BOLDsωimsuite: A new software suite for forward modeling of the BOLD fMRI signal

Abstract

In trying to elucidate the mechanisms of the blood-oxygenation level-dependent (BOLD) functional MRI (fMRI) techniques and to expand on the potential of the transverse relaxation time (T2*) in quantitative MRI, many methods for the forward modelling of the BOLD effect have been created and analyzed. Simulations of this nature can be difficult to implement without prior experience, and differences made by methodological choices can be unclear, which provides a significant barrier of entry into the field. In this paper, we present BOLDswimsuite, a toolbox for forward modeling of the BOLD effect, which collects many of the principal methods used in the literature into a single coherent package. Implemented as a Python package, simulations are built-in scripts by combining various simulation components, while providing flexibility in methodological choices. The goal of this toolbox is to provide an open-source, reproducible software suite that is adaptable for simulations in different MRI applications, and to which additional features can be added by the user with relative ease. This paper first provides an overview of the methods available in the package and how these methods can be constructed from the toolboxs modular code components. Then, a brief theoretical explanation of each simulation component is given, supported by the relevant contributors. Next, sample simulations and analyses that can be created using the package are presented to display its features. Finally, recommendations regarding computational requirements are included to help users choose the best simulation methods to fit their needs. This package has many use cases and significantly reduces methodological barriers to forward modeling. It can also be a good learning tool for MR physics as well as a powerful tool to promote reproducible science.