Anatomy-Guided Surface Diffusion Model for Alzheimer’s Disease Normative Modeling

Normative modeling has emerged as a pivotal approach for characterizing heterogeneityand individual variance in neurodegenerative diseases, notably Alzheimer’s disease (AD).One of the challenges of cortical normative modeling is the anatomical structure mismatchdue to folding pattern variability. Traditionally, registration is applied to address this issueand recently deep generative models are employed to generate anatomically aligned sam-ples for analyzing disease progression; however, these models are predominantly appliedto volume-based data, which often falls short in capturing intricate morphological changeson the brain cortex. As an alternative, surface-based analysis has been proven to be moresensitive in disease modeling such as AD. Yet, like volume-based data, it also suffers fromthe mismatch problem. To address these limitations, we propose a novel generative nor-mative modeling framework by transferring the conditional diffusion generative model tothe spherical domain. Furthermore, the proposed model generates normal feature mapdistributions by explicitly conditioning on individual anatomical segmentation to ensurebetter geometrical alignment which helps to reduce variance between subjects in norma-tive analysis. We find that our model can generate samples that are better anatomicallyaligned than registered reference data and through ablation study and normative assess-ment experiments, the samples are able to better measure individual differences from thenormal distribution and increase sensitivity in differentiating cognitively normal (CN), mildcognitive impairment (MCI), and Alzheimer’s disease (AD) patients.