Fig 1.
Acquisition strategies for the phantom and human experiments.
Phantom: A) Relative positioning of the slices (s0 to s5) across multiband (MB) factors. B) Slice shifting along z with increasing interslice gap. Note that the position of slice s3 (dashed green line) was kept fixed for all interslice gaps and MB factors. C) Adopted current (I) “injection” schemes: no currents (“noise floor” measurements) and 2 mA baseline-to-peak currents with alternating polarity (“MRCDI” measurements). D) Graphical summary of the experiment: for each interslice gap, a set of multiband factors was tested and compared to single-slice measurements of the corresponding slices. Real ΔBz,c measurements obtained with MB factors 3 and 6 and gaps of 6 and 18 mm are given as examples for each of the two current “injection” schemes. Human: E) Relative positioning of the slices (s0 to s4) across multiband factors. F) Slice shifting along z with increasing interslice gap. Note that slice s2 (dashed green line) was the one kept fixed this time. G) Illustration of how the cable loop was placed around subjects’ heads. The current “injection” schemes depicted in C) were also used in the human experiments, and the adopted strategy was similar to the one portrayed in D).
Fig 2.
Sensitivity of the ΔBz,c noise floors to different multiband (MB) factors and interslice gaps.
The relative position of the slices is shown in Fig 1A and 1B. A) Summary of the noise levels in the ΔBz,c measurements, as estimated from the spatial standard deviation of the ΔBz,c images obtained without current injection. Results are shown for SMS acquisitions with different MB factors and interslice gaps, and corresponding single-slice (SS) measurements. The SMS acquisitions are represented by data points connected by lines, while individual data points depict the SS results. Note that the slice pairs (s0,s3), (s1,s4) and (s2,s5) shared the same CAIPI shifts in the SMS acquisitions, and therefore the respective data points exhibit the same marker style. B) Magnitude and ΔBz,c images of the fixed slice (s3) obtained with a MB factor of 6 and different interslice gaps. Subtle artifacts are noticeable in the magnitude images for the smallest gap and are indicated by red arrows.
Fig 3.
Measurements of ΔBz,c induced by a 2-mA current loop around a phantom.
A) Root Mean Square (RMS) error between the measured (ΔBz,c2mA) and simulated (ΔBz,csim) current-induced magnetic fields. Results are shown for multi-slice measurements with different MB factors and interslice gaps, and the corresponding single-slice (SS) measurements. The relative position of the slices is shown in Fig 1A and 1B. Please note that as the interslice gap increased, the outermost slices got closer to the lead, where the current-induced magnetic fields were stronger, also causing increased RMS errors. B) Example of the simulated and measured current-induced magnetic fields, and respective differences for an acquisition with MB factor 6 and interslice gap of 18 mm. Slices sharing the same CAIPI-shift are shown side by side. The black arrows in s2 and s3 highlight the presence of subtle artifacts, likely due to signal leaking from s5 and s0, respectively.
Fig 4.
Sensitivity of the ΔBz,c noise floor measurements in the human brain to different multiband (MB) factors and interslice gaps.
The relative position of the slices is shown in Fig 1E and 1F. A) Summary of the noise levels in the ΔBz,c measurements, as estimated from the spatial standard deviation of the ΔBz,c images obtained without current injection. Results are shown for multi-slice measurements with different MB factors and interslice gaps, and the corresponding single-slice (SS) measurements. Each data point and vertical error bar represent the average and standard deviation across five subjects. B) ΔBz,c noise floor images from an SMS acquisition with a MB factor of 5 and an interslice gap of 12 mm (first row), and corresponding single-slice acquisitions (second row).
Fig 5.
Results from the ΔBz,c measurements in the human brain for a 2-mA current loop around the head.
The relative position of the slices and loop placement are shown in Fig 1E-1G. A) Root Mean Square (RMS) error between the measured (ΔBz,c2mA) and simulated (ΔBz,csim) current-induced magnetic fields. Results are shown for multi-slice measurements with different multiband (MB) factors and interslice gaps, and the corresponding single-slice (SS) measurements. Each data point and vertical error bar represent the average and standard deviation across subjects. B) Comparison between a multi-slice acquisition with a MB factor of 5 and an interslice gap of 12 mm, and corresponding single-slice acquisitions. Left: simulated and measured magnetic fields. Right: differences between the measured and simulated fields.