Interpretability in Linear Brain Decoding

TL;DR

This paper introduces a formal definition of interpretability for linear brain decoding models and proposes a multi-objective criterion combining interpretability and performance for model selection, demonstrated on toy data.

Contribution

It provides the first formal definition of interpretability in linear brain decoding and a new criterion for model selection balancing interpretability and accuracy.

Findings

Optimizing hyper-parameters with the new criterion yields more informative models.

The definition enables quantitative evaluation of interpretability.

Preliminary results on toy data support the approach.

Abstract

Improving the interpretability of brain decoding approaches is of primary interest in many neuroimaging studies. Despite extensive studies of this type, at present, there is no formal definition for interpretability of brain decoding models. As a consequence, there is no quantitative measure for evaluating the interpretability of different brain decoding methods. In this paper, we present a simple definition for interpretability of linear brain decoding models. Then, we propose to combine the interpretability and the performance of the brain decoding into a new multi-objective criterion for model selection. Our preliminary results on the toy data show that optimizing the hyper-parameters of the regularized linear classifier based on the proposed criterion results in more informative linear models. The presented definition provides the theoretical background for quantitative evaluation of interpretability in linear brain decoding.