Multiband multislice GE-EPI at 7 tesla, with 16-fold acceleration using partial parallel imaging with application to high spatial and temporal whole-brain fMRI

Abstract

Parallel imaging in the form of multiband radiofrequency excitation, together with reduced k-space coverage in the phase-encode direction, was applied to human gradient echo functional MRI at 7 T for increased volumetric coverage and concurrent high spatial and temporal resolution. Echo planar imaging with simultaneous acquisition of four coronal slices separated by 44mm and simultaneous 4-fold phase-encoding undersampling, resulting in 16-fold acceleration and up to 16-fold maximal aliasing, was investigated. Task/stimulus-induced signal changes and temporal signal behavior under basal conditions were comparable for multiband and standard single-band excitation and longer pulse repetition times. Robust, whole-brain functional mapping at 7 T, with 2 x 2 x 2mm(3) (pulse repetition time 1.25 sec) and 1 x 1 x 2mm(3) (pulse repetition time 1.5 sec) resolutions, covering fields of view of 256 x 256 x 176 mm(3) and 192 x 172 x 176 mm(3), respectively, was demonstrated with current gradient performance.

FIG. 1

Flow diagram for unaliasing multiband EPI data with GRAPPA. The single-band data are a single time point acquisition from a separate experiment. ACSs obtained during the multiband acquisition are not used.

FIG. 2

k-space with 4-fold undersampling in the PE direction. Top right images are k-space from a single channel for four individual slices, which are aliased slices in a multiband RF acquisition. Top left, the recombined k-space with an FOV extended by the number of RF bands. Bottom left, k-space for a single channel acquired with 4-fold undersampling in the PE direction and simultaneous acquisition of four slices with a multiband RF excitation. Bottom right, separation of acquired multiband k-space data into distinct slices, where the separation is performed with the ACS data from the top left k-space.

FIG. 3

Average from three subjects of the mean g-factors for 4-fold reduction in the PE direction, 4-fold aliasing in the slice direction, and 16-fold reduction in both PE and slice direction. The abscissa indicates the changes in the g-factor for different axial slices due to less aliasing and changes in sensitivity profiles.

FIG. 4

Comparison of temporal signal fluctuations between single-band and multiband acquisitions. a: Maps of the temporal fluctuations, where each pixel is the standard deviation through time, normalized by the mean signal. Left: a coronal slice from a multiband acquisition and (right) the same slice from a separate single-band acquisition. b: Histogram from three subjects, where each pixel from the signal band is compared with the same pixel from a multiband acquisition. The histogram is made with 80 equidistant bins utilizing all pixels in the brain/head, as determined using a signal intensity threshold from the single-band acquisition.

FIG. 5

Functional activation maps (t tests, filtered with a 2D cluster of four) from two sequential finger-tapping and checkerboard visual stimulation fMRI experiments. Top row, representative coronal slices from a single-band experiment with a reduction factor of 4 in the PE direction and a TR of 5 sec. Bottom row, same coronal slices from a multiband multislice experiment, acquired with a reduction factor of 4 in the PE direction, four RF bands, and a TR of 5 sec. In the multiband technique, TR was set to match the slower single-band acquisition and thus resulted in one-forth the duty cycle as compared with the single-band approach.

FIG. 6

a: Histograms of the normalized temporal fluctuations in each pixel (${\tilde{\sigma }}_t$), calculated experimentally from one study (the same as in Fig. 5), acquired with both multiband and single-band methods. b: Histograms of all the t scores in the brain (shown in Fig. 5) for both multiband and single-band acquisitions from an experiment with 2 × 2 × 2mm³ resolution and TR = 5 sec. The bulk random fluctuations are described by the bell-shaped distribution of t scores. c: Scatterplot of the t scores from the single-band versus multiband acquisitions in Fig. 5.

FIG. 7

Evaluation of correlated temporal fluctuations to residual unaliasing. a: Top: Four simultaneously excited and acquired slices. Bottom: signal change from four simultaneously acquired ROIs, chosen so that the ROI in the visual cortex (rightmost slice) shows strong stimulus-induced activation. The signal from the visual cortex is ~5 times larger than the signal in the aliased slices and is the only signal measured with the y-axis on the right of the plot. The signal changes in the ROIs in the visual cortex are ±0.05 (a.u.) and the mean signal in the aliased ROIs is 0.46, 0.28, and 0.56, respectively. b: Top: coronal slice with functional signal changes in a small ROI. Bottom, signal changes in the four aliased ROIs from 4-fold undersampling in the PE direction. The signal from the visual cortex is the only signal measured with the y-axis on the right of the plot.

FIG. 8

Functional activation maps based on cross-correlation and a 2D four-neighbor cluster, for a complex visuomotor dissociation task acquired for 88 slices in 1.25 sec with 2 × 2 × 2mm³ resolution. A total of 252 images were obtained with the subjects performing the task during three blocks of on-off periods. Inset: representative coronal slices from a functional scan with 1 × 1 × 2mm³ resolution and TR = 1.5 sec. The FOV was squeezed in the left-right direction, yielding an effective reduction and maximal aliasing of 16.