The concept of an autistic disorder is generally accepted to have originated with Leo Kanner and his classic 1943 account of ‘Autistic Disorders of Affective Contact’ (Kanner, 1943). He described 11 children with what would now be regarded as severe autism. Three of these 11 were girls, and as children their symptomatic profiles were very similar to those of the boys. It is less well-known that he conducted a follow-up study in 1971, and discovered that all those girls were in State Hospitals, virtually uncommunicative to both staff and their families (Kanner, 1971). As diagnostic criteria developed, and standardized instruments were devised to measure autistic traits, Kanner’s male-typical template – associated with severe to profound learning disability – was used for validation. For many years, diagnosticians regarded the typical child with autism as male with generalized learning difficulties. A high proportion were reported to have associated neurological complications such as early developmental regression and/or epilepsy. The male:female ratio is widely found to be 4:1 in ascertained cases, a figure that disguises the remarkable fact that among those with very low IQ the ratio is only 2:1. Leading clinical centres in the USA have almost never diagnosed ‘high functioning females’, so in North America there is a dearth of research on girls and women with autistic traits whose verbal IQ is above average.
Recent epidemiological evidence from the National Center for Health Statistics reports changes in the apparent prevalence of autism spectrum disorders in school age over the period from 2007 to 2014 (Blumberg et al, 2013; Zablotsky et al, 2015). This survey reflected diagnoses that were provided to families by paediatricians from around the United States. The figures show the expected rapid increase. They also reveal that by far the greatest change has been among children in mid to late adolescence, the great majority of whom do not have severe to profound learning disabilities. This increase in ascertainment has been predominantly of boys. By 2012 about 1 in 30 boys in the USA had been given a diagnosis of an ASD, compared with just 1 in 140 girls. In 2014, there was a change in the way the telephone survey was conducted. That change led many children who had previously been categorized as having primary generalized developmental delay to be reassigned to an ASD category. Reassignment impacted primarily on girls, where prevalence jumped over 60% from the 2011-2012 figure (compared with a mere 2% increase in boys), and the findings have two significant implications. First, ASD is still disproportionately recognised in boys; girls in adolescence are rarely newly diagnosed (at least, in the USA). Second, there is a persisting tendency to recognise autism in girls only if it is associated with generalized developmental delay.
The DSM-5 defines two dimensions of impairment associated with an ASD, comprising social communication deficits, and repetitive/stereotyped interests and behaviors, with sensory sensitivities. If there are girls out there in the general population with these characteristics, who have not been diagnosed, how would we recognise them? If there really is an ascertainment bias, one possibility is that those females with higher IQ, especially verbal IQ, are able to disguise their symptoms – a process that is known as ‘compensation’ (Lai et al, 2011). The motivation for compensation comes both from influences that are common to all females. These include, first, a biologically engendered greater awareness of other females’ social behaviour, plausibly related to the possession of a second X-chromosome (Skuse, 2000) and, second, the influence of socialization pressures that have evolved in concert with those biological differences (Adler et al, 1992). Clinical experience of interviewing many ‘high-functioning’ females with ASD in the UK confirms a widespread ‘wish to act normal’ which is not so prevalent among affected males at any age.
As we lack a biomarker for ASD, the opportunities to validate the hypothesis that there is a large number of undiagnosed females with ASD seem limited. One approach we have taken is to evaluate ASD traits by screening a general population sample of typical children, ascertained in the course of a UK birth cohort study (ALSPAC study team, 2001). Unlike most trait measures, our screening questionnaire of social communication competence was not developed from a male ASD prototype, but from a female prototype (Skuse et al, 2005). The provisional sex ratio of those boys and girls at very high risk of ASD was found to be 2:1. This is exactly the same sex ratio as has been reported for children with ASD who have severe to profound learning disabilities (Fombonne, 2003). We draw the inference that females who are not learning disabled are often motivated and capable of ‘compensation’. Even if they have significant autistic traits, their autistic vulnerabilities are not being recognised clinically; the higher their IQ, the greater the ascertainment bias.
We concluded that females with ASD traits and normal/high IQ are less likely to be identified clinically, and we have found evidence from this general population study (ALSPAC) that high verbal IQ is protective for females (Skuse et al, 2009). Genetic risk in the ALSPAC cohort was nevertheless shared with clinically identified samples of ASD, implying that the same neural substrates were likely to be responsible both for clinically significant manifestations and for more subtle ASD traits (Robinson et al, 2016). In other words, there is a continuum of genetic risk between clinically identified autism and autistic traits in the general population – the term ‘neurotypical’ is only relatively true for either sex.
If females at high ASD risk are really able to compensate at least in part for their difficulties, then this should mean they have learned how to interpret social cues by a conscious rather than by an intuitive process. Using a novel test of emotion recognition (Boraston et al, 2007) that they would not have seen before (and for which they could not have learned the appropriate response), we found that high-risk girls in our general population sample had great difficulty answering correctly. Their performance was no better than that of high-risk boys (Kothari et al, 2016). Yet on a comparison task of face-recognition emotion (a skill that could have been learned), they performed perfectly well (unlike the boys).
