Spectral Graph Wavelet Transform as Feature Extractor for Machine Learning in Neuroimaging

TL;DR

This paper demonstrates that Spectral Graph Wavelet Transform (SGWT) effectively extracts features from brain graphs, leading to improved machine learning performance in neuroimaging applications, validated through synthetic and real fMRI data.

Contribution

It introduces the use of SGWT as a feature extraction method for neuroimaging graph data, optimizing spectrum coverage and showing significant performance gains over existing methods.

Findings

SGWT improves regression accuracy on synthetic brain signals.

SGWT enhances classification performance on fMRI datasets.

Optimized wavelet shapes increase spectrum coverage and feature quality.

Abstract

Graph Signal Processing has become a very useful framework for signal operations and representations defined on irregular domains. Exploiting transformations that are defined on graph models can be highly beneficial when the graph encodes relationships between signals. In this work, we present the benefits of using Spectral Graph Wavelet Transform (SGWT) as a feature extractor for machine learning on brain graphs. First, we consider a synthetic regression problem in which the smooth graph signals are generated as input with additive noise, and the target is derived from the input without noise. This enables us to optimize the spectrum coverage using different wavelet shapes. Finally, we present the benefits obtained by SGWT on a functional Magnetic Resonance Imaging (fMRI) open dataset on human subjects, with several graphs and wavelet shapes, by demonstrating significant performance improvements compared to the state of the art.