Enhanced relative BOLD signal changes in T₂‐weighted stimulated echoes

Abstract

The origin of the stimulus/task-induced signal changes in spin echo (SE) functional MRI (fMRI) at high magnetic fields is dynamic averaging due to diffusion in the presence of field gradients surrounding deoxyhemoglobin-containing microvasculature. The same mechanism is expected to be operative in stimulated echoes (STE). Compared to SE-fMRI, however, STEfMRI has the potential for larger diffusion weighting and consequently larger stimulus/task-induced signal changes as a result of an additional delay, the mixing time, T M . In the present study, functional signal changes were quantified for both primary echo (PRE) and STE as a function of echo and mixing time. The relative blood oxygenation level dependent (BOLD) signal changes in STE were larger than in PRE at the same echo time and increased with both mixing and echo time. The contrastto-noise ratio (CNR) of the STE, however, is close to the CNR of the PRE, indicating an increase of physiological noise with longer mixing times. In addition, the signal attenuation due to diffusion in the presence of magnetic field gradients near blood vessels was modeled using Monte Carlo simulations. Key words: stimulated echo; fMRI; transverse relaxation; T 2 ; dynamic averaging Gradient-echo (GE) echo-planar imaging (EPI) is commonly applied to probe neuronal activity with relatively high spatial and temporal resolution because this fast imaging method has a high signal-to-noise ratio (SNR) and exhibits large blood oxygenation level dependent (BOLD) signal changes. However, GE techniques are limited in their accuracy or specificity relative to the site of neuronal activity. In such functional magnetic resonance imaging (fMRI) experiments, the main mechanism causing BOLD signal changes is static dephasing, i.e., signal variation originating from phase differences between spins experiencing different magnetic fields near deoxyhemoglobincontaining blood vessels. As larger draining veins also contribute to the relative signal changes, the BOLD response can spread noticeably beyond the area of neuronal activity. Intravascular T 2 -changes, which are caused by spins diffusing in local field variations around and within deoxyhemoglobin-loaded red blood cells, add to the spatial blurring of the BOLD response.The BOLD signal variations are much more confined to the site of neuronal activity using moderately diffusionweighted spin echo (SE) imaging techniques. In these sequences, static dephasing is refocused and does not contribute to the signal change. The apparent T 2 is modulated through diffusion-induced dynamic averaging of the susceptibility gradients in the vicinity of capillaries and venules. Simulations and experiments demonstrated that T 2 -BOLD signal changes predominantly occur in the capillary bed (1,2). SE EPI is, therefore, assumed to be more specific to the site of neuronal activity. Specificity is further improved by mild diffusion-weighting to suppress intravascular signal contributions.SE-based imaging techniques, however, are hampered by the fact that the stimulu...

Cortical Depth Dependent Functional Responses in Humans at 7T: Improved Specificity with 3D GRASE

“…However, the functional contrast in 3D GRASE images consists of both T 2 and stimulated echo weightings. It has been shown with modeling that the contrast generated by stimulated echoes originates from dynamic averaging due to diffusion and is thus similar to conventional SE-EPI with the potential advantage of higher signal changes due to longer sampling periods [39]. However, whether or not 3D GRASE provides any specificity advantages over conventional GE for high resolution applications of cortical layers or columns, as has been shown for conventional SE BOLD imaging at high fields, has not yet been established.…”

“…Using 3D GRASE imaging [37], [38] with a train of refocused pulses to encode a 3D slab combined with inner volume excitation, facilitates the efficient acquisition of high isotropic resolution images with higher SNR over extended volumes compared to inner volume based 2D SE-EPI, which has produced high resolution fMRI maps in humans. The resulting 3D images, however, do have stimulated-echo weighted contrast [39] in addition to T 2 weighting. Despite this, we were able to successfully map axis of motion columns in human area MT [10], demonstrating the high degree of spatial specificity achievable with 3D GRASE.…”

Sub-millimeter T2 weighted fMRI at 7 T: comparison of 3D-GRASE and 2D SE-EPI

“…The later (non-center) partitions in the 3D-GRASE echo-train are sampled at a later effective TE than the nominal TE, which was matched with the TE of 2D SE-EPI 1. Therefore, the T 2 contribution is higher in 3D-GRASE because later partitions experience more T 2 weighting and additional T 2 weighted stimulated echo weighting than the center partition (Goerke et al, 2007 ). In addition, blood vessels have a particularly sharp spatial profile (compared to smoother gray or white matter structures), which manifests itself in the k-space regions corresponding to high spatial frequencies.…”

Whole brain, high resolution spin-echo resting state fMRI using PINS multiplexing at 7T

“…Consequently, the use of T 2 -weighted fMRI is desirable for the sake of a higher spatial specificity of BOLD activation. Owing to its high sensitivity, ready availability and (in relation to pure spin echo sequences) low radiofrequency power deposition SE-EPI is the most promising candidate sequence, but several alternatives have also been tested [FSE (Constable et al, 1994), RASER (Chamberlain et al, 2007), nb-S2-SSFP (Barth et al, 2010), HASTE (Poser and Norris, 2007), STE (Goerke et al, 2007), bSSFP (Miller et al, 2003;Scheffler et al, 2001)]. Unfortunately, the application of a sufficiently fast, whole brain SE-EPI protocol faces several practical limitations that have impeded the common use of T 2 -weighting for fMRI:…”