Unlocking Neural Transparency: Jacobian Maps for Explainable Ai in Alzheimer’s Detection

Abstract

Alzheimer's disease (AD) causes progressive cognitive decline, where early detection is critical for effective intervention. While deep learning models have achieved high detection accuracy in AD diagnosis, their lack of interpretability has led to skepticism among medical professionals. This paper introduces a novel pre-model approach leveraging Jacobian Maps (JM) within a multi-modal framework to improve interpretability and trustworthiness in AD detection. By capturing localized brain volume changes, JMs enhance model explainability by correlating predictions with established neuroanatomical biomarkers of AD. We validate the effectiveness of JMs through experiments comparing the performance of a 3D CNN trained on JMs versus traditional preprocessed data, which demonstrates superior accuracy. Additionally, we provide both visual and quantitative insights using 3D Grad-CAM analysis, demonstrating improved interpretability and diagnostic accuracy.

Department(s)

Computer Science

Second Department

Electrical and Computer Engineering

Keywords and Phrases

Alzheimer's Disease (AD); Explainable AI (XAI); Jacobian Maps; Medical Image Analysis; Multi-Modal Data

International Standard Book Number (ISBN)

978-981968196-9

International Standard Serial Number (ISSN)

1611-3349; 0302-9743

Document Type

Article - Conference proceedings

Document Version

Citation

File Type

text

Language(s)

English

Rights

© 2025 Springer, All rights reserved.

Publication Date

01 Jan 2025

Link to Full Text

Share

 
COinS