Normative Modeling of Neuroimaging Data using Scalable Multi-Task Gaussian Processes

TL;DR

This paper introduces a scalable multi-task Gaussian process regression method for normative modeling of neuroimaging data, enabling efficient high-dimensional analysis while accounting for spatial covariance, thus improving sensitivity in detecting anomalies.

Contribution

The paper presents a novel scalable multi-task Gaussian process regression approach that incorporates spatial covariance, reducing computational complexity for high-dimensional neuroimaging data.

Findings

Substantial computational improvements enabling high-dimensional modeling

Higher sensitivity in novelty detection by modeling spatial and sample variances

Effective modeling of spatial structure in fMRI data

Abstract

Normative modeling has recently been proposed as an alternative for the case-control approach in modeling heterogeneity within clinical cohorts. Normative modeling is based on single-output Gaussian process regression that provides coherent estimates of uncertainty required by the method but does not consider spatial covariance structure. Here, we introduce a scalable multi-task Gaussian process regression (S-MTGPR) approach to address this problem. To this end, we exploit a combination of a low-rank approximation of the spatial covariance matrix with algebraic properties of Kronecker product in order to reduce the computational complexity of Gaussian process regression in high-dimensional output spaces. On a public fMRI dataset, we show that S-MTGPR: 1) leads to substantial computational improvements that allow us to estimate normative models for high-dimensional fMRI data whilst accounting for spatial structure in data; 2) by modeling both spatial and across-sample variances, it provides higher sensitivity in novelty detection scenarios.