Autism is a spectrum of neurological disorders that involve, primarily, reduced social aptitude. People with autism tend to make less eye contact, they have less of a response to viewing a human face, and they are less verbal. Half a century ago autism was blamed on bad parenting, but that view is now considered outdated and even cruel. Autism is a brain disorder. Neuroscientists are learning more and more about what is different about autistic brains from more typical brains. One feature seems to be reduced communication among neurons in the brain.

Autism is diagnosed clinically. It is usually first recognized by the parents, who then bring their child to medical attention and after an evaluation the diagnosis is made. At present there are no supporting laboratory tests – we don’t diagnose autism by an MRI scan, EEG, or blood test. It is diagnosed by clinical observation and some standardized questionnaires and cognitive tests. At the more subtle end of the spectrum the diagnosis may not be made right away, not until the child is a bit older and can be more thoroughly evaluated.

The median age at diagnosis was 4.4 years in 1992. This has steadily decreased, to less than 3.4 years by 2001. This effect is greater in higher socioeconomic status (SES) groups. Low SES children are diagnosed later than higher SES children, and this gap has widened in the last 20 years.

There has also been a linear increase in the number of autism diagnoses since 1992, aggregating in birth cohorts, with a greater effect for higher functioning children with autism. This suggests that more diagnoses are being made at the milder end of the autism spectrum, and at a younger age, with a strong social influence. The current dominant interpretation among experts (I’m not sure it’s robust enough to call it a consensus) is that the increase in autism diagnoses over the last 20 years is due to increased surveillance, widening the definition, diagnosing children at younger ages, and diagnosing milder cases.

A number of studies have also looked for signs of autism at younger and younger ages, with several showing differences between children who will go on to be clinically diagnosed on the autism spectrum from other children as early at 6 months of age.

Now another study has been published also demonstrating these results. Researchers looked at infants aged 6-10 months old and showed them pictures of faces with the eyes either looking toward or away from the infant. The researchers recorded brain wave activity with EEG, looking for differences in brain response between these two stimuli. A typical child should respond differently to the two images, as a human face looking toward the infant should evoke more of a response.

They found that there was indeed a difference in brain response between the two stimuli, but in those children who would later be confirmed to have autism the response was generally diminished compared to children who were not later diagnosed (at 36 months). The effect was robust, but not with sufficient sensitivity and specificity to be very useful clinically.

This is an important distinction to make. The results were statistically significant, meaning that there is very likely a real effect here. In other words, even as early as 6 months the brains of children with autism respond differently than children without autism in a way that one might predict from the symptoms of autism (diminished social responsiveness). But there were still false positives and false negatives, limiting such a test’s usefulness when applied to an individual child.

The false positives and negatives may be due to the heterogeneity of autism as a spectrum of disorders, or to limitations of the diagnostic technique itself. Probably this technique (in addition to being refined) will have to be combined with other methods of early diagnosis before application to the individual will be practical.

There are several implications of this research worth discussing. The first is that early diagnosis of autism will improve early access to intervention, which seems to make a difference in outcome. Laboratory confirmation can help to reduce doubt or confusion as to the proper diagnosis, and lead more quickly to services for young children.

There are also implications for our understanding of autism as a brain disorder. What, exactly, is happening in the brain of children with autism? This study has indirect implications for this question, but still adds to our pathophysiological knowledge of autism as a disorder.

Finally, there are implications from this research regarding the ongoing social (but not scientific) controversy of the role of vaccines and other environmental factors in autism. If the signs of autism are present at 6 months of age, then vaccines that are given after 6 months cannot be implicated as a cause of autism. This unavoidable implication, however, is unlikely to move the anti-vaccine community.

As with any single study, this new study is not  definitive. Follow up research is required, and it seems that this technique will likely have to be combined with other techniques to be clinically useful. It adds to existing research, however, increasingly pointing to the detection of differences in the brains of children with autism as early as 6 months of age.

This also fits with the dominant view that autism is likely a complex set of genetic disorders. Environmental factors cannot be entirely ruled out, but genes seem to be playing a dominant role in autism. If this is true it may be theoretically possible to push early detection of autism to even younger ages. Then again, perhaps not – perhaps the brain has to develop to a certain point before the differences in brain function are there.

Six months, so far, seems to be the point of earliest detection, and this new study supports that (although they did not study younger children). It would be interesting to apply these techniques to 2-6 month olds to see if the differences continue to show.

In any case – this and other studies like it are pointing the way toward much earlier diagnosis of autism spectrum disorders.