Summary: Alzheimer’s disease advances unevenly across the brain. New mathematical modeling developed by researchers at The University of Texas at Arlington and collaborators at UCSF helps explain why certain regions deteriorate faster while others resist damage. Their extended network diffusion model simulates how tau protein accumulates and spreads along brain connections and identifies genes that increase or decrease regional vulnerability.
The model shows that regions more tightly connected to tau-affected areas are likelier to incur damage, while more isolated regions tend to remain resilient. This network-based perspective creates a powerful framework for understanding disease progression and may inform targeted therapeutic strategies.
Key Facts:
- Network vulnerability: Brain regions that are more connected to tau-laden areas deteriorate sooner and more extensively.
- Gene classification: The model sorts Alzheimer’s risk genes into categories based on whether their effects align with network spread or act independently.
- Human data: The study uses PET data from human participants to model tau progression, increasing clinical relevance compared with many animal-only studies.
Source: UT Arlington
Mathematics is providing a fresh lens on Alzheimer’s disease. Pedro Maia, assistant professor of mathematics and data science at The University of Texas at Arlington, and colleagues at UCSF’s Raj Lab applied mathematical modeling to map how tau pathology propagates across the brain’s connectome.
Their published work in the journal Brain introduces an extended network diffusion model (eNDM) that captures both the network-mediated transmission of tau protein and the region-specific, gene-driven predisposition to pathology. The model can be fit to individual or group tau PET data to produce a reference pattern of network-based tau spread.
Using the model, the team examined how baseline gene expression across brain regions relates to observed tau accumulation. By comparing actual tau PET measurements to the model’s network-predicted pattern, they derived residual maps (observed minus model-predicted tau) and associated those residuals with spatial gene expression profiles. This strategy separates network-aligned vulnerability from network-independent effects.
From these analyses, researchers identified four functional classes of Alzheimer’s risk genes:
- Network-aligned susceptibility (SV-NA): Genes whose expression aligns with network-driven patterns of vulnerability and are associated with increased tau.
- Network-independent susceptibility (SV-NI): Genes that increase vulnerability through mechanisms not explained by network spread.
- Network-aligned resilience (SR-NA): Genes that protect regions in ways that correlate with the network pattern.
- Network-independent resilience (SR-NI): Genes that confer protection through local or cell-autonomous mechanisms.
Functional enrichment analyses revealed distinct biological roles for these classes. Network-aligned genes tend to participate in pathways such as cell death regulation, stress responses and metabolic processing. In contrast, network-independent genes are enriched in amyloid-β processing and immune-response pathways. These segregated roles suggest multiple routes by which genetic risk factors influence whether a region succumbs to or resists tauopathy.
The study fitted the eNDM to tau PET scans from 196 participants drawn from a research cohort. The cohort included individuals with early-stage mild cognitive impairment (102 participants), late-stage mild cognitive impairment (47 participants) and Alzheimer’s disease (47 participants). By grounding modeling in human imaging data, the research aims to improve relevance to clinical disease and therapy development.
Maia emphasized that while animal models remain valuable for controlled experiments, human data are essential to understand how Alzheimer’s unfolds in people. “If we want to develop treatments that work in humans, we need data that comes from humans,” he said.
The research highlights how the brain’s nonuniform structure—different cell types, regional gene expression and connectivity—shapes where and how pathology develops. Regions that are highly connected or lie close to affected hubs are more prone to tau accumulation and downstream damage, whereas more isolated or differently wired regions can show resilience.
Beyond advancing basic understanding, this modeling approach may help prioritize therapeutic targets by distinguishing genes and pathways that modulate disease via network spread versus those acting locally. Such distinctions could guide interventions aimed either at slowing transneuronal propagation or strengthening local resilience mechanisms.
About this math modeling and Alzheimer’s disease research news
Author: Drew Davison
Source: UT Arlington
Contact: Drew Davison, UT Arlington
Image: Image credit: Neuroscience News
Original research: Open access. Title: “Selective vulnerability and resilience to Alzheimer’s disease tauopathy as a function of genes and the connectome” by Pedro Maia et al., published in Brain.
Abstract (condensed)
Alzheimer’s disease shows marked regional differences in vulnerability: structures such as the entorhinal cortex and hippocampus develop tau tangles early, while primary sensory cortices often remain relatively spared. To explain selective vulnerability and resilience, this study models both regional genetic factors and network-mediated propagation of misfolded tau. Using an extended network diffusion model fitted to human tau PET data, the authors extracted model residuals to test associations with spatial expression profiles of known Alzheimer’s risk genes from brain transcriptomic atlases.
The eNDM captured the broad spatial distribution of tau pathology. After accounting for the network-predicted component, some risk genes aligned with the network pattern while others correlated more strongly with residual, network-independent tau. This led to the classification of four gene categories (SV-NA, SV-NI, SR-NA, SR-NI), each with distinct spatial and functional signatures. Gene ontology analysis revealed that network-aligned genes often relate to cell death, stress and metabolism, whereas network-independent genes associate with amyloid processing and immune functions. The findings illuminate multiple pathways by which genetic factors confer vulnerability or resilience and may help identify new intervention targets.