Trying to diagnose attention deficit hyperactivity disorder (ADHD) in accordance with neuroimaging data has long been a work in progress. According to a recent study published in Radiology, however, MRI may be able to more accurately diagnose the neurological disorder and differentiate between subtypes.
This study may also help in creating classification models for ADHD diagnosis and subtyping, according to researchers of the study from Sichuan University in China. According to the American Psychiatric Association, an estimated 5 to 8 percent of children and 4 percent of adults worldwide are diagnosed with some type of ADHD.
Researchers used MRI to identify all possible cerebral radiomics features related to ADHD diagnosis and subtyping, as well as examining neurodevelopmental relations that can discriminate between patients with and without ADHD or common ADHD subtypes.
"It is crucial to identify all the relevant features related to the disease during the classification process, because understanding the mechanisms behind the imaging phenotype is the aim," wrote lead author of the study Huaiqiang Sun, PhD, and colleagues.
According to Sun and his colleagues, previous imaging findings have no established diagnostic value for imaging patients.
"Studies typically report group-level differences between patients and control subjects," Sun said. "Analyzing brain imaging data under the framework of machine learning has the potential to address this challenge."
Researchers recruited 83 children newly diagnosed or never-treated with ADHD from the West China Hospital at Sichuan University and 87 children control subjects from local schools, all between the ages of 7 and 15 and maintained an equal gender ratio. Study participants were consecutively recruited from September 2009 to October 2015 and full consent was obtained by all participants parents, according to study methods.
All study participants underwent an anatomic and diffusion tensor MRI. Researchers then identified and extracted a total of 768 features representing the shape of gray matter and diffusion properties of white matter in the participants, additionally constructing and evaluating random forest classifiers in those with identified features, according to the study.
However, no notable difference was found between the participants with ADHD and control subjects in total brain volume or total gray and white matter volume, according to study results.
"The major finding of this study was that cerebral radiomics-based classification models provided discrimination of patients with ADHD from healthy control subjects, as well as separation of the two most common subtypes in a medication-naïve and relatively large sample-size single-imager MR imaging study," Sun said.
Additionally, a 73.7 percent mean classification accuracy was achieved when discriminating patients with ADHD and difference in shape in the left temporal lobe, bilateral cuneus and regions around the left central sulcus.
Due to its relation to cognitive and behavioral function, observed abnormalities in the left temporal lobe in participants with ADHD may be helpful in the diagnostic process. Sun and his colleagues also suggest that an observed thicker cuneus in participants with ADHD may point to developmental delays in the primary visual cortex. Overall, this study adds to the field of psychoradiology research according to Sun and, because of the automatic workflow used in the study, has the potential be useful in clinical settings.