Age-Stratified Differences in Morphological Connectivity Patterns in ASD: An sMRI and Machine Learning Approach

Area of Science:

• Neuroscience
• Medical Imaging
• Developmental Disorders

Background:

• Autism spectrum disorder (ASD) is a common neurological disorder with challenging early detection.
• Current diagnostic approaches rely on various physiological and medical imaging signals, often considering patient age and symptom severity.

Purpose of the Study:

• To investigate the influence of age on the diagnosis of autism spectrum disorders (ASD).
• To evaluate the effectiveness of morphological features (MF) and morphological connectivity features (MCF) in age-specific ASD diagnosis.

Main Methods:

• Utilized structural magnetic resonance imaging (sMRI) data from the ABIDE-I and ABIDE-II databases.
• Analyzed data from three age groups: 6-11, 11-18, and 6-18 years.
• Employed a random forest (RF) classifier on preprocessed MF (592 per subject) and MCF (10,878 per subject).

Main Results:

• The 6-11 age group demonstrated superior performance in ASD diagnosis compared to other age groups for both MF and MCF.
• Morphological connectivity features (MCF) showed particularly strong diagnostic potential in the youngest age group.
• The RF classifier achieved 75.8% accuracy, 83.1% F1 score, 86% recall, and 80.4% precision in the 6-11 age group.

Conclusions:

• Age-specific analysis, particularly focusing on morphological connectivity, is a promising avenue for enhancing ASD diagnosis.
• The findings suggest that distinct developmental stages may influence the manifestation of brain morphology relevant to ASD.