HypTI: Hyperbolic Geometry-Aware Multimodal Feature Fusion Framework for Dementia Detection Based on Tabular-Image Data

Abstract

AI-assisted dementia detection usually requires multimodal fusion of diverse data, such as clinical tabular assessments and neuroimaging data. However, most existing approaches perform feature fusion in Euclidean space. This overlooks the geometric structure of representation, including hierarchical organization in clinical assessments and multi-scale anatomical patterns in neuroimaging data. Recent studies have proved hyperbolic space provides a natural geometry for modelling hierarchical and multi-scale structures due to its exponentially expanding space. Motivated by these, we propose HypTI, a hyperbolic geometry-aware multimodal fusion framework for dementia detection based on clinical tabular assessments and MRI scans. HypTI extends hyperbolic representation learning to an underexplored tabular-MRI setting, jointly modelling clinical tabular, MRI scans, and cross-modal interactions in hyperbolic space. Specifically, clinical tabular variables are represented through learnable Lorentz prototypes to capture hierarchical relationships among assessments, while MRI scans are encoded using ROI-guided multi-scale representation learning to capture anatomical patterns across brain regions and spatial scales in Lorentz space. The multimodal representations are integrated through an adaptive fusion module generates subject-specific scale and shift coefficients to dynamically modulate and integrate the multimodal embeddings, enabling effective patient-level tabular-image representation learning for dementia detection. Experiments on OASIS-3 dataset show that the proposed framework achieves strong performance compared with representative approaches.