Morphometric Inverse Divergence Networks Combined with HYDRA Identify Parkinson's Disease Subtypes with Distinct...

Area of Science:

• Neuroscience
• Biomarkers
• Neurodegenerative Diseases

Background:

• Parkinson's disease (PD) is a common neurodegenerative disorder with unclear subtypes.
• Current understanding of PD lacks distinct molecular and serum biomarker signatures for patient stratification.

Purpose of the Study:

• To investigate if cortical architecture can identify biologically distinct Parkinson's disease (PD) subtypes.
• To link these subtypes to molecular signatures and serum biomarkers (NfL, GFAP).

Main Methods:

• Structural MRI was used to construct subject-specific cortical similarity networks (MIND).
• HYDRA clustering identified PD subtypes based on network differences.
• Spatial patterns were analyzed for gene expression, functional enrichment, and serum biomarkers (NfL, GFAP).

Main Results:

• Two PD subtypes with distinct cortical organization were identified, only partially overlapping with motor phenotypes.
• Cluster 1 showed temporo-parietal network increases linked to synaptic/oligodendroglial pathways, with no serum biomarker correlation.
• Cluster 2 exhibited fronto-cingulate network reductions linked to mitochondrial/lysosomal/proteostatic pathways, correlating with elevated serum NfL and GFAP.

Conclusions:

• Biologically distinct Parkinson's disease (PD) subtypes exist, characterized by divergent cortical organization and molecular signatures.
• These subtypes demonstrate distinct patterns of systemic neurodegeneration, offering a novel approach beyond conventional motor phenotyping.