Differentiating Functional Connectivity Patterns in ADHD and Autism Among the Young People: A Machine Learning...

Area of Science:

• Neuroscience
• Developmental Psychology
• Computational Psychiatry

Background:

• Attention-deficit/hyperactivity disorder (ADHD) and autism are common neurodevelopmental conditions with overlapping symptoms and potential shared origins.
• Differentiating between ADHD and autism, especially in young people, is challenging due to their complex nature and shared etiological factors.

Purpose of the Study:

• To differentiate between ADHD and autism in young individuals based on intrinsic brain connectivity patterns during resting state.
• To assess the efficacy of machine learning in distinguishing ADHD from autism using functional magnetic resonance imaging (fMRI) data.
• To identify specific brain networks that are differentially involved in ADHD and autism.

Main Methods:

• Analysis of resting-state fMRI data from the Autism Brain Imaging Data Exchange (ABIDE) and ADHD-200 Consortium.
• Selection of 330 participants (110 each for ADHD, autism, and healthy controls), excluding comorbidities.
• Application of linear discriminant analysis to region-to-region connectivity values to identify discriminative patterns.

Main Results:

• Machine learning models achieved an 85% accuracy in differentiating between ADHD and autism based on brain connectivity.
• Altered connectivity in the frontoparietal network was a key differentiator for ADHD compared to autism and controls.
• Autism diagnosis was associated with more heterogeneous network alterations, including language, salience, and frontoparietal networks.

Conclusions:

• The study reveals distinct neural connectivity signatures for ADHD and autism.
• High discriminability between ADHD and autism using brain-based metrics supports their potential role in differential diagnostics.
• Findings enhance the understanding of the neurobiological underpinnings differentiating these two conditions.