we should really stop talking about autism as if it were a single entity and instead talk about "the autisms" as a collection of distinct disorders.While this idea has certainly gained some traction, it's still the case that most autism research focuses on the differences between autistic and non-autistic people, ignoring the variation within the groups. Often, the implication is that the group average is somehow representative of all people with autism. But if, as Geschwind and Levitt argue, there really are many autisms, the average of the autism group may not actually be representative of anyone.A study out last week in the (oddly titled) journal, Molecular Autism, illustrates this point rather nicely in concrete terms. And it does it by looking at differences in facial characteristics.Facial characteristics of autistic boysThe study was conducted by Kristina Aldridge and colleagues at the University of Missouri. They took multiple photos of 64 boys with autism aged between 8 and 12 years, and then stitched the photographs together to create a 3D realisation of each boy's face. They then identified 17 different points on each face (as shown below) and measured the distance between them. The faces of the autistic boys were then compared with those of 41 typically developing (i.e., non-autistic) boys.The 17 points used by Aldridge et al to measure facial structureThe press release from the University of Missouri summarises the results as follows:Children with autism have a broader upper face, including wider eyes.Children with autism have a shorter middle region of the face, including the cheeks and nose.Children with autism have a broader or wider mouth and philtrum -- the divot below the nose, above the top lip.However, a second analysis lends a very different perspective on these results. Aldridge and colleagues performed a cluster analysis of the 64 autistic boys and identified two subgroups with distinctive facial characteristics.The first subgroup contained 12 boys, who had wider nose and mouth but a shorter distance between the upper and lower parts of the face (see the left face below for an example).The second subgroup (right face below) contained just 5 boys, and was characterised by a wide upper face.The faces of the remaining 47 boys were indistinguishable from those of typically developing boys.Black lines indicate a reduced distance, white lines indicate an increaseThese results rather undermine the claims of a "distinct facial phenotype" in autism. On average, the autistic boys had a shorter middle face region and a broader upper face than controls, but these characteristics, it would appear, were actually owned by different individuals. It's quite possible that no single individual actually possessed the average "autistic face". And most of the boys (almost three quarters of the sample) didn't show any of the "distinct" characteristics.Embryonic ideasAldridge et al.'s data show the problems inherent in relying on the group average, but they also demonstrate how much richer an understanding can be gained by considering individual differences within autism.Citing evidence that the brain and face both develop from the same embryonic structures, Aldridge and colleagues argue that atypical facial structure in autism is a clue to underlying differences in early prenatal development. If we take this at face value (pun intended), then the data actually suggest that there are multiple different ways in which embryonic development can diverge from the typical trajectory, each associated with different facial profiles.We don’t know (as yet) about the brain structure or function of the boys in this study. However, Aldridge et al do report that these two different facial subgroups differed in terms of their clinical and cognitive profiles. Boys in the first subgroup tended to have more severe autism symptoms, lower IQ, and a higher incidence of regression (loss of skills).An obvious next question involves the origin of these facial and behavioural differences. Williams syndrome and Down syndrome provide obvious comparisons and suggest a possible genetic cause. The authors note that boys with Fragile X syndrome, chromosomal disorders (like Down syndrome), or copy number variations (like Williams syndrome) were excluded, but it's not clear (to m... Read more »
Aldridge K, George ID, Cole KK, Austin JR, Takahashi TN, Duan Y, & Miles JH. (2011) Facial phenotypes in subgroups of pre-pubertal boys with autism spectrum disorders are correlated with clinical phenotypes. Molecular autism, 2(1), 15. PMID: 21999758
“You made a circle”, exclaimed Ethan proudly as he looked up from his drawing. “You did make a circle”, his mum acknowledged, ignoring the fact that, not for the first time, Ethan had reversed the pronoun, saying “you” when he should have said “I”. Ethan was one of six children from Providence, Rhode Island taking part in a study of child language development. Every couple of weeks, a researcher from Brown University would visit him and his mum at home, record, and then transcribe their conversations in painstaking detail. The transcriptions would show that Ethan was a prolific reverser of pronouns; frequently saying “you” when he meant “I” and “your” instead of “my” or “mine”. This curious habit began as soon as pronouns entered his vocabulary and he was still reversing pronouns when, just before his third birthday, the study came to an end.
