Development of a novel task-based functional magnetic resonance imaging phantom based on a bubble-compression approach.

Abstract

Phantoms used in previous functional magnetic resonance imaging (fMRI) studies have drawbacks, such as a complicated circuit and equipment use, a single signal-change rate, and T2 * values that do not correspond to those of living human brains. We aimed to develop a phantom for use in task-based fMRI studies (gradient-echo echo-planar imaging; GRE-EPI) with bioequivalent T1 and T2 * values, using an innovative method to control the rate of signal change. A gel phantom with T1 and T2 * values equivalent to that of the living brain gray matter was fixed in a 150mm diameter container, with five holes, each of which could hold a 30-ml syringe. The gel phantom contained microscopic air bubbles; this made it possible to control the percent signal change by injector-induced water pressure changes. Using this phantom, we investigated the percent signal change, derived an equation that can approximately reproduce an arbitrary percent signal change, compared different gel phantom samples, investigated the change in relaxation time and bubble size during signal change, and assessed the change in values in each sample over time. The relaxation time of the gel phantom was similar to the literature values for gray matter. The percent signal change achieved was approximately 0%-13.51% and was dependent on the water pressure change. The derived equation was y=0.000008x3 - 0.000771x2 + 0.034222x - 0.026054, with y being the percent signal change and x being the pressure in kPa; the reproducibility was high. No significant difference was detected among samples of gray matter gel phantoms (P>0.05). The change in the rate of signal change with the change in water pressure was due to the change in T2 * value with the change in bubble size. With pressure increasing from 0 to 151.7kPa, the T2 * value increased from 52ms to 85ms. The newly developed gel phantom was stable for 60 days, but its bubble size changed after 21 days. We developed a novel phantom for use in fMRI, which could reproduce minute signal changes similar to the blood-oxygen-level-dependent effect and with bioequivalent T1 and T2 * values, and used an innovative method to control the percent signal change by compressing the air contained in the phantom for validation of fMRI using GRE-EPI. This phantom reproduced the percent signal change due to changes in T2 * values, which is very similar to scanning a human body. This phantom is expected to be a powerful tool for advancing the study of task-based fMRI.