Partial-k-space acquisition method for improved SNR efficiency and temporal resolution in 3D fMRI

Department of Physics, Stanford University, Stanford, California 94305-5488, USA.
Magnetic Resonance in Medicine (Impact Factor: 3.4). 05/2006; 55(5):1106-13. DOI: 10.1002/mrm.20877
Source: PubMed

ABSTRACT Previous studies have shown the relative importance of physiological noise and thermal noise in 2D MR images. Since physiological noise is proportional to the signal, it can be the dominant component at the center of k-space. In this study we demonstrate that the signal-to-noise ratio (SNR) efficiency and temporal resolution for 3D functional MRI (fMRI) are increased by the use of a partial-k-space acquisition method. In partial-k-space methods, the high-spatial-frequency components are doubled in amplitude during reconstruction, resulting in twice as much noise from those components. However, in sum these contributions are relatively small compared to those at the low spatial frequencies, where physiological noise is dominant. Therefore, the effect on the final MR images is almost negligible due to the square summation rule. Thus, the partial-k-space 3D method sacrifices much less SNR than is expected from the thermal noise model, and the SNR efficiency is increased compared to a full-k-space acquisition since more time frames can be collected for the same scan time. Accordingly, the temporal resolution can be increased in 3D acquisitions because only partial coverage of k-space is necessary. Experimental results confirm that more activation with a higher average t-score is detected by this method.

  • [Show abstract] [Hide abstract]
    ABSTRACT: A new acquisition scheme for T2-weighted spin-echo BOLD fMRI is introduced. It uses a T2-preparation module to induce blood-oxygenation-level-dependent (BOLD) contrast, followed by a single-shot three-dimensional (3D) fast gradient-echo readout with short echo time (TE). It differs from most spin-echo BOLD sequences in that BOLD contrast is generated before the readout, which eliminates the "dead time" due to long TE required for T2 contrast, and substantially improves acquisition efficiency. This approach, termed "3D T2prep-GRE," was implemented at 7 Tesla (T) with a typical spatial (2.5 × 2.5 × 2.5 mm(3) ) and temporal (TR = 2.3 s) resolution for functional MRI (fMRI) and whole-brain coverage (55 slices), and compared with the widely used 2D spin-echo EPI sequence. In fMRI experiments of simultaneous visual/motor activities, 3D T2prep-GRE showed minimal distortion and little signal dropout across the whole brain. Its lower power deposition allowed greater spatial coverage (55 versus 17 slices with identical TR, resolution and power level), temporal SNR (60% higher) and CNR (35% higher) efficiency than 2D spin-echo EPI. It also showed smaller T2* contamination. This approach is expected to be useful for ultra-high field fMRI, especially for regions near air cavities. The concept of using T2-preparation to generate BOLD contrast can be combined with many other sequences at any field strength. Magn Reson Med, 2013. © 2013 Wiley Periodicals, Inc.
    Magnetic Resonance in Medicine 12/2014; 72(6). DOI:10.1002/mrm.25055 · 3.40 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: The somatosensory fMRI response to electrical stimulation of the middle phalange of the second digit of four rats at a spatial resolution of 200 micron cubic at 9.4 T is reported. At high threshold (P < 0.002), activated voxels encompass a penetrating vein that passes across gray matter. These voxels lie mostly in three contiguous slices perpendicular to the pial surface. This activation is assigned to the representation in the forepaw barrel subfield (FBS) of a single cortical column of this phalange. In addition, activation of the Indusium Grieseum (IG) is visualized robustly. Voxels revealed by fMRI were used to observe functional connectivity to other voxels of the sensorimotor cortex using fcMRI. Results of this experiment were analyzed as a function of decreasing threshold, which exhibited spreading connectivity that revealed S2, M1/M2 and contralateral S1. Noting that every cubic mm of tissue contains 125 voxels, connectivity patterns are complex. It is hypothesized that they reflect connections within gray matter by association fibers. S2 and IG revealed connectivities with many voxels across the sensorimotor cortex. These regions also showed sub-regional variation of connectivity. A 1 cm diameter surface coil with a local low-noise rf amplifier was used in these studies. The usual region of sensitivity (ROS) of such a coil is 1 cm diameter by 0.5 cm depth. Significant connectivity was observed between time courses of voxels that were within the ROS and voxels that were outside, which extends the volume of tissue that can be observed by the methods of this paper.
    08/2014; DOI:10.1089/brain.2014.0281
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Although 2D echo-planar imaging (EPI) remains the dominant method for functional MRI (FMRI), 3D readouts are receiving more interest as these sequences have favorable signal-to-noise ratio (SNR) and enable imaging at a high isotropic resolution. Spoiled gradient-echo (SPGR) and balanced steady-state free-precession (bSSFP) are rapid sequences that are typically acquired with highly segmented 3D readouts, and thus less sensitive to image distortion and signal dropout. They therefore provide a powerful alternative for FMRI in areas with strong susceptibility offsets, such as deep gray matter structures and the brainstem. Unfortunately, the multi-shot nature of the readout makes these sequences highly sensitive to physiological fluctuations, and large signal instabilities are observed in the inferior regions of the brain. In this work a characterization of the source of these instabilities is given and a new method is presented to reduce the instabilities observed in 3D SPGR and bSSFP. Rapidly acquired single-slice data, which critically sampled the respiratory and cardiac waveforms, showed that cardiac pulsation is the dominant source of the instabilities. Simulations further showed that synchronizing the readout to the cardiac cycle minimizes the instabilities considerably. A real-time synchronization method was therefore developed, which utilizes parallel-imaging techniques to allow cardiac synchronization without alteration of the volume acquisition rate. The implemented method significantly improves the temporal stability in areas that are affected by cardiac-related signal fluctuations. In bSSFP data the tSNR in the brainstem increased by 45%, at the cost of a small reduction in tSNR in the cortical areas. In SPGR the temporal stability is improved by approximately 20% in the subcortical structures and as well as cortical gray matter when synchronization was performed.
    NeuroImage 06/2011; 57(4):1364-75. DOI:10.1016/j.neuroimage.2011.05.070 · 6.13 Impact Factor


Available from