Interpretable discriminant analysis for functional data supported on random nonlinear domains with an application to Alzheimer's disease
These videos have been matched automatically. Contact us if they are not relevant.
These videos have been matched automatically. Contact us if they are not relevant.
View abstract on PubMed
Summary
This summary is machine-generated.This study presents a new method for classifying Alzheimer's disease using brain scans. The approach analyzes cortical surface geometry and thickness to identify disease markers effectively.
Area Of Science
- Neuroscience
- Medical Imaging
- Statistics
Background
- Alzheimer's disease (AD) diagnosis relies on identifying subtle changes in brain structure.
- Functional data on complex manifold domains, like cortical surfaces, pose analytical challenges.
- Existing methods for analyzing such data are often computationally intensive.
Purpose Of The Study
- To develop a novel framework for classifying functional data on nonlinear manifold domains.
- To apply this framework to identify Alzheimer's disease subjects using cortical surface geometry and thickness.
- To establish a computationally efficient method that avoids prior covariance structure estimation.
Main Methods
- Reformulation of classification as a regularized multivariate functional linear regression model.
- Direct estimation of discriminant directions with differential regularization.
- Theoretical analysis of out-of-sample prediction error and finite sample performance evaluation.
Main Results
- The proposed method effectively classifies subjects based on functional data on manifold domains.
- Identified discriminant directions capture both geometric and thickness features predictive of Alzheimer's disease.
- The approach demonstrates computational feasibility by bypassing covariance structure estimation.
Conclusions
- The novel framework offers an effective and computationally efficient approach for classifying functional data on manifold domains.
- The method successfully identifies key neuroimaging features associated with Alzheimer's disease.
- Findings align with existing neuroscience literature, validating the approach's clinical relevance.
These videos have been matched automatically. Contact us if they are not relevant.
These videos have been matched automatically. Contact us if they are not relevant.
