Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Combining imaging and shape features for prediction tasks of Alzheimer's disease classification and brain age regression

Nairouz Shehata, Carolina Piçarra, Ben Glocker

TL;DR

This work tackles predicting brain age and Alzheimer's disease (AD) classification from structural MRI by fusing image-based features from ResNet-18 with shape features derived from brain surface meshes via a multi-graph network across 15 brain structures. The approach concatenates image and shape embeddings and passes them through an MLP, trained on datasets CamCAN, IXI, and OASIS-3 with skull stripping, bias correction, and mesh-based shape representations using FPFH descriptors. Evaluation shows that image–shape fusion yields substantial gains in AD classification (highest AUC of 0.861 with GCNConv and improved low-FPR performance) and modest improvements in brain age regression (best MAE ≈ 4.362, R² ≈ 0.893). These results underline the value of integrating geometric brain information with appearance features for clinically relevant neuroimaging tasks and motivate future work on larger datasets and more sophisticated fusion techniques.

Abstract

We investigate combining imaging and shape features extracted from MRI for the clinically relevant tasks of brain age prediction and Alzheimer's disease classification. Our proposed model fuses ResNet-extracted image embeddings with shape embeddings from a bespoke graph neural network. The shape embeddings are derived from surface meshes of 15 brain structures, capturing detailed geometric information. Combined with the appearance features from T1-weighted images, we observe improvements in the prediction performance on both tasks, with substantial gains for classification. We evaluate the model using public datasets, including CamCAN, IXI, and OASIS3, demonstrating the effectiveness of fusing imaging and shape features for brain analysis.

Combining imaging and shape features for prediction tasks of Alzheimer's disease classification and brain age regression

TL;DR

This work tackles predicting brain age and Alzheimer's disease (AD) classification from structural MRI by fusing image-based features from ResNet-18 with shape features derived from brain surface meshes via a multi-graph network across 15 brain structures. The approach concatenates image and shape embeddings and passes them through an MLP, trained on datasets CamCAN, IXI, and OASIS-3 with skull stripping, bias correction, and mesh-based shape representations using FPFH descriptors. Evaluation shows that image–shape fusion yields substantial gains in AD classification (highest AUC of 0.861 with GCNConv and improved low-FPR performance) and modest improvements in brain age regression (best MAE ≈ 4.362, R² ≈ 0.893). These results underline the value of integrating geometric brain information with appearance features for clinically relevant neuroimaging tasks and motivate future work on larger datasets and more sophisticated fusion techniques.

Abstract

We investigate combining imaging and shape features extracted from MRI for the clinically relevant tasks of brain age prediction and Alzheimer's disease classification. Our proposed model fuses ResNet-extracted image embeddings with shape embeddings from a bespoke graph neural network. The shape embeddings are derived from surface meshes of 15 brain structures, capturing detailed geometric information. Combined with the appearance features from T1-weighted images, we observe improvements in the prediction performance on both tasks, with substantial gains for classification. We evaluate the model using public datasets, including CamCAN, IXI, and OASIS3, demonstrating the effectiveness of fusing imaging and shape features for brain analysis.
Paper Structure (12 sections, 3 figures, 2 tables)

This paper contains 12 sections, 3 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Methods
  3. Imaging datasets
  4. Brain shape representation
  5. Feature extraction & fusion
  6. Experiments & Results
  7. Brain age regression
  8. Alzheimer's disease classification
  9. Discussion
  10. Conclusion
  11. Acknowledgments
  12. Compliance with ethical standards

Figures (3)

  • Figure 1: Fusion model overview, consisting of image-based and graph-based feature extractors, embeddings concatenated and passed through a multi-layer perceptron (MLP).
  • Figure 2: Scatter plots of true age against predicted age for different brain age prediction models.
  • Figure 3: ROC curves reveal substantial performance improvements for fusion models in the low FPR regime.