AI reveals an association between individual differences within the anatomy of the brain and individual differences within the symptoms of autism spectrum disorders

Overview: Different symptoms of ASD can affect different brain regions and neuroanatomy. AI technology has allowed researchers to detect brain differences at the individual level of ASD patients.

sauce: Boston University

Differences in behavior in people with autism spectrum disorders (ASD) are closely related to differences in neuroanatomy (brain shape). A team of neuroscientists at Boston University reports in a journal today. Chemistry..

This finding may help us understand the causes of ASD and develop individualized interventions.

The team uses artificial intelligence (AI) to study magnetic resonance imaging data from more than 1,000 individuals with ASD and compare those images with AI-generated simulations of what the brain looks like in the absence of ASD. Did.

Aidas Aglinskas, a post-doctor researcher at Boston University, said: And co-author of the report.

“Furthermore, separating ASD-related changes in the anatomy of the brain from irrelevant changes reveals a hidden relationship between the anatomy of the brain and individual differences in symptoms.”

Autism is individual, both symptomatic and neuroanatomical. Previous studies have hypothesized that there may not be a single set of neuroanatomical correlations common to all individuals with ASD.

According to Aglinskas, identifying ASD-specific neural changes has been a difficult task, making it difficult to confirm these suggestions. The brain depends on many factors, including genetic variation that is not due to ASD, which is difficult to control in research.

The team overcame that barrier by using AI to identify patterns of ASD-specific neural fluctuations. This allowed the team to identify the neural pathways that are particularly affected by ASD. Hartshorne and Stefano Anzellotti.

“ASD-related differences in the anatomy of the brain can be’hidden’among non-ASD-related differences,” Aglinskas said. “As a result, it was difficult to identify differences in the anatomy of the brain associated with different symptoms. AI was used to distinguish between differences related to ASD and those unrelated.”

The team set out to determine whether the ASD-specific functions of the anatomy of the brain differ between individuals in a symptom-related manner.

Previous studies investigating individual differences in the anatomy of the brain within the ASD did not separate ASD-specific features from other unrelated individual differences in neuroanatomy, thus studying the relationship between neuroanatomy and symptoms. It was difficult to do.

According to the report, the team used MRI data from 1,103 study participants and used an analysis method that was roughly similar to “deepfake.” ..

Instead, the team used computer-detected patterns to create a simulation of what each ASD individual’s brain would look like in the absence of ASD. The team reported that this was made possible by a new AI technology that separates individual differences in the anatomy of the brain into ASD-specific and non-ASD-related functions.

“Despite the large variations in brain anatomy between individuals with ASD along multiple dimensions, individuals were not grouped into distinct category subtypes as previously thought. I was particularly surprised, “said Aglinskas.

“At the level of brain anatomy, individual differences within the ASD may be better captured in continuous dimensions than category subtypes, but importantly, this is functional imaging, etc. It does not rule out the possibility of finding category subtypes in other types of brain measurements. “

In the future, researchers point out that we need to understand in more detail how these neuroanatomical differences affect behavior.

According to Anzerotti, the team will look for ways to use AI tools to better understand the structure of the brain, as well as the diagnosis of ASD and the behavior of individuals with ASD.

“The two brains can have very similar shapes, but they still work differently,” said Anzerotti.

See also

“There are many other aspects of the brain that we need to see to get the big picture. Now we are on a measure of functional connectivity, that is, how the brain is” wired “. I’m focusing. The big question is whether it shows us something new about individual differences within ASD. The goal of this type of work is to be able to use brain imaging data to assist in the development of personalized healthcare approaches for people with ASD. “

About this AI and ASD research news

author: Press office
sauce: Boston University
contact: Press Office-Boston University
image: The image is in the public domain

Independent research: Closed access.
“”Contrast machine learning reveals the structure of neuroanatomical changes within autismAidas Aglinskasetal. Chemistry

Overview

Contrast machine learning reveals the structure of neuroanatomical changes within autism

Autism Spectrum Disorders (ASD) are highly heterogeneous. Identifying systematic individual differences in neuroanatomy can inform diagnosis and personalized interventions.

The challenge is that these differences are intertwined with variability due to other causes: individual differences that are not related to ASD or measurement artifacts.

Contrast deep learning was used to unravel the ASD-specific neuroanatomical changes from those shared with typical control participants. ASD-specific variability correlated with individual differences in symptoms.

This structure of ASD-specific variations also corresponds to many years of debate about the nature of ASD. Individuals are not clustered into separate subtypes, at least from a neuroanatomical point of view. Instead, they are organized along a continuous dimension that affects different sets of regions.