A comparison of multiband and multiband multiecho gradient‐echo EPI for task fMRI at 3 T

Abstract

A multiband (MB) echo‐planar imaging (EPI) sequence is compared to a multiband multiecho (MBME) EPI protocol to investigate differences in sensitivity for task functional magnetic resonance imaging (fMRI) at 3 T. Multiecho sampling improves sensitivity in areas where single‐echo‐EPI suffers from dropouts. However, It requires in‐plane acceleration to reduce the echo train length, limiting the slice acceleration factor and the temporal and spatial resolution Data were acquired for both protocols in two sessions 24 h apart using an adapted color‐word interference Stroop task. Besides protocol comparison statistically, we performed test–retest reliability across sessions for different protocols and denoising methods. We evaluated the sensitivity of two different echo‐combination strategies for MBME‐EPI. We examined the performance of three different data denoising approaches: “Standard,” “AROMA,” and “FIX” for MB and MBME, and assessed whether a specific method is preferable. We consider using an appropriate autoregressive model order within the general linear model framework to correct TR differences between the protocols. The comparison between protocols and denoising methods showed at group level significantly higher mean z‐scores and the number of active voxels for MBME in the motor, subcortical and medial frontal cortices. When comparing different echo combinations, our results suggest that a contrast‐to‐noise ratio weighted echo combination improves sensitivity in MBME compared to simple echo‐summation. This study indicates that MBME can be a preferred protocol in task fMRI at spatial resolution (≥2 mm), primarily in medial prefrontal and subcortical areas.

How do tasks impact the reliability of fMRI functional connectivity?

“…The literature suggests a MB acceleration factor of four improves cortical FC reliability, while a single band sequence is preferred for subcortical regions (M. Cahart et al, 2023;M.-S. Cahart et al, 2022). Further, the combination of multiband and multiecho sequences promise to yield higher reliability estimates than MB sequences alone (Cohen et al, 2021;Fazal et al, 2023;Lynch et al, 2020) (Finn, 2021;Meer et al, 2020;Sonkusare et al, 2019), thus leading to increases in reliability. Such improvements in reliability were found across all brain regions in a study that compared test-retest reliability of FC measures between movie-watching and rest (Wang, Ren, et al, 2017).…”

“…Cahart et al, 2022). Further, the combination of multiband and multiecho sequences promise to yield higher reliability estimates than MB sequences alone (Cohen et al, 2021; Fazal et al, 2023; Lynch et al, 2020). Congruent to increasing scan lengths and emerging neuroimaging protocols, our findings postulate that task conditions also increase FC‐TRC in specific regions and the choice of task used in a study may directly impact FC reliability of those regions.…”

Comparing the effect of multi gradient echo and multi band fMRI during a semantic task

“…In a follow up study, Kirilina et al, (2016) showed greater task activation during emotional and reward-based learning in orbitofrontal cortices, most likely due to effective recovery of signal in these susceptible areas, but less activation in deep brain regions, most likely due to higher acceleration and/or smaller voxel size. Similarly, Fazal et al, (2023) showed that multi-echo multi-band had greater sensitivity in areas affected by signal inhomogeneity but multi-band only did show greater sensitivity in visual areas (not deep brain regions).…”

“…This is most likely due to lower tSNR in brain regions that are further from the receiver coils, which is exaggerated when using higher gfactors for in-plane acceleration (Kirilina et al, 2016). Although, Fazal et al, (2023) showed that multi-echo multi-band had reduced sensitivity in visual and anterior cingulate regions compared to multi-band only during a Stroop task.…”

“…Thus while comparing results from a single versus multiple echoes from the same multi-echo sequence provides tight control of other sequence parameters, thereby isolating the advantage of multi-echoes, it is not a fair comparison for practical decisions about whether to use a multi-echo versus standard single-echo protocol. Critically, we found only two studies that compared a multi-echo protocol against a typical single-echo protocol (Poser, Versluis, Hoogduin, & Norris, 2006; Kirilina, Lutti, Poser, Blankenburg, & Weiskopf, 2016); remaining studies used a multi-band accelerated single-echo protocol (Cohen, Chang, & Wang, 2021; Cohen, Jagra, Visser, et al, 2021; Cohen, Jagra, Yang, et al, 2021; Cohen, Yang, Fernandez, Banerjee, & Wang, 2021; Fazal et al, 2023; Lynch et al, 2020) and/or compared different denoising strategies (Lombardo et al, 2016). Using resting-state fMRI (rsfMRI), Poser et al (2006) demonstrated that multi-echo data had better functional contrast-to-noise ratio (CNR) across the brain than a typical protocol, particularly in susceptibility-prone regions.…”