Diagnostic classification of autism using resting-state fMRI data improves with full correlation functional brain...

Area of Science:

• Neuroscience
• Medical Imaging
• Machine Learning

Background:

• Autism Spectrum Disorder (ASD) is a neurodevelopmental condition characterized by altered brain network connectivity.
• Understanding these neural differences is crucial for diagnosis and intervention.

Purpose of the Study:

• To investigate and differentiate functional connectivity (FC) patterns in the brains of individuals with ASD and typical development (TD).
• To develop and evaluate machine learning models for classifying ASD based on brain connectivity features.

Main Methods:

• Analysis of Resting-state functional Magnetic Resonance Imaging (Rs-fMRI) data from ASD and TD participants.
• Application of correlation methods (GLASSO, MDMC, PCCE) to assess functional connectivity across 238 brain regions.
• Development of classifier models including Random Forest (RF), Support Vector Machine (SVM), and Convolutional Neural Network (CNN) using feature selection techniques.

Main Results:

• The PCCE-CNN pipeline achieved the highest average test accuracy (70.31%) and an Area Under the Curve (AUC) of 0.73.
• Key differentiating FC features were identified within the Dorsal Attention (DA) and Cingulo-Opercular Task Control (COTC) networks.
• Specific connections between COTC and DA networks were significant in discriminating between ASD and TD groups.

Conclusions:

• Machine learning models, particularly PCCE-CNN, demonstrate significant potential for classifying ASD based on brain functional connectivity.
• The findings highlight specific network alterations in ASD, offering insights into the neurobiological underpinnings of the disorder.
• The developed models show improved performance compared to previous studies, suggesting robustness in classifying heterogeneous ASD populations.