Keywords: Alzheimer's Disease, Disease Progression Modeling, Deep Generative Model
TL;DR: We propose a novel interpretable framework named Prototypical Brain Maps (ProtoBrainMaps) for modeling Alzheimer's disease progressions through a set of prototypes in the latent space via clinically-guided topological maps.
Abstract: Discovering the brain progression over a lifetime is beneficial for identifying the subject affected by neurodegenerative disorders, such as Alzheimer's disease (AD) which require detection at the earliest possible time for the sake of delaying the progression by the virtue of particular treatments. As brain progressions in terms of both normal aging and AD-pathology tend to be entangled to each other, distinguishing the progression pathways of AD over the normal aging brains is quite an intricate task. To this end, we propose Prototypical Brain Maps (ProtoBrainMaps) for modeling the AD progressions through the established prototypes in the latent space via clinically-guided topological maps. Having devised as an interpretable network, it possesses the ability to establish and synthesize a set of well-interpolated prototypical brains, each corresponding to certain health conditions in terms of neurodegenerative diseases.
Paper Type: methodological development
Primary Subject Area: Interpretability and Explainable AI
Secondary Subject Area: Integration of Imaging and Clinical Data
Paper Status: original work, not submitted yet
Source Code Url: The source code will be released upon the publication of the paper.
Data Set Url: ADNI Dataset : http://adni.loni.usc.edu/
Registration: I acknowledge that publication of this at MIDL and in the proceedings requires at least one of the authors to register and present the work during the conference.
Authorship: I confirm that I am the author of this work and that it has not been submitted to another publication before.