Fig 1.
Illustration of interslice MT effects.
The application of a gradient varies the Larmor frequency f(z) linearly in space (z). During excitation, the slice of interest (slice 0) receives on resonance excitation. With a positive gradient polarity and descending slice order (shown above), the next slice to be acquired (slice −1) receives off-resonance irradiation at a frequency offsets of 3840 Hz.
Table 1.
Two-Pool MT Parameters at 3T.
Fig 2.
MTR images of a 10% agar phantom (left) and saline phantom (right).
Fig 3.
Comparison of MTR images generated with interslice MT effects and with presaturation.
Offset irradiation frequencies of the presaturation pulses corresponded to the offeset frequencies of the first (+3200 Hz), second (+6400 Hz), and third (+9600 Hz) prior slices of the interslice method. Average RF-power of saturation was equivalent in both methods. Baseline (MT-weighted) and MTR images (a) from a representative subject are shown. Both MTR (b) and SNR (c) were calculated for white matter. Error bars show the 95% confidence interval of the group average.
Fig 4.
Center slices of MTR images from a representative subject are shown for varying flip angles and for linear phase encoding (top) and centric phase encoding (bottom).
Fig 5.
Mean MTR values across subjects from regions of interest for white (a) and gray matter (b).
Predicted values from simulating the two-pool model (solid lines) with parameters from the literature show close agreement with the in vivo values. Centric phase encoding shows substantially better SNR (c) than linear phase encoding. Error bars show the 95% confidence interval of the group average.
Fig 6.
Saturation of the longitudinal magnetization accumulates over multiple prior slices with the majority of saturation due to the first prior slice.
For white (a) and gray (b) matter, simulations show the longitudinal magnetization as a function of the number of prior slices for varying number of phase encoding steps per slice and for varying flip angles. One prior slice = saturation at 3840 Hz, two prior slices = saturation at 7680 Hz followed by 3840 Hz, three prior slices = saturation at 11520 Hz, 7680 Hz, and 3840 Hz, etc.
Fig 7.
Simulated MTR values for white (solid line) and gray matter (dashed line) for varying interslice delay time for reference image acquisition.
Simulations were performed using sequence parameters that matched the bSSFP acquision for images in Fig. 8.
Fig 8.
Comparison of interslice MTR imaging with bSSFP and SSFP-FID sequences.
The SSFP-FID sequence significantly reduced banding artifacts in slices 3–5, but SNR was 22% lower than with bSSFP.
Fig 9.
Interslice MTR images of a brain tumor (meningioma).
Distinct signal characteristics in the MTR images were visible in the brain tumor regions.
Fig 10.
Interslice MTR images with short interslice delay times.
Matrix size = 128 × 128, FA = 60°, TR/TE = 4.15/2.08 ms, RF-pulse duration = 1.24 ms, slice thickness = 5 mm, number of slices = 24 (excluding 6 dummy slices), total scan time = 56 s (0.7 s delay) and 87 (2.0 s delay). Two scans (each with 12 number of slices excluding 3 dummy slices) were spatially interleaved, in order to provide near whole brain coverage with no gap.