AI machine learning reveals how brain anatomy changes with autism

We have come a long way in our understanding of autism since it was first used as a clinical description in 1943. Scientists have identified some genes that appear to play a role, and have developed treatments that can improve the quality of an individual with autism. life. Despite these achievements, modern science has revealed only the tip of the iceberg of the neurodevelopmental disorder. For example, scientists have yet to map all the parts of the brain that are affected and how these structural differences lead to myriad symptoms and behaviors.

The main challenge standing in the way is autism itself – there is a mountain of individual differences when it comes to the underlying biology of the disorder. However, one group of neuroscientists aims to find all the points in the autistic brain and connect them with some help from artificial intelligence.

in Thursday paper was published in the magazine SciencesResearchers at Boston College used machine learning — a type of artificial intelligence that learns and improves from experience — to determine the types of anatomical differences in the brain caused by autism versus other factors such as age or gender.

“It’s a technological innovation,” James McBartland, a clinical psychologist and autism researcher at Yale University, who was not involved in the study, told The Daily Beast. “It’s really hard to spot autism in a lot of signals [noisy data]. And the more ways you can meaningfully analyze this noise, the more power you have to detect signals that are meaningful in understanding the neuroscience of autism.”

When it comes to autism neuroscience, there’s been this move toward precision medicine where the diversity of the disorder can be categorized and defined based on shared symptoms, Aidas Aglinskas, a Boston College neuroscientist and lead author of the new research paper, told The Daily Beast.

“[There’s] This is a common idea in precision medicine to try to find subgroups of autism — let’s say autism A and autism B — that might have different symptoms,” Aglinskas said. a or b. Instead, what we saw, surprisingly, was that the enormous amount of variance was greater than these individual subspecies could capture.”

To reveal this huge asymmetry in the brain, Aglinskas and colleagues studied and compared functional MRI scans between autistic and non-autistic individuals, taken from the Autism Brain Imaging Data Exchange (ABIDE). Boston College Group AI has learned to filter for common neuroanatomical differences that all brains have in common (such as those related to age and gender) and specific brain regions that appear to be associated with autism.

“[Our machine learning] It highlighted a large number of areas of the brain, many of which have been associated with known symptoms of autism,” Aglenskas said. “So things like the areas in the brain that are involved in social cognition, thinking about others, the motor and sensory cortex, areas in language.”

These findings are a wonderful first step in mapping the autistic brain and creating a more accurate roadmap for future research, Mac Bartland said. But he cautioned that this study, while recent, has no immediate clinical applications largely due to the current lack of treatments targeting a specific part of the brain.

“Regardless of how well we understand the potential neuroanatomical difference, we as clinicians working with neuroscientists have a long way to go before we are in a place where we can say, ‘Aha, there is a structural difference in this brain region and this is how we can be helpful.'”

Aglinskas agreed and admitted that going forward, his team wants to include a wide range of brain measurements, such as an electroencephalogram, which captures the brain’s electrical activity, and genetic data to see if these complement what their AI has found or perhaps provide a new perspective.

While using machine learning to break the black box of autism is still a work in progress, Aglinskas hopes it will be of great value and possibly have more clinical impact on other neurological disorders that need a more nuanced understanding.

“Motivation is something that is particularly evident in autism, but it is not, by far, the only heterogeneous disorder,” he said. “[There is] Attention deficit disorder or depression as there is a lot of variance out there. These methods can easily be applied to study variance in those areas as well.”