Yes you can according to this new study on brain development in infants. Autism can now be detected in infants aged between 6-24 months – at least one year earlier than when the symptoms of Autism start to show. So does this mean that Autism can be detected in infants before they get their MMR vaccine?
Just FYI, this post is not an explanation about what vaccines do, what Autism is and why the number of measles, mumps and rubella are on the rise. You can find all this out from this excellent and easy to digest blog post by Jennifer Raff and this video by the Healthcare Triage. What this post is about is the brand new finding that Autism can be detected in children before the symptoms appear.
Bit of background information, psychiatrists found that children and adults with Autism have slightly overgrown brains! As weird as that sounds, this was first shown over 20 years ago. And since the development of new scanning and imaging technology, like MRI, psychiatrists can now detect these brain size differences in infants as young as 6 months old. This isn’t the only study that is researching the link between brain-size development and Autism, but what makes this study so interesting is that they’ve used their knowledge and findings to predict more accurately than ever if a child is Autistic.
Time for a quick and dirty anatomy lesson. There’s an outer layer of the brain called the cerebral cortex. It’s made up of grey matter (giving it a greyish-pink in colour), is about 2-3mm thick, contains a huge mass of neuronal cells and is responsible for lots of different functions, like memory, speech, muscle control, seeing, listening, creativity, judgement, the list goes on. So the cerebral cortex is a pretty insanely complex piece of brain. Below you can see the four different lobes (areas) of the cerebral cortex and their main functions.
What this Autism study did was scan and measure the brain size (total brain volume) and how much the surface area of the cerebral cortex (surface area expansion) in children when they were 6 to 12 months old. There were two groups of children: high-risk and low-risk. The high-risk children were children who were from families with a family member(s) who was also on the Autism spectrum. Low-risk children were children who did not have any history of Autism in the family (here’s some more info if you want to know more about Autism running in families). The researchers then later noted whether (or not) the children were diagnosed with Autism when they reached 24 months old.
What this study found was that the children who were diagnosed with Autism had expanded cerebral cortices back when they were between 6 and 12 months old. So yes! There’s a link between the brain size of children and whether they will be diagnosed with Autism or not. But not all the children who had enlarged cerebral cortices were later diagnosed with Autism. So the researchers went one step further and developed a Deep-learning algorithm to predict if a child will be diagnosed with Autism by measuring the size of their cerebral cortex.
Now I’m not a Deep-learning expert (in fact I’ve only recently heard of the idea). But I’ll do my best to explain what Deep-learning is.
Deep-learning algorithms work (roughly) like neural networks in the brain, except the network is artificial. These artificial networks recognise patterns in real-world data, such as images, sound, text, and so on. When the data is input into a Deep-learning algorithm, the network checks the data for similarities, and then clusters and classifies the data. The network can then be trained, using this data, to learn. Just like your brain does. Ultimately, the more information you feed (or input) into a Deep-learning algorithm, the more accurate your results will be. So by feeding this Deep-learning algorithm information about Autism, diagnosing Autism, brain development, expanded cerebral cortices, and so on, the more data it has and the more the network can learn. And the more it learns, the better and more accurate its predictions will be. If you’d like to know more about Deep-learning you can watch the video below and there’s also some great resources on this blog post by Jason Brownlee.
The researchers developed this algorithm to predict if high-risk children would be diagnosed with Autism at 24-months. Their algorithm had an incredible 81% positive predictive value, which means that the algorithm will accurately diagnose 81% of children with Autism at 6 to 12 months before their symptoms appear. And the algorithm is still learning!
And just to throw in that final burning question: If children can be diagnosed with Autism before their symptoms appear and before they’re given the MMR vaccine, how can the MMR vaccine cause Autism?
Exactly.
Written by Alison Holland
Updates
I came across this interesting article about a neuroscientist and his son whose brain could unlock Autism. What caught my eye about this article is it talks about a drug called VPA used to control seizures (also known as Depakote), which when taken in high doses while pregnant can increase the risk of Autism seven fold! His research opened doors to other theories and models about neuro-connectivity, Autism, social and cognitive disorders, and much more that have also been and are continually being tested and validated.
Why am I talking about this article? Because in science it is important to consider all possibilities. This is exactly what this particular neuroscientist did and what the man who fueled the anti-vaccine movement and caused the MMR outbreak did not do. Enjoy reading their story.
References
Hazlett et al. (2017) Early brain development in infants at high risk for autism spectrum disorder
https://deeplearning4j.org/neuralnet-overview
https://www.quora.com/What-is-deep-learning
http://www.bbc.com/news/health-38955872
https://www.dartmouth.edu/~rswenson/NeuroSci/chapter_11.html
Image courtesy of the NIH Office of Science Education
The Digestible Brain 