Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

DiGAN: Diffusion-Guided Attention Network for Early Alzheimer's Disease Detection

Maxx Richard Rahman, Mostafa Hammouda, Wolfgang Maass

TL;DR

DiGAN tackles early Alzheimer's disease detection under limited, irregular longitudinal data by integrating latent diffusion-based trajectory synthesis with an attention-guided convolutional encoder. The diffusion component generates realistic neuroimaging trajectories to enrich temporal context, while the SAC-based attention-convolutional network extracts discriminative structural–temporal embeddings, followed by a max-pooling subject-level aggregation. Empirical results on synthetic and ADNI datasets show DiGAN outperforms baselines in NO vs. MCI and NO vs. SCD/AD tasks, with interpretable embedding maps that align with known neuroanatomical progression markers. This work offers a robust, data-efficient framework for early AD detection with practical potential for clinical deployment and risk stratification.

Abstract

Early diagnosis of Alzheimer's disease (AD) remains a major challenge due to the subtle and temporally irregular progression of structural brain changes in the prodromal stages. Existing deep learning approaches require large longitudinal datasets and often fail to model the temporal continuity and modality irregularities inherent in real-world clinical data. To address these limitations, we propose the Diffusion-Guided Attention Network (DiGAN), which integrates latent diffusion modelling with an attention-guided convolutional network. The diffusion model synthesizes realistic longitudinal neuroimaging trajectories from limited training data, enriching temporal context and improving robustness to unevenly spaced visits. The attention-convolutional layer then captures discriminative structural--temporal patterns that distinguish cognitively normal subjects from those with mild cognitive impairment and subjective cognitive decline. Experiments on synthetic and ADNI datasets demonstrate that DiGAN outperforms existing state-of-the-art baselines, showing its potential for early-stage AD detection.

DiGAN: Diffusion-Guided Attention Network for Early Alzheimer's Disease Detection

TL;DR

DiGAN tackles early Alzheimer's disease detection under limited, irregular longitudinal data by integrating latent diffusion-based trajectory synthesis with an attention-guided convolutional encoder. The diffusion component generates realistic neuroimaging trajectories to enrich temporal context, while the SAC-based attention-convolutional network extracts discriminative structural–temporal embeddings, followed by a max-pooling subject-level aggregation. Empirical results on synthetic and ADNI datasets show DiGAN outperforms baselines in NO vs. MCI and NO vs. SCD/AD tasks, with interpretable embedding maps that align with known neuroanatomical progression markers. This work offers a robust, data-efficient framework for early AD detection with practical potential for clinical deployment and risk stratification.

Abstract

Early diagnosis of Alzheimer's disease (AD) remains a major challenge due to the subtle and temporally irregular progression of structural brain changes in the prodromal stages. Existing deep learning approaches require large longitudinal datasets and often fail to model the temporal continuity and modality irregularities inherent in real-world clinical data. To address these limitations, we propose the Diffusion-Guided Attention Network (DiGAN), which integrates latent diffusion modelling with an attention-guided convolutional network. The diffusion model synthesizes realistic longitudinal neuroimaging trajectories from limited training data, enriching temporal context and improving robustness to unevenly spaced visits. The attention-convolutional layer then captures discriminative structural--temporal patterns that distinguish cognitively normal subjects from those with mild cognitive impairment and subjective cognitive decline. Experiments on synthetic and ADNI datasets demonstrate that DiGAN outperforms existing state-of-the-art baselines, showing its potential for early-stage AD detection.
Paper Structure (25 sections, 7 figures, 2 tables)

This paper contains 25 sections, 7 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Preliminaries
  3. DiGAN Architecture
  4. Diffusion-based Generative Network
  5. Attention-Convolutional Network
  6. 1. Self-Attention Layer:
  7. 2. Convolutional Layer:
  8. 3. Batch Normalization and Output:
  9. Subject-Level Aggregation
  10. Model Training and Optimization
  11. Experiments
  12. Datasets
  13. Synthetic dataset.
  14. ADNI. petersen2010adni
  15. Experimental Settings
  16. ...and 10 more sections

Figures (7)

  • Figure 1: Overview of the DiGAN architecture comprising (i) a diffusion process for synthesizing neuroimaging profiles, (ii) an attention-convolutional network for extracting structural–temporal embeddings, and (iii) a subject-level aggregation for AD identification.
  • Figure 2: Comparison between real and synthetic ADNI data distributions of different parameters.
  • Figure 3: Performance of DiGAN for NO vs. MCI across profile lengths of 2, 3, and 4 visits.
  • Figure 4: Performance of DiGAN for NO vs. SCD (synthetic) and NO vs. AD (ADNI) across profile lengths of 2, 3, and 4 visits.
  • Figure 5: The differential correlation heatmap (left) and a PCA projection (right), showing substantial overlap between synthetic and real profiles.
  • ...and 2 more figures