Transformer-Based Multi-Modal Temporal Embeddings for Explainable Metabolic Phenotyping in Type 1 Diabetes

2026-01-07

arXiv.org

AI Summary

Type 1 diabetes exhibits substantial metabolic heterogeneity that cannot be fully captured by conventional metrics such as HbA1c. This study proposes the first interpretable, multimodal temporal embedding framework that integrates continuous glucose monitoring with laboratory data. By leveraging a Transformer encoder to model temporal dependencies and a Gaussian mixture model to identify latent metabolic phenotypes, the approach enables robust subtyping while incorporating attention mechanisms and SHAP values for transparent feature attribution. Applied to a cohort of 577 patients, the method reveals five physiologically coherent metabolic subgroups spanning a spectrum from stable profiles to those conferring high cardiovascular risk. These subgroups show significant associations with hypertension, myocardial infarction, and heart failure, with key drivers including HbA1c, triglycerides, cholesterol, creatinine, and thyroid-stimulating hormone (TSH).

Technology Category

• Machine Learning: Transparent, Interpretable, Explainable ML
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Data Mining & Knowledge Management: Mining of Spatial, Temporal or Spatio-Temporal Data

Application Category

• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Security and Privacy: Data transparency and provenance

Abstract

Type 1 diabetes (T1D) is a highly metabolically heterogeneous disease that cannot be adequately characterized by conventional biomarkers such as glycated hemoglobin (HbA1c). This study proposes an explainable deep learning framework that integrates continuous glucose monitoring (CGM) data with laboratory profiles to learn multimodal temporal embeddings of individual metabolic status. Temporal dependencies across modalities are modeled using a transformer encoder, while latent metabolic phenotypes are identified via Gaussian mixture modeling. Model interpretability is achieved through transformer attention visualization and SHAP-based feature attribution. Five latent metabolic phenotypes, ranging from metabolic stability to elevated cardiometabolic risk, were identified among 577 individuals with T1D. These phenotypes exhibit distinct biochemical profiles, including differences in glycemic control, lipid metabolism, renal markers, and thyrotropin (TSH) levels. Attention analysis highlights glucose variability as a dominant temporal factor, while SHAP analysis identifies HbA1c, triglycerides, cholesterol, creatinine, and TSH as key contributors to phenotype differentiation. Phenotype membership shows statistically significant, albeit modest, associations with hypertension, myocardial infarction, and heart failure. Overall, this explainable multimodal temporal embedding framework reveals physiologically coherent metabolic subgroups in T1D and supports risk stratification beyond single biomarkers.

Problem

• Type 1 diabetes
• metabolic heterogeneity
• multimodal data
• metabolic phenotyping
• temporal embedding

Innovation

• Transformer
• multimodal temporal embedding
• explainable AI
• metabolic phenotyping
• Gaussian mixture modeling