Original contribution
Hadamard-encoded sub-slice fMRI for reduced signal dropout

https://doi.org/10.1016/j.mri.2011.07.019Get rights and content

Abstract

T2⁎-weighted blood oxygen level-dependent functional magnetic resonance imaging is adversely affected by susceptibility-induced field gradients in brain regions adjoining air interfaces that cause image distortion and signal dropout. Reducing slice thickness diminishes signal dropout but is accompanied by reduced signal-to-noise ratio (SNR). This study proposes simultaneous excitation of subslices with total width equal to the desired slice thickness, employing alternating Hadamard-encoded radiofrequency pulses coupled with incoherent addition of the subslices to achieve reduction of through-plane dephasing with minimal SNR loss but at the expense of a reduction in temporal resolution. Using a sensory task and hypercapnic challenge with breathholding (BH), results with two subslices per slice and a twofold reduction in temporal resolution show improved activation relative to a conventional acquisition. Average (eight subjects) T-scores in the BH task increased by 16% (P<.0003), and activation extent increased by 12% (not significant). In frontal brain regions, significant improvements in BH activation extent (11.4%, P<.05) and T-scores (18%, P<.0002) were demonstrated. Higher temporal resolution can be achieved by tradeoff of SNR.

Introduction

Blood oxygen level-dependent (BOLD) functional magnetic resonance imaging (fMRI) [1], [2], [3] typically uses T2-weighted imaging to provide sensitivity to small changes in deoxyhemoglobin concentration consequent to neural metabolism changes. However, this sensitization also renders the acquisition subject to image distortion and signal dropout from localized off-resonance conditions due to susceptibility induced field gradients (SFGs), especially in brain regions adjoining air interfaces [4].

Many approaches have been developed to reduce signal loss in BOLD fMRI. These include 2D [5] and 3D [6] zshimming, which cancel SFGs in the slice select direction by acquiring images with several values of compensatory zshim gradients, but at severe signal-to-noise ratio (SNR) efficiency loss in uniform brain. Another approach uses multiple echoes to synthesize the optimal TE for each voxel [7], but the longer echo train length increases TR and again reduces SNR efficiency. Multidimensional tailored radiofrequency (RF) pulses can reduce signal loss by customizing the selective excitation pulse for each brain region, but the pulses are too long to be practical [8]. Accelerated methods use multi-channel coils to shorten the readout duration and reduce signal loss when small voxels are acquired [9], but again, the SNR loss can be substantial. Spiral-in/out methods acquire two separate images before and after TE and combine them adaptively [10]. While this method is very effective, some signal losses remain. The simple expedient of using smaller voxels is effective in reducing signal loss from SFGs but, again, at SNR penalty. Some of this SNR loss is recovered by combining reconstructed voxels incoherently as, for example, in acquiring thin slices and combining adjacent slices' magnitude images [11]. However, the full SNR of a single thick slice is not obtained in the combination because additional scan time is allocated to the sub-slice acquisitions, leading to a loss of temporal resolution as well. For example, halving the slice thickness gives a uniform-brain SNR loss by the factor √2/4 in each thin slice or 1/2 after combination. Other techniques have also been developed, e.g., Refs. [12], [13], [14], but none fully solve the problem of signal dropout without incurring a substantial uniform-brain SNR penalty.

A 3D method was demonstrated whereby M slices are simultaneously encoded, using M Hadamard excitation pulses rather than phase encoding [15]. Reconstruction of the individual slices uses linear inversion based on the Hadamard matrix. Because the excitation volume is increased M-fold, the SNR advantage is identical to that achieved by conventional 3D methods.

In the following sections, we adapt this Hadamard approach for a 2D acquisition to obtain simultaneous subslice sampling of M thinner slices with the same total slice thickness as the conventional excitation. The thinner subslices result in less through-plane dephasing but reduced SNR, since each subslice has 1/M of the original voxel volume. However, during reconstruction, the subslices are combined incoherently to achieve signal recovery, at a loss in temporal resolution in uniform brain.

Section snippets

Slice excitation

Consider a slice that has a SFG B(z) perpendicular to the plane of the slice. Let us assume the slice can be modeled as a combination of M thinner subslices, each of which is homogeneous and in a piecewise uniform magnetic field. Using conventional slice excitation, within a given voxel, the signal In at time frame n is the complex sum of the subslice signals:In=eiϕm=1MSm,neiθm,where Sm,n is the real-valued signal intensity in the mth subslice, θm=γΔBmTE is the phase difference between

RF pulse design

A Hamming-weighted sinc pulse E(t) was employed both for conventional excitation and as the basis for L(t) and R(t), with bandwidth 2πω0=1.25kHz and duration T=6.4 ms. Then, the Hadamard pulses E1,2(t) are generated by adding two shifted pulses together (small flip angle approximation):E1,2(t)=L(t)±iR(t)=[eiω0t/2±ieiω0t/2]E(t),|t|T/2.

The slice select gradient has amplitude Gz0/λw for the conventional excitation and twice that for the Hadamard method with subslice width of w/2. The two

Results

Typical SM task activation results are shown in Fig. 3 for one subject. Note the overall apparent increase in activation in most slices for the Hadamard method; this is confirmed by the histograms showing whole-brain voxel counts. As shown in Table 1, the average voxel counts were greater for the Hadamard method by 5.7%, but this was not significant (P=.295). The T-scores did not differ. Thus, even though most subjects exhibited the same or slightly greater SM activation with the Hadamard

Discussion

This study introduces the concept of simultaneous sub-slice excitation using complex Hadamard encoding coupled with a novel subslice combination method, with the aim of providing enhanced signal recovery while partially mitigating the usual severe SNR penalty of thin slices. By alternating the Hadamard slice encoding pulse phases with time frame, the intraslice dephasing signal Yu is cast to the Nyquist frequency (for the two-subslice case here) and removed using temporal-domain filtering.

Group

Acknowledgments

The authors thank anonymous reviewers for comments, gratefully acknowledge support from NIH grants F31-AG032168 (CC) and P41-RR009784 (GHG) and thank their volunteer scan subjects.

References (24)

  • V.A. Stenger et al.

    Three-dimensional tailored, RF pulses for the reduction of susceptibility artifacts in T2⁎-weighted functional MRI

    Magn Reson Med

    (2000)
  • M. Weiger et al.

    Sensitivity-encoded single-shot spiral imaging for reduced susceptibility artifacts in BOLD fMRI

    Magn Reson Med

    (2002)
  • Cited by (3)

    • Spiral imaging in fMRI

      2012, NeuroImage
      Citation Excerpt :

      Over the years since the early experiments, more powerful and higher field strength systems have become available, and the burgeoning use of fMRI in cognitive and clinical neuroscience has continued to drive the development of advanced fMRI methods using spiral trajectories (Bammer et al., 2005; Guo and Song, 2003; Sha et al., 2003; Truong and Song, 2008). Other fMRI developments using spiral readouts include 3D (Hu and Glover, 2007, 2009; Lai and Glover, 1998; Yang et al., 1996; Yang et al., 1999), Hadamard sub-slice encoding (Glover and Chang, 2011), diffusion (Liu et al., 2004; Liu et al., 2005; Song et al., 2004), and perfusion (Hsu and Glover, 2006). Implementation of spiral methods, however, is more complex than EPI.

    View full text