Quantifying the spatial resolution of the gradient echo and spin echo BOLD response at 3 Tesla

Abstract

The blood oxygen level dependent (BOLD) response, as measured with fMRI, offers good spatial resolution compared to other non-invasive neuroimaging methods. The use of a spin echo technique rather than the conventional gradient echo technique may further improve the resolution by refocusing static dephasing effects around the larger vessels, so sensitizing the signal to the microvasculature. In this work the width of the point spread function (PSF) of the BOLD response at a field strength of 3 Tesla is compared for these two approaches. A double echo EPI pulse sequence with simultaneous collection of gradient echo and spin echo signal allows a direct comparison of the techniques. Rotating multiple-wedge stimuli of different spatial frequencies are used to estimate the width of the BOLD response. Waves of activation are created on the surface of the visual cortex, which begin to overlap as the wedge separation decreases. The modulation of the BOLD response decreases with increasing spatial frequency in a manner dependent on its width. The spin echo response shows a 13% reduction in the width of the PSF, but at a cost of at least 3-fold reduction in contrast to noise ratio. Good spatial resolution is one of the principle advantages of fMRI compared to other neuroimaging methods, such as MEG or EEG. However, compared to most MRI techniques, the typical resolution of the blood oxygenation level dependent (BOLD) response is quite poor. The resolution is limited by physiologic rather than technical considerations, with the vascular response to neural activity extending over several millimeters. The change in deoxyhaemoglobin content in the draining veins and venules leads to inaccurate localization of neural activity (1), and also to poor precision by widening the spatial extent or point spread function (PSF) of the response, resulting in an inability to resolve activity from close sources (2,3). In general, localization and resolution are not related; for example, it is perfectly possible to have very high resolution signal in the wrong location. However, in this case, deoxyhaemoglobin changes in venous vessels distant from the site of neuronal activation will degrade both measures.Recent work suggests that at a field strength of 3T, a spin echo (SE) sequence could improve the spatial resolution of the BOLD response compared to the standard gradient echo (GE) technique (4,5). To understand this we need to consider the relative signal contribution from both the intra-and extravascular spaces. The extravascular signal change is due to the dephasing effect of local field gradients surrounding the blood vessels. Water protons surrounding capillaries will move a considerable distance relative to the capillary diameter during the echo time and, hence, will experience a range of field gradients. This dynamic dephasing is a random process that cannot be refocused by a spin-echo. Water protons surrounding large vessels, however, will tend to remain in the same magnetic field during the echo time, resulting in little dynamic d...

The effect of large veins on spatial localization with GE BOLD at 3 T: Displacement, not blurring

“…The negative responses were also observed in many voxels that appeared to be well within the cortical representation of the stimulus and that responded positively to the stimulus in a differential protocol. Presumably because of their proximity to large veins (marked by high mean-normalized variance in the HIRF scans), these voxels responded negatively We conclude that the negative BOLD signal is being displaced, rather than blurred, because we measured inconsistent responses in voxels within the cortical representation of a stimulus, while a blurring or point-spread function model (Engel et al, 1997;Parkes et al, 2005) would predict inconsistent responses outside of the cortical representation. The key difference is measured in voxels that are within the cortical representation of the stimulus, but which contain (or are dominated by) a large vein that collects blood from outside the cortical representation of the stimulus.…”

“…Even though the BOLD response is a spatially and functionally indirect measure of neural activity, many studies have demonstrated that BOLD experiments can provide excellent localization of neural activity (e.g., Cheng et al, 2001; for review see Heeger and Ress, 2002). To characterize the precision of the BOLD response, several studies have measured approximate cortical point-spread functions (Engel et al, 1997;Parkes et al, 2005). However, the accuracy with which BOLD fMRI can localize neural responses is known to be dependent on both experiment design and data analysis.…”

“…The true spatial resolution of BOLD fMRI, however, is degraded by correlation between neighboring regions of cortex (Grinvald et al, 1994), T 2 ( * ) blurring of the images (Farzaneh et al, 1990), and the presence of large draining veins (Duvernoy et al, 1981). While spin echo (SE) techniques can reduce the contribution of signal from large veins on the pial surface of the cortex (Yacoub et al, 2003), the gain in spatial sensitivity comes at the cost of a significant loss in signal-to-noise ratio (Norris et al, 2002;Parkes et al, 2005). While ultra high-field experiments can clearly benefit from the increased spatial specificity of SE BOLD (Yacoub et al, 2005), the low signal-to-noise ratio at 3 T means that the technique can fail to yield useful data.…”

Interstudy reproducibility of SSFP cine magnetic resonance: Impact of magnetic field strength and parallel imaging

“…Since the SSFP sequence exhibits T2/T1 contrast weighting, reducing T2 of the RV blood pool would effectively decrease signal intensity (1). Blood oxygen level-dependent (BOLD) effects on T2 are well described (21), and are more significant at high field strength (22). Nevertheless, a more comprehensive characterization of right ventricular CNR at 3T is an area for further study.…”

Whole brain, high resolution multiband spin-echo EPI fMRI at 7T: A comparison with gradient-echo EPI using a color-word Stroop task

“…Given that gradient echo (GE) EPI BOLD contrast is augmented by the dephasing from the venous blood (draining vessels), GE EPI and SE EPI have been labeled as being more sensitive and more specific to the true site of neuronal activation, respectively (Norris, 2012;Olman and Yacoub, 2011;Parkes et al, 2005;Uludağ et al, 2009;Yacoub et al, 2005).…”