Gaussian Processes for HRF estimation for BOLD fMRI

TL;DR

This paper introduces a non-parametric method using Gaussian processes to jointly estimate the HRF and task activation in fMRI, enabling continuous evaluation and variance estimation for jittered paradigms.

Contribution

It presents a novel Gaussian process-based approach for HRF estimation that improves flexibility and provides uncertainty quantification.

Findings

Enables continuous HRF evaluation in jittered paradigms

Provides variance estimates at each time point

Improves accuracy of task activation detection

Abstract

We present a non-parametric joint estimation method for fMRI task activation values and the hemodynamic response function (HRF). The HRF is modeled as a Gaussian process, making continuous evaluation possible for jittered paradigms and providing a variance estimate at each point.