Towards end-to-end optimisation of functional image analysis pipelines

TL;DR

This paper proposes converting neuroimaging data processing pipelines into end-to-end trainable neural networks, enabling joint optimization of parameters for improved task-specific performance, demonstrated through adaptive spatial smoothing in brain decoding.

Contribution

It introduces a novel approach to optimize entire neuroimaging pipelines using deep learning, moving beyond isolated module parameter tuning.

Findings

Adaptive spatial smoothing improves decoding accuracy.

End-to-end pipeline optimization enhances overall neuroimaging analysis.

Proof of concept validated on standard brain decoding task.

Abstract

The study of neurocognitive tasks requiring accurate localisation of activity often rely on functional Magnetic Resonance Imaging, a widely adopted technique that makes use of a pipeline of data processing modules, each involving a variety of parameters. These parameters are frequently set according to the local goal of each specific module, not accounting for the rest of the pipeline. Given recent success of neural network research in many different domains, we propose to convert the whole data pipeline into a deep neural network, where the parameters involved are jointly optimised by the network to best serve a common global goal. As a proof of concept, we develop a module able to adaptively apply the most suitable spatial smoothing to every brain volume for each specific neuroimaging task, and we validate its results in a standard brain decoding experiment.