Does it matter if girls with ASD traits are not diagnosed, if they are able to compensate for their difficulties? Yes, it does. They cannot cover up those deficits forever, and in any event the process of doing so is highly stressful. There is a developmental pattern, by which reasonably successful adjustment to the social environment is often achieved during the preschool and the prepubertal period. Once puberty and adolescence supervenes, the social world becomes so complex for girls with ASD that they find it hard to cope any longer. They typically respond with internalizing problems, including depression and anxiety, perhaps with self-harm and school-refusal, with psychosomatic features. One striking characteristic, rarely recognised by clinicians, is that there can be major differences in their behavior at home and at school (Mandy et al, 2012). Intense efforts are put into compensating for social difficulties at school, the typical high functioning female with ASD being the ‘perfect child’, quiet and well behaved. In contrast, the child who is lauded at school for her diligence and excellent deportment is, at home, sullen, aggressive and even violent. Failure to recognise that this home-related behavior is not due to inadequate or inappropriate parenting, but rather a reaction to the stress and daily exhaustion of disguising her social-communication difficulties in the school environment, can lead to months of inappropriate family therapy. Parent-blaming does not help anyone to deal with the underlying issues.
In summary, there is emerging evidence that the true sex-ratio in ASD is 2:1 across the entire range of verbal intelligence. Ascertainment bias has bedeviled the interpretation of research studies into genetic risk, because there has been an over-representation of girls with developmental learning difficulties in such samples – generalized learning disorders can have their own genetic etiology and confound studies into ASD-risk genes (Skuse, 2007). Research into female-typical ASD has been hampered by the fact that the diagnostic template for the condition has, for 70 years, been male. Most standardized instruments that are widely used in ASD research in North America reflect this bias, thus studies that use those instruments are destined to perpetuate the myth that autism is extraordinarily rare among ‘high functioning females’ – what used to be called Asperger syndrome. We are failing those female children and women if we do not recognise and treat their social communication problems; it is a matter of urgency to do so.
David H. Skuse, MD, FRCP, FRCPsych, FRCPCH, is Professor of Behavioural and Brain Sciences at the University College London and Honorary Consultant in Developmental Neuropsychiatry at the Great Ormond Street Hospital for Children, London. For more information, please visit https://www.ucl.ac.uk/ich or email Dr. Skuse at firstname.lastname@example.org.
Adler, P. A., Kless, S. J., & Adler, P. (1992). Socialization to gender roles: Popularity among elementary school boys and girls. Sociology of education, 169-187.
ALSPAC Study Team. (2001). ALSPAC–the Avon longitudinal study of parents and children. Paediatric and perinatal epidemiology, 15(1), 74-87.
Blumberg, S. J., Bramlett, M. D., Kogan, M. D., Schieve, L. A., Jones, J. R., & Lu, M. C. (2013). Changes in prevalence of parent-reported autism spectrum disorder in school-aged US children: 2007 to 2011–2012. National health statistics reports, 65(20), 1-7.
Boraston, Z., Blakemore, S. J., Chilvers, R., & Skuse, D. (2007). Impaired sadness recognition is linked to social interaction deficit in autism. Neuropsychologia, 45(7), 1501-1510.
Fombonne, E. (2003). Epidemiological surveys of autism and other pervasive developmental disorders: an update. Journal of autism and developmental disorders, 33(4), 365-382.
Kanner, L (1943) Autistic disturbances of affective contact. Nervous Child, 2, 217-250
Kanner, L. (1971). Follow-up study of eleven autistic children originally reported in 1943. Journal of autism and childhood schizophrenia, 1(2), 119-145.
Kothari, R., Skuse, D., Wakefield, J., & Micali, N. (2013). Gender differences in the relationship between social communication and emotion recognition. Journal of the American Academy of Child & Adolescent Psychiatry, 52(11), 1148-1157.
Lai, M.C., Lombardo, M.V., Pasco, G., Ruigrok, A.N., Wheelwright, S.J., Sadek, S.A., Chakrabarti, B., Baron-Cohen, S. and MRC AIMS Consortium (2011). A behavioral comparison of male and female adults with high functioning autism spectrum conditions. PloS one, 6(6), e20835.
Mandy, W., Chilvers, R., Chowdhury, U., Salter, G., Seigal, A., & Skuse, D. (2012). Sex differences in autism spectrum disorder: evidence from a large sample of children and adolescents. Journal of autism and developmental disorders, 42(7), 1304-1313.
Robinson, E.B., St Pourcain, B., Anttila, V., Kosmicki, J.A., Bulik-Sullivan, B., Grove, J., Maller, J., Samocha, K.E., Sanders, S.J., Ripke, S. and Martin, J (2016). Genetic risk for autism spectrum disorders and neuropsychiatric variation in the general population. Nature genetics, 48(5), 552-555.
Skuse, D. H. (2000). Imprinting, the X-chromosome, and the male brain: explaining sex differences in the liability to autism. Pediatric Research, 47(1), 9-9.
Skuse, D. H. (2007). Rethinking the nature of genetic vulnerability to autistic spectrum disorders. TRENDS in Genetics, 23(8), 387-395.
Skuse, D. H., Mandy, W. P., & Scourfield, J. (2005). Measuring autistic traits: heritability, reliability and validity of the Social and Communication Disorders Checklist. The British Journal of Psychiatry, 187(6), 568-572.
Skuse, D.H., Mandy, W., Steer, C., Miller, L.L., Goodman, R., Lawrence, K., Emond, A. and Golding, J. (2009). Social communication competence and functional adaptation in a general population of children: preliminary evidence for sex-by-verbal IQ differential risk. Journal of the American Academy of Child & Adolescent Psychiatry, 48(2), 128-137.
Zablotsky, B., Black, L. I., Maenner, M. J., Schieve, L. A., & Blumberg, S. J. (2015). Estimated Prevalence of Autism and Other Developmental Disabilities Following Questionnaire Changes in the 2014 National Health Interview Survey. National health statistics reports, (87), 1-21.