In a recent study that was presented at the annual meeting of the Radiological Society of North America, researchers examined brain scans of teenagers with and without attention deficit hyperactivity disorder (ADHD) using artificial intelligence.

This method revealed variations in the white matter tracts of the brains of ADHD sufferers, offering new perspectives on the illness.

Approximately 6 million children and teenagers in the US suffer from ADHD, thus early diagnosis and treatment are essential for bettering well-being in a culture where distractions are a growing factor.

Symptoms of ADHD

The symptoms of attention deficit hyperactivity disorder (ADHD) can include trouble focusing, controlling energy, and impulse control.

It usually manifests in childhood and has a substantial impact on a person’s health and capacity to engage with society.

Approximately 6 million kids and teens in the US between the ages of 6 and 17 have been diagnosed with ADHD.

According to experts, diagnosing ADHD can be difficult as doctors frequently rely on subjective self-reported questionnaires. They claim that there is a definite need for more impartial diagnostic techniques.

Researchers reported on using a deep learning form of artificial intelligence (AI) to analyze MRI scans of teenagers with and without ADHD at the Radiological Society of North America’s annual meeting in November.

The white matter tracts, which are specific brain regions, showed significant changes in individuals with ADHD, according to the researchers.

AI deep learning for identifying signs of ADHD

The study, which hasn’t yet been released in a peer-reviewed journal, according to the researchers, is significant since it’s the first to employ deep learning to find signs of ADHD.

Artificial intelligence (AI) that uses deep learning is able to automatically identify links and patterns in enormous volumes of data.

Medical News Today was informed by Justin Huynh, MS, co-author of the study, researcher in the Department of Neuroradiology at the University of California San Francisco, and medical student at the Carle Illinois College of Medicine at Urbana-Champaign, Illinois, that they examined a sizable dataset of adolescent brain images of ADHD and non-ADHD patients.

“We discovered that, generally, there were statistically significant variations in imaging between research subjects who had and did not have [attention deficit],” he stated.

Our results should be seen as a positive first step toward a more uniform, impartial, and precise method for the treatment of ADHD as well as a deeper biological understanding of the disorder.

MS Justin Huynh

Analysis of clinical surveys and MRI scan data

Brain scan data, clinical survey data, and other information collected from 21 US research locations were used in the study.

Diffusion-weighted imaging (DWI), a specialized form of magnetic resonance imaging (MRI), was one of the methods employed to get the brain imaging data.

The researchers noted that the complex nature of ADHD and small sample numbers have made previous attempts to use AI for detection of the illness difficult.

The research team intentionally selected 1,704 participants for this study, including both ADHD-positive and non-ADHD-positive teenagers.

They extracted fractional anisotropy (FA) data along 30 major white matter routes in the brain using DWI images. FA quantifies the flow of water molecules across these tracts’ fibers.

A deep-learning AI model was trained using the FA values of 1,371 individuals. The model was then tested on 333 participants, 193 of whom had been diagnosed with ADHD and 140 of whom had not.

The researchers claimed to have made a noteworthy finding using AI. They found that FA values in nine white matter tracts were significantly greater in people with ADHD.

These unique MRI patterns in individuals with ADHD have never been seen before in such fine detail.

Generally speaking, the abnormalities found in these white matter tracts matched the symptoms of ADHD.

“I agree with [the researchers’] basic framing of ADHD as a complex disorder with potential structural and functional variations underpinning psychopathology,” said Dr. David Lefkowitz, a neuroradiology specialist and medical director of MRIs at SimonMed Imaging, who was not involved in this study, in an interview with Medical News Today. This represents a breakthrough in the study on ADHD brain scans.

“Attempts to find structural correlations revealed by MRI to diagnose ADHD have been largely unsuccessful historically, despite significant effort,” Lefkowitz added.

He clarified, “But they might still exist, and the researchers here are using the best tools available to find such correlations by combining deep learning and DTI.”

If we look closely enough, we might find structural anomalies in ADHD, but this is not the most promising avenue for further research. Ultimately, ADHD is a disorder of behavior. It would seem more reasonable that an imaging method that evaluates function rather than structure would be more promising. Therefore, my bias would be to research brain metabolism (PET) or functional networks (fMRI). Still, I believe it’s critical to maintain an open mind.

David Lefkowitz, M.D.

Lefkowitz stated, “My skepticism should not be viewed as dismissive. Discoveries happen in unexpected places.” “I am eager to follow this research, particularly as it progresses towards a peer-reviewed publication.”

Technology advancements could increase the accuracy of ADHD diagnosis

“This research represents a significant advancement in the application of AI and imaging data analysis to the field of ADHD diagnosis,” said Livia Lifes, the chief executive officer of Neuroute and an expert in artificial intelligence who was not involved in the study, to Medical News Today.

Autoencoders, an unsupervised deep learning methodology, might potentially reveal subtle structural patterns that conventional diagnostic tools could overlook. This has the potential to significantly improve the diagnosis accuracy of ADHD and offer insightful information about the underlying neurobiology of the condition.

“An accurate non-invasive imaging technique for ADHD patients could be very helpful in clinical management and also in drug trials,” Livia Lifes Lefkowitz concurred.

“Patient selection is one of the challenges with proving drug efficacy,” he stated. “Clinical drug trials are very expensive, in part because many patients must participate in order to obtain results that are statistically significant.”“More precise diagnosis of ADHD and, additionally, the capacity to group patients according to severity may lessen the necessary size and, consequently, expense of such trials,” he continued.“The implications are beyond the patient population (which obviously would benefit), but to society more broadly,” Lefkowitz concluded.