"In solving a problem of this sort, the grand thing is to be able to reason backward."
Ethan’s language skills were otherwise exceptionally good. When assessed at 18 months, his scores put him in the top 1% for children his age. However, some years after the study finished, it transpired that Ethan had Asperger syndrome. Pronoun reversal is common amongst children on the autism spectrum. Leo Kanner noted as much in the first systematic description of autism and, to this day, it is considered an important marker when conferring an autism diagnosis. But the underlying cause of this highly specific problem remains something of a mystery. Ethan’s diagnosis made sense of his pronoun reversal, but it didn’t exactly explain it.While pronoun reversal is relatively common in autism, it certainly isn’t unique to the disorder. Deaf children in particular are prone to reversal, despite the fact that in many sign languages, pronouns simply involve pointing to the person in question. And while most typically developing children appear to have little difficulty with pronouns, there have also been several case reports of children who go through a prolonged phase of pronoun reversal. By coincidence, Naima, one of the five other children in the Providence study, was one such child. Aware of the serendipitous nature of their data, two of the researchers, Karen Evans and Katherine Demuth, returned to their transcriptions. Forensically re-examining the evidence, they tried to work out why the two children had encountered such difficulties with pronouns. The results of their enquiries provide some intriguing insights into the multiple challenges facing both typically and atypically developing linguists.The pronoun problemPersonal pronouns represent an unusual problem for the young language learner. Most words they encounter will have a constant reference, at least within the context of the ongoing conversation. “Mummy” will refer to their own mother. “Dog” will refer to the animal that is sat on the carpet right in front of them. But the meanings of “I” and “you” change, depending on who it is that is speaking. My “you” is your “me”.
It is a capital mistake to theorize before you have all the evidence.
In Naima’s case, it seems that she simply failed to grasp this concept, thinking that “you” was really just another name for herself. It wasn’t that she sometimes got it right and sometimes got it wrong. Between the ages of 19 and 28 months, virtually every time she used “you” or “your”, she was actually referring to herself, sometimes with amusing results:Naima: "I think you peed in your diaper."Mother: "Just now?"Naima: "I think you did."Then, all of a sudden, something clicked. In Naima’s final two sessions at 29 and 30 months, every single pronoun was used correctly. But why did she make this mistake in the first place? And what happened for the penny to drop?Are you experienced?Yuriko Oshima-Takane, a psychologist at McGill University in Montreal, has argued that children can only deduce the principles of pronoun use by listening in on other people’s conversations. Pronoun reversers, she suggests, are children who, for one reason or another, have missed out on this vital linguistic experience.Naima appears to be a perfect illustration of this theory. She was an only child at the time of the study and spent most of her time alone with either her mother or her father. As a result, most of the speech she heard was directed at her. This in turn meant that almost every time she heard the word “you” it referred to her. It would be perfectly understandable if she thought of "you” as simply another name for herself.Evans and Demuth note that the abrupt end of Naima’s pronoun reversal coincided with a family holiday. They speculate that the time spent with both mum and dad is what gave her the learning experience necessary to finally grasp the concept of “you”.Oshima-Takane suggests a similar explanation for the high rates of pronoun reversal in deaf and autistic children. For deaf kids, having to rely on visual communication or poor quality auditory input makes it much more difficult to follow other people’s conversations. For autistic kids, the argument goes, the problem is more that they are disinterested in other people and so fail to pay attention to their conversations. Like Naima, both groups of children will only learn from speech that directly engages them and will mistakenly jump to the conclusion that “you” only ever refers to themselves.
One should always look for a possible alternative, and provide against it.
So could this explain Ethan’s difficulties? Evans and Demuth suggest not, pointing out that, although he often used “you” to refer to himself, he used it appropriately on enough occasions to demonstrate that he’d grasped the concept.The trail led elsewhere. Say it againKanner’s explanation for pronoun reversal in autism came from another observation - that children with autism often repeat entire phrases verbatim, inappropriately and out of context. This so-called ‘echolalia’ would lead to reversals as the pronouns are repeated exactly as heard. British child psychiatrist, Michael Rutter g... Read more »
Evans KE, & Demuth K. (2011) Individual differences in pronoun reversal: Evidence from two longitudinal case studies. Journal of Child Language, 1-30. PMID: 21669013
The advent of neuroimaging techniques such as magnetic resonance imaging (MRI) has revolutionized autism research. We can now look into the brain and see the "neural correlates" of autism. But, as with any form of correlation, identifying a neural correlate doesn't necessarily mean that we have identified a neural cause.
