Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

GKAN: Explainable Diagnosis of Alzheimer's Disease Using Graph Neural Network with Kolmogorov-Arnold Networks

Tianqi Ding, Dawei Xiang, Keith E Schubert, Liang Dong

TL;DR

This work targets Alzheimer's disease diagnosis using a single-modal structural MRI graph model. By integrating Kolmogorov-Arnold Networks with Graph Convolutional Networks, the approach introduces spline-based nonlinearities to better capture brain-region interactions. On the ADNI subset, GCN-KAN yields a 4–8% accuracy improvement over traditional GCNs and provides interpretable ROI-level insights, notably highlighting the hippocampus, parietal gyrus, and amygdala as key regions. The method offers an explainable and more accurate tool for early AD diagnosis, with future extensions to multi-modal data and larger cohorts to enhance robustness and clinical applicability.

Abstract

Alzheimer's Disease (AD) is a progressive neurodegenerative disorder that poses significant diagnostic challenges due to its complex etiology. Graph Convolutional Networks (GCNs) have shown promise in modeling brain connectivity for AD diagnosis, yet their reliance on linear transformations limits their ability to capture intricate nonlinear patterns in neuroimaging data. To address this, we propose GCN-KAN, a novel single-modal framework that integrates Kolmogorov-Arnold Networks (KAN) into GCNs to enhance both diagnostic accuracy and interpretability. Leveraging structural MRI data, our model employs learnable spline-based transformations to better represent brain region interactions. Evaluated on the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset, GCN-KAN outperforms traditional GCNs by 4-8% in classification accuracy while providing interpretable insights into key brain regions associated with AD. This approach offers a robust and explainable tool for early AD diagnosis.

GKAN: Explainable Diagnosis of Alzheimer's Disease Using Graph Neural Network with Kolmogorov-Arnold Networks

TL;DR

This work targets Alzheimer's disease diagnosis using a single-modal structural MRI graph model. By integrating Kolmogorov-Arnold Networks with Graph Convolutional Networks, the approach introduces spline-based nonlinearities to better capture brain-region interactions. On the ADNI subset, GCN-KAN yields a 4–8% accuracy improvement over traditional GCNs and provides interpretable ROI-level insights, notably highlighting the hippocampus, parietal gyrus, and amygdala as key regions. The method offers an explainable and more accurate tool for early AD diagnosis, with future extensions to multi-modal data and larger cohorts to enhance robustness and clinical applicability.

Abstract

Alzheimer's Disease (AD) is a progressive neurodegenerative disorder that poses significant diagnostic challenges due to its complex etiology. Graph Convolutional Networks (GCNs) have shown promise in modeling brain connectivity for AD diagnosis, yet their reliance on linear transformations limits their ability to capture intricate nonlinear patterns in neuroimaging data. To address this, we propose GCN-KAN, a novel single-modal framework that integrates Kolmogorov-Arnold Networks (KAN) into GCNs to enhance both diagnostic accuracy and interpretability. Leveraging structural MRI data, our model employs learnable spline-based transformations to better represent brain region interactions. Evaluated on the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset, GCN-KAN outperforms traditional GCNs by 4-8% in classification accuracy while providing interpretable insights into key brain regions associated with AD. This approach offers a robust and explainable tool for early AD diagnosis.

Paper Structure

This paper contains 19 sections, 10 equations, 4 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related Work
  3. Graph Neural Networks for Neuroimaging
  4. Kolmogorov-Arnold Networks (KAN)
  5. Methodology
  6. Graph Construction
  7. GCN-KAN Architecture
  8. Training and Interpretability
  9. Experimental Setup
  10. Dataset
  11. Implementation Details
  12. Accuracy Calculation
  13. AUC-ROC Calculation
  14. F1-Score Calculation
  15. Results and Analysis
  16. ...and 4 more sections

Figures (4)

  • Figure 1: Overview of the GCN-KAN architecture, showing GCN and KAN layer integration.
  • Figure 2: Training and validation loss curves for the GCN model.
  • Figure 3: Training and validation loss curves for the GCN-KAN model.
  • Figure 4: Visualization of salient ROIs identified by GCN-KAN. The color intensity reflects the relative importance of each region, with brighter colors indicating higher importance.