Adapting HFMCA to Graph Data: Self-Supervised Learning for Generalizable fMRI Representations

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Frac, Jakub, Schmatz, Alexander, Li, Qiang, Van Wingen, Guido, Yu, Shujian
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918155449794560
author Frac, Jakub
Schmatz, Alexander
Li, Qiang
Van Wingen, Guido
Yu, Shujian
author_facet Frac, Jakub
Schmatz, Alexander
Li, Qiang
Van Wingen, Guido
Yu, Shujian
contents Functional magnetic resonance imaging (fMRI) analysis faces significant challenges due to limited dataset sizes and domain variability between studies. Traditional self-supervised learning methods inspired by computer vision often rely on positive and negative sample pairs, which can be problematic for neuroimaging data where defining appropriate contrasts is non-trivial. We propose adapting a recently developed Hierarchical Functional Maximal Correlation Algorithm (HFMCA) to graph-structured fMRI data, providing a theoretically grounded approach that measures statistical dependence via density ratio decomposition in a reproducing kernel Hilbert space (RKHS),and applies HFMCA-based pretraining to learn robust and generalizable representations. Evaluations across five neuroimaging datasets demonstrate that our adapted method produces competitive embeddings for various classification tasks and enables effective knowledge transfer to unseen datasets. Codebase and supplementary material can be found here: https://github.com/fr30/mri-eigenencoder
format Preprint
id arxiv_https___arxiv_org_abs_2510_05177
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Adapting HFMCA to Graph Data: Self-Supervised Learning for Generalizable fMRI Representations
Frac, Jakub
Schmatz, Alexander
Li, Qiang
Van Wingen, Guido
Yu, Shujian
Image and Video Processing
Machine Learning
Functional magnetic resonance imaging (fMRI) analysis faces significant challenges due to limited dataset sizes and domain variability between studies. Traditional self-supervised learning methods inspired by computer vision often rely on positive and negative sample pairs, which can be problematic for neuroimaging data where defining appropriate contrasts is non-trivial. We propose adapting a recently developed Hierarchical Functional Maximal Correlation Algorithm (HFMCA) to graph-structured fMRI data, providing a theoretically grounded approach that measures statistical dependence via density ratio decomposition in a reproducing kernel Hilbert space (RKHS),and applies HFMCA-based pretraining to learn robust and generalizable representations. Evaluations across five neuroimaging datasets demonstrate that our adapted method produces competitive embeddings for various classification tasks and enables effective knowledge transfer to unseen datasets. Codebase and supplementary material can be found here: https://github.com/fr30/mri-eigenencoder
title Adapting HFMCA to Graph Data: Self-Supervised Learning for Generalizable fMRI Representations
topic Image and Video Processing
Machine Learning
url https://arxiv.org/abs/2510.05177