Shedding new light on the relationship between learning difficulties and autism
Learning difficulties are common in autism. Until recently, it was thought that the majority of people with autism also had an intellectual disability (defined as having an estimated IQ below 70). Because autism and intellectual disability often co-occur, some researchers have assumed that both conditions must share the same causes. Autism can, however, be diagnosed in people with all types of intellectual ability; some have profound learning difficulties, but others are extremely intelligent.
The recognition that autism can also occur in people with average intelligence has resulted in an increased rate of diagnoses in able people in recent years. Although autism awareness has greatly increased, some suggest that autism spectrum conditions may still often be missed in people with average or above average intelligence (Skuse, 2007). Children who have both intellectual disability and autism may be more likely to be referred to a diagnostic clinic than children with autism only. This potential bias makes it hard to pin down the exact relationship between autism and intellectual disability.
Because there are still so many unknowns about the association between autism and intellectual impairment, we designed a study to examine the relationship between these conditions in a large group of children from England and Wales. Rather than studying the association in children with diagnosed autism, in which the results may be biased, we looked at the association between intelligence and the behavioural and personality characteristics related to autism, also referred to as autistic traits, in 8,000 community-based twin pairs. Since 1994, the development of these twins has been tracked in the Twins Early Development Study at King’s College London, led by Professor Robert Plomin and funded by the Medical Research Council.
When the twins were around eight, nine and twelve years of age, parents and teachers were asked to rate the autistic traits of the twins, using a specially devised questionnaire called the Childhood Autism Spectrum Test (CAST; Williams et al., 2008). A high score on this questionnaire predicts an autism spectrum diagnosis. At age seven, nine and twelve years, the twins themselves were asked to complete a series of intelligence tests. Using these data we could assess the relationship between autistic traits and intelligence at three ages in childhood.
We were interested, initially, in whether children with a lot of autistic traits showed compromised intellectual functioning. To examine this, we selected the five per cent highest scorers on the autistic traits measure and looked at their performance on the IQ test. Likewise, we selected the five per cent of children with the lowest IQ scores and examined their autistic trait scores.
We found that children with extreme autistic traits performed somewhat less well on the IQ test. However, the association was only modest. This is illustrated in Figure 1a: In the graph on the left hand side, the proportion of children with extreme scores on the CAST are represented in red. The red line in the right hand graph shows the average score on the IQ test of these children. The average IQ score of all the twins together is set to zero in this graph; most twins obtain scores between -2 and +2. The IQ scores in children with many autistic traits were found to be slightly below average, but most of them still scored within the normal range of the IQ test. Similarly, children with low IQ scores scored only slightly higher than average on the CAST (Figure 1b).
We then explored whether the link between autistic traits and IQ changed over time. One might imagine that having extreme autistic traits may have detrimental effects on intellectual development. Likewise, intellectual problems could have a knock-on effect on autistic-like difficulties. The association between autistic traits and intelligence was, however, found to be very stable over time; at all time points in childhood the association remained modest, and a change in autistic-like difficulties did not lead to significant change in IQ scores or vice versa.
Because the study focussed on twins, we could also examine the extent to which autistic traits and IQ are influenced by genetic or environmental factors. This is because identical twins share all their genetic code, whilst non-identical twins on average only share half of their genes (just like ordinary brothers and sisters). If identical twins resemble each other more closely than do non-identical twins, this provides evidence for genetic influences.
From previous studies we already knew that autistic traits are highly heritable (Ronald et al., 2006a; 2006b); identical twins resemble each other much more strongly in their autistic traits than do non-identical twins. But what was new in our current study was that we could test whether autistic traits and intellectual difficulties were caused by the same genetic influences. We found only a modest genetic overlap between low IQ and high autistic traits (Hoekstra et al., 2009; in press). So, although autistic traits are strongly genetic, most of these genetic effects are independent of the genetic influences on IQ.
These results challenge the assumption that autism and intellectual disability share many of the same causes, and this has important implications. In recent years, research groups around the world have started to collaborate in large multi-million dollar initiatives to find the genes involved in autism. These studies have resulted in some promising discoveries, and a variety of genes have been detected that are thought to play some, although often small, part in autism. Most of the genes that have thus far been detected have a broad role in the development and function of the brain. It is unclear how these genes could influence the risk for autism without also compromising intellectual functioning. Our new results suggest it will be vital to also search for genes that influence the risk for autism without impacting on general intellectual ability.
Our finding that the characteristics of autism are largely independent of intellectual functioning stands in marked contrast with what is seen in clinics, where intellectual disability is common in people with an autism diagnosis. It remains a puzzle why intellectual impairment is so common in autism. In part, this may reflect a bias to diagnose autism in lower functioning individuals more readily than in intellectually able people. Highly intelligent children may have developed compensatory strategies that mask their autistic traits. They may, for instance, spend their lunch break in the computer room rather than in the play ground. Such strategies might make it harder to detect autism in these children. Diagnosis in children with average or above average intelligence is often established late, after a long referral process. It is important that teachers and health workers remain alert to the signs of autism in able children, so that problems can be detected early and appropriate services can be provided for the children and their families.
Dr Rosa Hoekstra is a lecturer in psychology and teaches the course Understanding the Autism Spectrum (SK124) at The Open University:
Francesca Happé is a professor in cognitive neuroscience at King’s College:
Simon Baron-Cohen is a professor in developmental psychopathology at the University of Cambridge:
Dr Angelica Ronald is a lecturer in behaviour genetics at Birkbeck College:
We are indebted to Professor Plomin for the use of the Twins Early Development Study data and gratefully acknowledge the twin families for their participation in this study. For more information about the Twins Early Development Study, visit:
The twins depicted in the photographs did not take part in the study. Photographs courtesy of the National Autistic Society:
Hoekstra, R. A., Happé, F., Baron-Cohen, S., & Ronald, A. (2009). Association between extreme autistic traits and intellectual disability: insights from a general population twin study. The British Journal of Psychiatry, 195, 531-536.
Hoekstra, R. A., Happé, F., Baron-Cohen, S., & Ronald, A. Limited genetic covariance between autistic traits and intelligence: Findings from a longitudinal twin study. American Journal of Medical Genetics. Part B, Neuropsychiatric Genetics. In press.
Ronald, A., Happe, F., Bolton, P., Butcher, L. M., Price, T. S., Wheelwright, S., Baron-Cohen, S., et al. (2006a). Genetic heterogeneity between the three components of the autism spectrum: a twin study. Journal of the American Academy of Child and Adolescent Psychiatry, 45(6), 691-699.
Ronald, A., Happe, F., Price, T. S., Baron-Cohen, S., & Plomin, R. (2006b). Phenotypic and genetic overlap between autistic traits at the extremes of the general population. Journal of the American Academy of Child and Adolescent Psychiatry, 45(10), 1206-1214.
Skuse, D. H. (2007). Rethinking the nature of genetic vulnerability to autistic spectrum disorders. Trends in Genetics, 23(8), 387-95.
Williams, J., Allison, C., Scott, F., Bolton, P., Baron-Cohen, S., Matthews, F., & Brayne, C. (2008). The Childhood Autism Spectrum Test (CAST): Sex Differences. Journal of Autism and Developmental Disorders, 38(9), 1731-9.