Radiomics neurobiomarkers for distinguishing mathematical-related areas in children with elevated ADHD symptoms

Abstract

Children with ADHD symptoms are at higher risk of experiencing mathematical difficulties. Individuals with ADHD often exhibit lower mathematical accuracy and rely on more basic strategies, even without comorbid learning disorders. In this study, we aim to identify whether this higher risk is associated with general (e.g. executive function) or specific (e.g. calculation) mathematical abilities by analyzing brain areas in structural imaging with radiomics. We used a public brain imaging dataset of mathematical development in children to define two subgroups: one with ADHD symptoms and a control group, matched for age mean and variance. Then, we explore the differences in brain regions related to mathematical performance between these groups using radiomics features in structural imaging. Our results confirm a lower mathematical performance in the group with ADHD symptoms in the behavioral analysis. Structurally, we found differences between the groups in key brain regions associated with attention and memory, including the inferior and middle frontal cortex bilaterally, the right posterior cingulate cortex, and the left hippocampus. Features related to shape demonstrated the highest predictive power. We hypothesize that children with ADHD symptoms might have difficulties with 'number attention' and propose future studies in this area.