Visual Explanations From Deep 3D Convolutional Neural Networks for Alzheimer's Disease Classification

TL;DR

This paper introduces three methods for visual explanations of 3D-CNNs used in Alzheimer's disease classification, enhancing interpretability by identifying key brain regions and analyzing network activations.

Contribution

It presents novel approaches for visual explanation of 3D-CNNs, combining sensitivity analysis and activation visualization to improve understanding of model decisions in Alzheimer's diagnosis.

Findings

All methods identify important brain regions for Alzheimer's.

Sensitivity analysis struggles with loosely distributed cortex.

Activation visualization is limited by convolutional layer resolution.

Abstract

We develop three efficient approaches for generating visual explanations from 3D convolutional neural networks (3D-CNNs) for Alzheimer's disease classification. One approach conducts sensitivity analysis on hierarchical 3D image segmentation, and the other two visualize network activations on a spatial map. Visual checks and a quantitative localization benchmark indicate that all approaches identify important brain parts for Alzheimer's disease diagnosis. Comparative analysis show that the sensitivity analysis based approach has difficulty handling loosely distributed cerebral cortex, and approaches based on visualization of activations are constrained by the resolution of the convolutional layer. The complementarity of these methods improves the understanding of 3D-CNNs in Alzheimer's disease classification from different perspectives.