Your browser is no longer supported

For the best possible experience using our website we recommend you upgrade to a newer version or another browser.

Your browser appears to have cookies disabled. For the best experience of this website, please enable cookies in your browser

We'll assume we have your consent to use cookies, for example so you won't need to log in each time you visit our site.
Learn more

Is it harder to detect autism in girls?

  • Comment

“Thousands of girls may have autism that has never been diagnosed because they cover-up the signs so well,” the Mail Online reports.

The headline is prompted by a study focusing on one of the key symptoms of autistic spectrum disorder (ASD): a failure to recognise the emotional states of others.

Researchers wanted to see if there was a gender difference associated with this symptom.

Children were assessed as having ASD-like traits if they scored highly on a well-validated checklist assessing social reciprocity (responding to others’ emotional states) and other verbal and nonverbal social traits.

The children were also given two tests to assess how well they could recognise emotion. The first was a test of how well they could distinguish emotions from photos of faces of other children expressing emotion such as happiness of fear.

The second test was more subtle. They were shown an animation of a triangle and a circle animated in such a way as to convey emotions; such as a jaunty bouncing to convey happiness or a slow slouching movement to convey sadness.

This test, known as the Emotional Triangles Task, is designed to assess the ability to detect recognise emotion from movement (a real-world equivalent would be trying to judge somebody’s likely mood by the way they are walking).

Researchers found that girls with ASD-like traits could recognise emotion from photos of faces as well as girls without ASD-like traits. However boys with ASD-like traits performed worse than boys without ASD-like traits on this test.

But both girls and boys with ASD-like traits struggled with the Emotional Triangles Task.

The concern now is that girls with ASD may be going misdiagnosed and are not receiving the support they need. This finding may have implications on how ASD is diagnosed in girls. 

Where did the story come from?

The study was carried out by researchers from University College London and was funded by the National Institute of Health research and Wellchild, a UK charity for sick children.

The study was published in the peer-reviewed Journal of the American Academy of Child and Adolescent Psychiatry.

This story was covered by the Mail Online. The majority of the research was well-reported. Although it should be noted that the assertion that untreated girls with autism may be prone to eating disorders and depression in later life was based on a quote from the National Autistic Society.

Long-term outcomes of the participants were not assessed by the study.

What kind of research was this?

This was a cross-sectional study that aimed to investigate the potential association between autistic-type traits and emotion recognition, and whether this differs in boys and girls. This is the ideal study design to investigate this question.

What did the research involve?

The researchers used data from 3,666 children participating in the Avon Longitudinal Study of Parents and Children (ALSPAC); an on-going cohort study of children and parents.

The children were assessed using the Diagnostic Analysis of Non-Verbal Accuracy (DANVA) – the facial recognition test at 8.5 years of age and by the Emotional Triangles Task at 13.5 years.

The parents had also returned the results of the Social Communication Disorders Checklist (SCDC), a well-validated tool designed to detect for autistic-like social communication deficits at 13.5 years of age.

The SCDC measures social reciprocity, for example responding positively to positive actions such as kindness, and other verbal/nonverbal social traits.

A higher SCDC score is an indication of more deficits in social communication, and a score of nine or above is predictive of ASD. DANVA included a test assessing facial emotion recognition, by showing photographs of children expressing happiness, sadness, anger, or fear.

The Emotional Triangles Task assessed emotion recognition from the movement of objects, in this case a triangle and a circle moving around a screen during five-second animations.

The researchers compared performance on the DANVA and the Emotional Triangles task between children who scored higher or lower on the social communication disorders checklist. They then analysed boys and girls separately.

What were the basic results?

Children who scored nine or more on the SCDC had higher odds of:

  • making at least seven errors in facial emotion recognition (all faces)
  • making at least three errors in facial emotion recognition when shown ‘high-intensity’ faces
  • making at least five errors in facial emotion recognition when shown ‘low-intensity’ faces
  • making at least three errors in facial emotion recognition when shown fearful faces
  • making a least two errors in facial emotion recognition when shown sad faces
  • misattributing faces as happy at least four times

However, when boys and girls were analysed separately, it was found that girls scoring nine or more on the SCDC were not at increased odds of making any type of error in facial emotion recognition than girls who scored less than nine.

However, boys who scored nine or more on the SCDC had higher odds of making errors in the recognition of emotion in all types of faces compared to boys who had scores of less than nine. They also had higher odds of misattributing faces as happy.

Children with high SCDC scores performed worse on the Emotional Triangles Task, where they had to recognise emotions from the movement of objects.

Higher SCDC scores were associated with poorer emotion recognition in the happy and sad conditions. When boys and girls were analysed separately it was found that girls with high SCDC scores had poorer emotion recognition in the happy and sad conditions than girls with low scores. Boys with high SCDC scores had poor emotion recognition in the happy condition.

How did the researchers interpret the results?

The researchers conclude that “autistic-like social communication difficulties were associated with poorer recognition of emotion from social motion cures in both genders, but were associated with poorer facial emotion recognition in boys only”.

They speculate that this might be because girls learn to compensate for facial emotion recognition difficulties and it is this speculation that caught the media’s attention.

They go on to say that “the implications of this are far reaching with regard to the diagnosis of autism spectrum disorders in females, suggesting that more subtle assessment may be required to identify those individuals with difficulties”.


This was a large interesting and useful study. It suggests that both boys and girls with defects in social communication skills suggestive of ASD have difficulties in emotion recognition from social motion, tested using the Emotional Triangles Task.

However, only boys with ASD-like social communication deficits have difficulties in recognising emotion from faces – girls with ASD-like traits perform as well as girls without ASD-like traits on this task.

This suggests that there may be differences in the characteristic traits associated with ASD between boys and girls, and implies that the criteria used to diagnose ASD may need to be gender-specific.

This study has one main strength in that it used data from a large number of children. However, the authors point out that it also has limitations, including that the group of children studied (from Avon, England) have been found not to be representative of the UK child population as whole.

So it is not known whether the findings are applicable to other groups across the country. Unfortunately, the study doesn’t specify how the children living in Avon were different from the general UK population but one might speculate there may be differences in ethnic diversity or socioeconomic background in this area, compared to the UK average.

They also point out that fewer girls than boys scored above the threshold of nine on the SCDC.

This may mean that the numbers of girls with scores above nine wasn’t big enough to detect a significant difference in facial emotion recognition, rather than there not being one.

It should also be noted that this study looked at autistic-type traits, but did not confirm a diagnosis of ASD.

Further research will be required to confirm the findings of the study, and to investigate if, how or why, girls might be able to compensate for facial emotion recognition difficulties. 

  • Comment

Have your say

You must sign in to make a comment

Please remember that the submission of any material is governed by our Terms and Conditions and by submitting material you confirm your agreement to these Terms and Conditions. Links may be included in your comments but HTML is not permitted.