A case in point. Earlier this week I stumbled across a press release doing the rounds of the internet, proclaiming that "Brain imaging research reveals why autistic individuals confuse pronouns". Pronouns are the words like "he", "she", "you" and "I" that can stand in for real names. Kids with autism often struggle with them (there goes another one). In particular, they'll say "you" to refer to themselves and "I" to refer to other people.Various theories have been put forward over the years to try and explain pronoun "reversal". Leo Kanner thought it happened just because the autistic kids were echoing things other people had said. Bruno Bettelheim (he of 'refrigerator mother' fame) reckoned kids with autism didn't have a sense of self, and so "you" and "I" were indistinguishable to them. An intriguing theory, proposed more recently by Yuriko Oshima-Takane is that kids with autism don't learn how pronouns work because they don't attend to other people's conversations.So what does brain imaging add to this debate?The studyThe study was conducted by Akiko Mizuno, a graduate student working with Marcel Just at Carnegie Mellon Uni. She tested a group of 15 high-functioning adults with autism on what is known in the trade as a first-order visual-perspective-taking task. On each trial, they saw a series of photographs in which a woman (called Sarah) first showed them a card with different pictures on each side and then asked "What can you see now?" or "What can I see now?" Participants had to press a button on the left or right to give the correct answer.Interpreting Sarah's questions required the participants to comprehend the pronouns "you" and "I". The adults with autism were slower and less accurate at this task than non-autistic adults. They were also a little slower on control questions that didn't involve pronouns, such as "What can Sarah see?" and "Who can see the carrot?" but the group differences weren't quite as marked. This is crucial because it suggests that the adults with autism had specific problems with the pronoun condition.These results in themselves are really interesting. They suggest that subtle difficulties with pronouns are apparent, even amongst high functioning adults with autism. It's not clear whether these individuals ever reversed pronouns themselves in their speech, and it's important to remember that the study looked at comprehension of pronouns rather than production. But it's nevertheless striking that there are group differences, even on such a simple task.The focus, however, was on the brainy stuff.While the participants were completing the task, their brains were being scanned using fMRI. The headline finding was that, in the autism group, there was reduced "connectivity" between two brain regions, the right anterior insula and the precuneus. Furthermore, within the autism group, there was a significant correlation between brain connectivity and reaction time. People who were slower had weaker connectivity.Mizuno et al. imply that this is what ultimately causes pronoun reversal:"The observed lower functional connectivity between those two neural nodes in the autism group, therefore, may result in disturbed perspective-taking processes in shifting a centre of reference between self and other."Finer grained analyses showed that group differences in "connectivity" were observed only when Sarah asked "what can you see?" and not when she asked "what can I see?". The author's explanation is as follows:"These findings indicate that the critical disturbance... may be dysfunctional processing when recognizing the self as a referent of ‘you’, and shifting to map self onto the pronoun ‘I’."In other words, when Sarah says "What can you see?", the participant has to translate that into "What can I see?" and this translation process is reliant on the "connectivity" between the precuneus and anterior insula .Reasons to be cautious:It's possible that Mizuno and colleagues are correct in their interpretation. In fact, I'd really like them to be right, because I've been waffling on about brain connectivity in autism for ages. But there are a number of reasons to query their conclusions.1. Everyone uses the term "functional connectivity" in the context of fMRI scans, but it's pretty misleading. fMRI measures brain activity indirectly via changes in blood oxygen levels. Here's an example of the time course of oxygen level changes for a control participant in Just et al.'s original fMRI "connectivity" in autism study.The two brain regions in this figure are considered to be "functionally connected" because their activation goes up and down at roughly the same time. What isn't obvious from the figure (and is rarely acknowledged) is the fact that the changes are happening really slowly - roughly one cycle of activation and deactivation every trial.If the two brain regions are 'talking' to each other in order to complete the pronoun task, they're doing it a much faster rate than anything fMRI can hope to measure.2. Since that first paper, Just and colleagues (as well as several other research groups) have published a large number of studies demonstrating changes (usually reductions) in "functional connectivity" throughout the