High angular resolution diffusion imaging with stimulated echoes: compensation and correction in experiment design and analysis.

Henrik Lundell,Daniel C Alexander,Tim B Dyrby

doi:10.1002/nbm.3137

Abstract

Stimulated echo acquisition mode (STEAM) diffusion MRI can be advantageous over pulsed-gradient spin-echo (PGSE) for diffusion times that are long compared with T2. It therefore has potential for biomedical diffusion imaging applications at 7T and above where T2 is short. However, gradient pulses other than the diffusion gradients in the STEAM sequence contribute much greater diffusion weighting than in PGSE and lead to a disrupted experimental design. Here, we introduce a simple compensation to the STEAM acquisition that avoids the orientational bias and disrupted experiment design that these gradient pulses can otherwise produce. The compensation is simple to implement by adjusting the gradient vectors in the diffusion pulses of the STEAM sequence, so that the net effective gradient vector including contributions from diffusion and other gradient pulses is as the experiment intends. High angular resolution diffusion imaging (HARDI) data were acquired with and without the proposed compensation. The data were processed to derive standard diffusion tensor imaging (DTI) maps, which highlight the need for the compensation. Ignoring the other gradient pulses, a bias in DTI parameters from STEAM acquisition is found, due both to confounds in the analysis and the experiment design. Retrospectively correcting the analysis with a calculation of the full B matrix can partly correct for these confounds, but an acquisition that is compensated as proposed is needed to remove the effect entirely.

Highlights

Stimulated echo acquisition mode (STEAM) diffusion MRI (1,2) offers advantages over the more common pulsed-gradient spin-echo (PGSE) diffusion MRI when T2 is short compared with the diffusion time and T1 ≫ T2
We propose a simple compensation for the STEAM sequence, referred to as compensated acquisition, that accounts for the unwanted directional bias caused by the butterfly gradients
We demonstrate the need and effectiveness of the compensation for STEAM through High angular resolution diffusion imaging (HARDI)–diffusion tensor imaging (DTI) experiments in simulation and on data acquired from a fixed monkey brain

Summary

Introduction

Stimulated echo acquisition mode (STEAM) diffusion MRI (1,2) offers advantages over the more common pulsed-gradient spin-echo (PGSE) diffusion MRI when T2 is short compared with the diffusion time and T1 ≫ T2. The STEAM signal decays with rate T1 during the mixing time τm, which determines the diffusion time, whereas T2 decay occurs throughout the whole PGSE sequence. STEAM diffusion MRI is common in tissue with short T2, such as muscle or cartilage (3). STEAM is useful for measuring diffusivity in water compartments with significantly lower T2, like myelin water, as previously shown in preclinical settings with PGSE (5). Ex vivo q-space studies of brain tissue (7–9) usually prefer STEAM over PGSE, as fixation and lower temperature reduce diffusivity compared with in vivo studies, which increases the necessary diffusion times (10,11). Dyrby et al (12) demonstrated the need for long diffusion times to ensure sensitivity to large axon diameter, which is important for microstructure imaging techniques such as ActiveAx (13) and AxCaliber (8)

Methods

Results

Discussion

Conclusion