Pregnancy and child

Can autism really be detected in babies?

“Autism can be identified in babies as young as two months, early research suggests,” BBC News reports. Using eye-tracking technology, researchers claim to have identified subtle differences in the way affected babies respond to visual prompts, it says.

The news was based on a small study involving baby boys thought to be at high risk of autistic spectrum disorder (ASD) due to them having a brother or sister with the condition (around one in 20 cases of ASD are related to family history).

By monitoring the eye movements of babies from two months to 24 months old, while they were shown videos of a female actor inviting them to play, the researchers found that children who were subsequently confirmed as having ASD had more limited eye contact than their peers.

The researchers found that the average level of eye looking (in other words looking at the eyes of the actor on-screen) was the same in both groups at two months, but after this time the ASD group had reduced eye contact.  

The researchers did not suggest that this will be a new test for ASD in babies – a diagnosis of ASD is made by looking at a wide range of tests and behavioural interactions, usually from the age of two (see box).

Encouragingly, if eye contact is normal at two months, other functions may also be normal. So, with early intervention, it may be possible to prevent ASD developing. Exactly how this could be achieved remains unclear.

This study does not provide any practical advice for parents wanting to confirm or disprove a diagnosis of ASD. The researchers used specialised eye-tracking technology in an artificial environment. There may have also been many other confounding factors that could have accounted for the results.

If you are worried about your child’s behaviour or development then talk to your GP.

Where did the story come from?

The study was carried out by researchers from the Emory University School of Medicine, Atlanta, US and was funded by grants from the Simons Foundation and the National Institute of Mental Health and support from the Marcus Foundation, the Whitehead Foundation and the Georgia Research Alliance.

The study was published in the peer-reviewed journal Nature.

The BBC News reporting of the study was initially of variable quality and gave a mixed representation of the study. It usefully provided expert opinion that “autism is a very complex condition… no two people with autism are the same, and so a holistic approach to diagnosis is required that takes into account all aspects of an individual’s behaviour”. It also claimed that autism could be identified in babies as young as two months, which is not the case. However, it later clarified this point and commendably updated its story.

The BBC story also did not report the most interesting aspect of the study. The fact that, contrary to expectations, babies with ASD have unaffected eye contact ability at the two-month mark.

What kind of research was this?

This was a cohort study following infants from age two months to 24 months. It aimed to find out when the reduced eye contact that is typically seen in ASD occurs, by comparing the results of children diagnosed with ASD at 36 months to typically developing children.

What did the research involve?

The researchers recruited 59 infants who were at high risk of an ASD (they had a sibling with a diagnosis of ASD), and 51 infants who were at low risk of ASD (no first, second or third degree relative with ASD). They did this to try and make sure that their group of infants would contain some who would develop ASD.

The researchers showed the infants videos of a female actress inviting them to play and measured the amount of time the infants looked at the eyes, mouth, body and objects using eye-tracking equipment called ISCAN. They performed this test 10 times, at two, three, four, five, six, nine, 12, 15, 18 and 24 months of age.

By the time the infants were 36 months old, 13 children had been diagnosed with ASD (12 from the high-risk group and one from the low-risk group). The majority of the cases diagnosed were male. To avoid differences due to gender, the researchers only analysed boys. The researchers compared the eye-tracking results of 11 boys who had been diagnosed with ASD (10 from the high-risk group and one from the low-risk group) to 25 boys who had not (all from the low-risk group). They wanted to see when the reduced eye contact, which can be a feature of an ASD, was present before obvious symptoms of an ASD.

To prevent bias of results, the researchers administering the tests were not told which infants were at high or low risk, or whether any had already had a diagnosis of ASD, and the clinicians who diagnosed ASD were not aware of the results of the eye-tracking tests.

What were the basic results?

They found that the average amount of eye looking was the same in both groups at two months. The group that was developing typically showed increased eye contact of 3.6% per month from two to six months, while the ASD group showed reduced eye contact of 4.8% per month during this time.

The infants developing typically looked more at the eyes than at the mouth, body or objects from two to six months, but the infants with ASD showed a reduction in eye fixation from two to 24 months, which was half that of typically developing infants by 24 months.

Mouth fixation increased during the first year and peaked at 18 months in both groups. In the ASD group, eye fixation on the body reduced at less than half the rate seen in the typically developing infants and stabilised at 25% higher. Fixation on an object also reduced more slowly and increased during the second year. By 24 months it was twice the level of typically developing infants.

How did the researchers interpret the results?

In infants later diagnosed with ASD, early levels of eye looking are normal, but then levels fall. This contradicts prior hypotheses of an absence of social adaptive orientation from birth – that is babies with ASD are born with “hardwired” poor social skills.

Instead, the results showed that some social adaptive behaviours may initially be intact in newborns later diagnosed with ASD. If confirmed in larger samples, this would offer a remarkable opportunity for treatment. The developing brain has a great degree of what is known as plasticity – that is the “neural circuits” can be changed through treatment and training.

Or as the researchers put it there may well be “a neural foundation that may be built upon, offering far more positive possibilities than if that foundation were absent from the outset”.


This was a very small study that suggests that infants who develop ASD may not have the symptom of reduced eye contact from birth. The results should be interpreted with caution though, as it took place in a highly artificial environment. Eye contact was measured in response to a video of a woman and not a real, live person, and there may have been many other confounding factors that could have accounted for the results.

Limitations to this study include:

  • very small number of participants
  • only male infant data was analysed, and although boys are three to four times more likely to develop ASD than girls, it meant that the sample size was even smaller
  • watching a video is not the same as human interaction, and eye-tracking of infants may not be entirely accurate

We do not know the causes of ASD, but they are believed to be multifactorial, including environmental and genetic factors. A diagnosis at 24 months would be provisional and made on the basis of a number of symptoms, not just reduced eye contact. There is no evidence of the impact of earlier diagnosis, though this could improve the level of support that is provided to parents.

The researchers suggest that the eye fixation ability was normal at birth but declined by six months. While this offers hope that there may be a window of opportunity to prevent the onset of ASD, far more research is required to understand the causes of ASD and a way of preventing it.

Still, due to the fact that ASD is currently incurable, a potential window of opportunity is a very exciting prospect.

NHS Attribution