autistic brain. Their new study adds to this impressive body of evidence. But this in turn raises a second concern.As mentioned before, Mizuno et al. looked at connectivity between two brain regions - the right anterior insula and the precuneus. Importantly, this was the only pair of regions they considered looked at . Based on their previous findings, there's a fair chance that they could have chosen any number of brain regions and would have found "underconnectivity" between them too. They may be right and this is the only connection that relates to pronoun comprehension difficulties. But we don't know this.3. The claim is that differences in "connectivity" are responsible for difficulties in comprehending pronouns. But it could just as easily be the other way around. People with autism struggle to comprehend pronouns, so they have to work harder or (as the reaction time data suggests) for longer, so it's no surprise that their brain activity while they're doing the task is different.Brains vs MindsNeuroimaging studies have provided many important insights into the workings of autistic brains. But sometimes, it's easy to be seduced by the fancy gadgets, the pretty pictures, and the funny words and think that brain imaging is somehow more scientific than good old-fashioned cognitive psychology (as exemplified by Mizuno et al.'s reaction time data), or that it offers privileged insights into the autistic mind. ... Read more »
Mizuno A, Liu Y, Williams DL, Keller TA, Minshew NJ, & Just MA. (2011) The neural basis of deictic shifting in linguistic perspective-taking in high-functioning autism. Brain : a journal of neurology. PMID: 21733887
I'm famous. Well, sort of. Earlier this week, one of my colleagues sent me a link to a YouTube video in which chiropractic doctor David Sullivan discusses one of my papers on autism and how it influences his "evidence based practice". It's a classic of its genre. The video starts off with a spinning brain and funky science-o-mercial music. And Sullivan somehow manages to equate autism with a dodgy dial-up internet connection whilst weaving our hypothesis in with Einstein and the space-time continuum. I'm flattered, but also more than a little baffled.Our paper was called "The temporal binding deficit hypothesis of autism" and came out in the journal Development and Psychopathology nine years ago (now there's a scary thought). In it we suggested that autism might be caused, at least in part, by a reduced interaction between different brain regions. Based on the idea that communication within the brain involves synchronization (or 'temporal binding') of oscillatory neural activity, we predicted that there would be reduced synchronization of brain oscillations in autism.I'm always pleased when people read and share my papers, especially when they greet it with this degree of enthusiasm. Sullivan gets a lot of things wrong but, as someone who likewise blogs on papers I find interesting but maybe don't completely understand, I can't be too critical .But there are a number of things that it's important to clarify.First and most importantly, our paper does not endorse chiropractic as a treatment for autism. We don't even mention it. To be fair, Sullivan doesn't say that we do, but if you were distracted by the spinning brain, the white coat, and the fancy neuro-terminology then you might come away with that impression.Second, while I'm prepared to admit that I know very little about chiropractic, I really can't see how people with autism might benefit from someone fiddling about with their spines. Last time I checked, autism was considered to be a form of back problem. Sullivan doesn't provide any evidence that chiropractic is a suitable treatment. He doesn't explain how it might be beneficial, even in theory. More to the point, he doesn't elaborate on how the insights gained from our paper are at all relevant to his practice.Digging around on his website, we do however get this mission statement:"Therapeutically, the goal is to restore optimal synchronization and inter-communication between all brain areas. As this process evolves, the brain becomes more cohesive in its function, and the child is able to perform at a more age-appropriate and higher functional level."But there still nothing as to how his chiropractic treatments would actually achieve this goal."There you go sir, that should sort out your social interaction difficulties"Third, our paper presented a hypothesis. We didn't show anything; there was no evidence, no data; we had an idea and ran with it. As it happens, there have since been a number of studies suggesting that autistic brains on the whole are less well-connected than your average brain. But, it's not nearly as simple or straightforward as we initially hypothesized. Different studies find that different neural pathways are disconnected. Some studies even suggest heightened connectivity. And while there's lots of evidence for abnormal brain oscillations, look more closely and the actual pattern of abnormality isn't very consistent. Another big problem is that evidence for abnormal brain connectivity has been found for umpteen other disorders that are quite different to autism. And there's