Integrated multimodal network approach to PET and MRI based on multidimensional persistent homology

Area of Science:

• Neuroscience
• Medical Imaging
• Computational Topology

Background:

• Analyzing multimodal brain imaging data presents challenges in understanding integrated network relationships.
• Topological data analysis offers novel methods for characterizing complex network structures.

Purpose of the Study:

• To develop and apply a novel multidimensional persistence approach for analyzing integrated brain networks from PET and MRI data.
• To investigate topological differences in brain connectivity associated with Autism Spectrum Disorder (ASD) and Attention Deficit Hyperactivity Disorder (ADHD).

Main Methods:

• Utilized a threshold-free persistent homology method extended to multidimensional persistence.
• Implemented a bimodal integration technique using 1D projection to combine PET and MRI connectivity data.
• Varied thresholds and mixing ratios between modalities to visualize topological changes in integrated brain networks.

Main Results:

• Brain networks in ASD and ADHD children exhibit distinct topological features compared to control subjects.
• ASD showed differences primarily in metabolic connectivity, while ADHD displayed maximum network divergence at a 0.6 mixing ratio of PET and MRI data.
• Asymmetrical changes in connected structures between metabolic and morphological connectivities were observed in ASD and ADHD.

Conclusions:

• Multidimensional persistence provides a robust framework for understanding disease-related brain networks from multimodal imaging.
• The study highlights unique network alterations in ASD and ADHD, offering insights into their underlying neurobiology.
• The findings underscore the potential of integrating PET and MRI data for improved diagnostic and research applications in pediatric neurological disorders.