a fairly compelling counter-argument that anomalous brain connections might be a consequence of autism rather than its cause.As in so many other fields of autism research, progress is being made, but each new finding generates as many questions as it does answers. We're still a long way from understanding the neurobiological basis of autism in its various manifestations. Changes in brain connectivity and neural oscillations are, I believe, part of the story. But it's going to be a very complicated story.Sullivan claims to have "a thorough understanding of the science and neurology behind ASD". If he does, he's in a minority of one.Footnote: I do make a point of emailing the authors of the paper I'm blogging and giving them the opportunity to comment and make corrections. And if there's something I'm not sure about, I run it past them before posting.Reference:Brock J, Brown CC, Boucher J, & Rippon G (2002). The temporal binding deficit hypothesis of autism. Development and psychopathology, 14 (2), 209-24 PMID: 12030688Further reading:Gimpy: The [not] libellous Simon Singh article on chiropractorsStuff & Nonsense: Chiropractic for autismThe Autism Blog: Separating the Wheat from the Chaff- How to Decide on Treatments and Therapies for Your ChildKim Wombles: Before you buy - a woo primer for parents ... Read more »
Brock J, Brown CC, Boucher J, & Rippon G. (2002) The temporal binding deficit hypothesis of autism. Development and psychopathology, 14(2), 209-24. PMID: 12030688
Another day, another autism diagnosis media storm. Last week, it was speech patterns. Before that it was urine. Today's story is that researchers in London have found a way to diagnose autism by performing multi-dimesional analyses of magnetic resonance imaging (MRI) brain scans.So what did they actually do? And how close really are we to being able to use brain scanning to diagnose autism?The studyChristine Ecker and her colleagues at the Institute of Psychiatry tested a group of 20 high functioning adults with autism, who were assessed the traditional way using either a structured interview with the parent (the ADI-R) or by video-recording a structured interaction between the person with autism and the experimenter and then coding their behaviour (the ADOS). The authors note that many of the people with autism could be considered to have Asperger syndrome but they didn't attempt to explore this issue further due to the small sample size.The 20 adults with autism, together with 20 control adults, all underwent a 15-minute MRI scan. The set-up would have been something like the picture below. From the scans, the authors extracted five different dimensions for each hemisphere of the brain that they thought might differentiate between the two groups.One of these factors, left hemisphere cortical thickness, did a pretty good job. If the thresholds were set optimally, the algorithm could achieve a 90% accuracy. In other words, based on cortical thickness, 18 of the 20 people with autism were correctly classified as having autism and only 2 of the controls were incorrectly classified as being autistic. Other measures fared slightly less well and in general the right hemisphere measures were much worse at differentiating between the two groups. They also looked at combining all the different factors, but this didn't actually improve the discrimination performance. The authors conclude their paper by stating that this is a "proof of concept". It shows that it is feasible to use these kinds of analytic techniques to investigate differences in the autistic brain.So what are the challenges to be overcome in order to go from this to an MRI-based diagnostic protocol? Here are a few off the top of my head:Does it generalise to new populations? In the current study, the choice of measures and weights were optimised to provide the best discrimination between the two groups of people in the study, whose diagnoses were already known. But we need to know that the MRI algorithm works just as well when it's given data from new people it hasn't seen before. To address this issue, the authors removed one person from each group, and optimised the MRI algorithm using data only from the remaining 38 people, so the two people removed were effectively new subjects. They then looked to see whether the two 'new' people were still correctly 'diagnosed'. They did this 20 times, removing a different pair each time and showed that they were statistically above chance at classifying the 'new' people. However, it's not clear (to me at least) how much better than chance.Does it work with kids? The participants with autism in this study were aged between 20 and 68. A diagnostic tool for autism that only works with adults is obviously of limited use. The authors are confident that the MRI strategy will work as well if not better with children, but this needs to be demonstrated. It's also worth bearing in mind that brain scanning young children (especially autistic children) is far from straightforward and often has to be done when the child is asleep or sedated. The quick and easy 15 minute scan suddenly becomes less of a reality.Can it discriminate between autism and similar disorders? Here is the crux. The 90% accuracy headline figure is in comparison to typically developing brains. I don't think the authors are really suggesting that we should screen the entire population for autism using MRI. Apart from the cost and ethical issues, it wouldn't be at all effective. Given that 10% of non-autistic people are classified as having 'autistic brains' and that only 1% of the population have autism, this means that in a random sample, most of the people identified as having 'autistic' brains wouldn't actually have autism.More useful, perhaps, would be if people who were already suspected of having autism could be given a brain scan to confirm or deny the diagnosis. People tend to think of autism as being a clear cut condition, but in truth it's a messy affair, with much overlap between different disorders and fuzzy boundaries. A 'brain' test could be really useful, but it would have to pass a much more stringent evaluation process. This would involve comparing two groups of people suspected of having autism - those who go on to have their diagnoses confirmed and those who are not ultimately diagnosed with autism. If these two groups differed consistently in their brain structures then MRI might prove to be a useful diagnostic tool.Ecker and colleagues make a step in this direction by applying their algorithm to adults with ADHD. It does OK, but still mis-diagnoses 21% of them as having autism. And there's no suggestion that anyone ever thought these people might have had autism.So why bother looking at brains?Overall, I have to question whether there really is any value in trying to use brain scans to diagnose autism. Ecker et al.'s paper begins by acknowledging the fact that autism is a highly heterogeneous disorder and should perhaps be referred to as 'the autisms' rather than a single condition. As I've mentioned before, two people can get the same diagnosis by ticking completely different boxes and recent research only goes to strengthen the impression that there are many different causes of autism at the genetic level. So does it then make sense to try and find a brain 'fingerprint' for autism? Probably not.Where I think this kind of approach might well come in useful in the future is in identifying subgroups within autism. There are already a huge number of autism brain scans that have been collected over the years and it might be possible to run similar analyses and identify clusters in multi-dimensional 'brain space' that correspond to meaningful subgroups.Perhaps one day, it will be possible for a child newly diagnosed with autism to undergo a brain scan which will indicate what subtype of autism they have. The outcome of this could help determine what intervention strategies and support are likely to be the most beneficial. If and when that day arrives, the media excitement will truly be justified.Reference:Ecker C, Marquand A, Mourão-Miranda J, Johnston P, Daly EM, Brammer MJ, Maltezos S, Murphy CM, Robertson D, Williams SC, & Murphy DG (2010). Describing the brain in autism in five dimensions--magnetic resonance imaging-assisted diagnosis of autism spectrum disorder using a multiparameter classification approach. The Journal of neuroscience : the official journal of the Society for Neuroscience, 30 (32), 10612-23 PMID: 20702694Download a free copy of the original article from the... Read more »
Ecker C, Marquand A, Mourão-Miranda J, Johnston P, Daly EM, Brammer MJ, Maltezos S, Murphy CM, Robertson D, Williams SC.... (2010) Describing the brain in autism in five dimensions--magnetic resonance imaging-assisted diagnosis of autism spectrum disorder using a multiparameter classification approach. The Journal of neuroscience : the official journal of the Society for Neuroscience, 30(32), 10612-23. PMID: 20702694
Something slightly unusual happened this week. In a paper in the journal Vision Research, Simon Baron-Cohen and colleagues reported that they had failed to find any statistically significant difference between the visual acuity of individuals with and without autism. The study was a follow-up to a 2009 paper that claimed to show enhanced (or "eagle-eyed") visual acuity in autism. Following two particularly damning commentaries by experts in vision science, the Baron-Cohen group got together with the critics, fixed up the problems with the study, and tried to replicate their original findings. They failed.While it's slightly concerning that the original study ever made it to publication, it's heartening that the authors took the criticism seriously, the concerns were addressed, and the scientific record was set straight fairly quickly. This is how science is supposed to work. But it's something that happens all too rarely.In a brilliant piece in last weekend's New York Times, Carl Zimmer highlighted the difficulty science has in correcting itself. Wrong hypotheses are, in principle, there to be disproven but it's not always that straightforward in reality. In particular, as Zimmer points out, scientists are under various pressures to investigate new hypotheses and report novel findings rather than revisit their own or other people's old studies and replicate (or not) their results. And many journals have a policy of not publishing replication studies, even if the outcomes should lead to a complete reassessment of the original study's conclusions. There is, however, a deeper problem that Zimmer doesn’t really go into.Most of the time, at least in the fields of science I'm familiar with, we’re in the business of null hypothesis testing. We're looking for an effect - a difference between two conditions of an experiment or two populations of people, or a correlation between two variables. But we test this effect statistically by seeing how likely it is that we would have made the observations we did if our hypothesis was wrong and there wasn’t an effect at all. If the tests suggest that it’s unlikely that this null hypothesis can account for the data, we conclude that there was an effect.The criteria are deliberately strict. By convention, there has to be less than a 5% chance that the null hypothesis can explain your data before you can confidently conclude that an effect exists. This is supposed to minimize the occurrence of people making grand claims based on small effects that could easily have come about purely by chance. But the problem is that it doesn’t work in reverse. If you don’t find a statistically significant effect, you can’t be confident that there isn’t one. Reviewers know this. Editors know this. Researchers know that reviewers and editors know this. Rather than being conservative, null hypothesis testing actually biases the whole scientific process towards spurious effects entering the literature and biases against publication of follow-up studies that don't show such an effect. Failure to reject the null hypothesis is seen as just that - a failure.This is something with which I'm well acquainted. My PhD was essentially a series of failures to replicate. To cut a very long story very short, a bunch of studies in the mid 90s had apparently shown that, during memory tasks, people with Williams syndrome rely less on the meanings of words and more on their sounds. I identified a number of alternative explanations for these results and, like a good little scientist, designed some experiments to rule them out. Lo and behold, all the group differences disappeared.Perhaps not surprisingly, publishing these studies turned out to be a major challenge. One paper was rejected four times before being finally accepted. By this time, I'd finished my PhD, completed a post-doc on similar issues in Down syndrome, and published two papers arising from that study. In some ways, they were much less interesting than the Williams syndrome studies because they really just confirmed what we already knew about Down syndrome. But they contained significant group differences and were both accepted first time.So the big question. How do you get a null result published?One helpful suggestion comes from Chris Aberson in the brilliantly titled Journal of Articles in Support of the Null Hypothesis. He points out that you can never really say that an effect doesn’t exist. What you can do, however, is report confidence intervals on the effect size. In other words, you can say that, if an effect exists, it’s almost certainly going to be very small.Another possibility is to go Bayesian. Rather than simply telling you that there is not enough evidence to reject the null hypothesis, Bayesian statistics provides information on how likely it is that the null hypothesis versus the experimental hypothesis is correct given the observed data. I haven't attempted this yet myself so I'd be interested to hear from anyone who has.The strategy I've found really helpful is to look at factors that contribute to the size of the effect you're interested in. For example, in one study on context effects in language comprehension in autism, we were concerned that group differences in previous studies were really down to confounding group differences in language skills. Sure enough, when we selected our control group to have similar language skills to our autism group, we found no difference between the two groups. But more importantly, within each group, we were able to show that an individual's language level predicted the size of their context effect. This gave us a significant result to report and in itself is quite an interesting finding.This brings me neatly to my final point. At least in research on disorders such as autism or Williams syndrome, a significant group difference is considered to be the holy grail. In terms of getting the study published, it certainly makes life easier. But there is another way of looking at it. If you find a group difference, you’ve failed to control for whatever it is that has caused the group difference in the first place. A significant effect should really only be the beginning of the story.Reference: Tavassoli T, Latham K, Bach M, Dakin SC, & Baron-Cohen S (2011). Psychophysical measures of visual acuity in autism spectrum conditions. Vision research PMID: 21704058Further reading:Neuroskeptic: Eagle-Eyed Autism? No.Carl Zimmer: It's Science, But Not Necessarily Right BPS Research Digest: Statistical significance explained in plain EnglishCosmos and Culture: Working within the error bars ... Read more »
Tavassoli T, Latham K, Bach M, Dakin SC, & Baron-Cohen S. (2011) Psychophysical measures of visual acuity in autism spectrum conditions. Vision research. PMID: 21704058
It’s widely believed that early intervention is crucial for long-term prognosis in autism and that the earlier the intervention begins the better. Getting in early, of course, requires that autistic children are identified at a young age. But even for more severe forms of autism, children are rarely diagnosed before three to four years of age. With this in mind, the American Academy of Pediatrics has recommended screening all toddlers for autism.However, writing in next July’s issue of Pediatrics (the academy’s own journal), Mona Al Qabandi and colleagues argue against routine population-based screening for autism. Chief amongst their objections is that existing screening tools are simply not up to the task. Most of these screens involve a questionnaire given to parents, sometimes augmented with a brief phone interview. But they all have their problems. Some are insensitive, missing a large number of kids who go on to get an ASD diagnosis further down the line. Others are sensitive but not specific, hoovering up all kinds of kids, many of whom don’t have autism, and may not have any kind of developmental problems at all.Al Qabandi et al. conclude that “none of the autism screening tests currently available has been shown to be able to fulfill the properties of accuracy… in a population-wide screening program”.Similar conclusions were reached in an earlier review by Josephine Barbaro and Cheryl Dissanayake at the Olga Tennison Autism Research Centre in Melbourne. So they tried a different approach. Rather than relying on parental questionnaires, they set up a 'surveillance program', training community nurses to spot the signs of autism during regular infant health checks.Each nurse attended a short two-and-a-half-hour workshop in which they were shown how to complete the screen. They were given a checklist with key behaviours to monitor, depending on the child’s age, and were trained how to score each item as either typical, atypical, or absent. For instance, the item for “eye contact” read as follows:"Has the child spontaneously made eye contact with you during the session? If not, interact with the child to elicit eye contact. Does s/he make eye contact with you?"From an initial sample of almost 21 thousand children, 216 were identified as “at risk” of ASD by 24 months of age. Of these, 110 completed further assessment, including the ADOS and ADI-R. 89 of these kids received an ASD diagnosis, giving the surveillance program a positive predictive value of 81%. Of the remaining 21 children, all but one had developmental language disorders.Calculating the screening program’s sensitivity is an inexact process at this stage. But assuming that the rates were similar for the children who did not undergo further assessment, Barbaro and Dissanayake estimated that approximately 175 ASD children would have been picked up. Dividing this by the total number of kids in the program gave an estimated prevalence of 1 in 119. This is reassuringly close to recent estimates of approximately 1 in 100 kids having an ASD, suggesting that the screen managed to pick up the majority of ASD kids in the initial sample.To get a more accurate indication of sensitivity, however, the researchers will have to wait until the children enter school. Only then will they be able to work out how many children end up with an ASD diagnosis but weren’t picked up by the screening measure.While it’s still early days, the Melbourne study suggests that population-wide screening for autism is possible, at least in areas that already have comprehensive child health checks. References:Barbaro, J., & Dissanayake, C. (2010). Prospective Identification of Autism Spectrum Disorders in Infancy and Toddlerhood Using Developmental Surveillance: The Social Attention and Communication Study Journal of Developmental & Behavioral Pediatrics, 31 (5), 376-385 DOI: 10.1097/DBP.0b013e3181df7f3cAl-Qabandi M, Gorter JW, & Rosenbaum P (2011). Early Autism Detection: Are We Ready for Routine Screening? Pediatrics PMID: 21669896Links:Olga Tennison Autism Research CentreFurther reading:Irresponsible Pediatrics article argues against autism screeningNot ready for introduction of routine screening for autism? Clarifying the issuesThe case against caution... Read more »
Barbaro, J., & Dissanayake, C. (2010) Prospective Identification of Autism Spectrum Disorders in Infancy and Toddlerhood Using Developmental Surveillance: The Social Attention and Communication Study. Journal of Developmental , 31(5), 376-385. DOI: 10.1097/DBP.0b013e3181df7f3c
Al-Qabandi M, Gorter JW, & Rosenbaum P. (2011) Early Autism Detection: Are We Ready for Routine Screening?. Pediatrics. PMID: 21669896
Do you write about peer-reviewed research in your blog? Use ResearchBlogging.org to make it easy for your readers — and others from around the world — to find your serious posts about academic research.
If you don't have a blog, you can still use our site to learn about fascinating developments in cutting-edge research from around the world.
Research Blogging is powered by SMG Technology.
To learn more, visit seedmediagroup.com.