Introduction

Due to the well-known wavelength reduction at 7T, birdcage head coils suffer from B₁⁺ inhomogeneities that translate to higher signal at the center of the image and lower signal in the area close to the coil. As a simpler alternative to active B₁⁺ shimming (pTx^1,2), passive RF shimming approaches were proposed using passive elements such as dielectric pads^3-5 and metamaterials^6,7 to control B₁⁺ within the head coils. We previously proposed a meander dipole (MD) based on two 3^rd-order Hilbert fractals connected in series as a promising shimming solution⁸. Although they enhanced B₁⁺ in the region of interest, the effect was very local and moderate in amplitude. In this work, we present a new design of meander dipole based on six 2^nd-order Hilbert fractals connected in series to improve on these limitations. The new meander dipoles were numerically studied and experimentally tested in vivo.

Material and Methods

Six 2^nd-order Hilbert fractal metallic strips connected in series were designed (Figure 1). A cross pattern was added between the fractals to decrease their quality factor. The copper structures were simulated in a 16-leg birdcage coil and in presence of a homogeneous anthropomorphic phantom of permittivity ɛ= 45.3 and conductivity σ= 0.87 S/m. Their effect on B₁⁺ was compared with that of the 3^rd-order MDs proposed earlier. They were printed on a 0.5-mm-thick Rogers 3003 substrate and tested in vivo in a Tx/Rx proton head coil (Invivo Corp., Gainesville, FL) in a 7T Magnetom MRI scanner (Siemens Healthineers, Erlangen, Germany) on one healthy volunteer. Informed consent was obtained from the volunteer prior the scan. Two MDs were placed at 1 cm distance on both lateral sides of the head (Figure 2). We measured B₁⁺ maps using the XFL MRI sequence⁹ ( TR=20s, TE=3.06ms, FA=4°, Shinnar-Leroux-minimum-phase-saturation-pulse=60°, FOV=256mm, matrix=64x64 and 3 slices of thickness=4mm). One central slice was acquired in each of the three main orientations. Three GRE image slices were acquired with the same geometry (parameters TR=20s, TE=3.06ms, FA=3°, FOV=256mm, matrix=128x128, slice thickness=4mm). All images were acquired using the same reference voltage of 260V.

Results

Figure 3 shows the numerical B₁⁺ maps with the birdcage coil alone (top), with the earlier proposed 3^rd-order MDs (center) and with the 2^nd-order MDs (bottom). The 2^nd-order MDs (gray rectangles) produced a stronger B₁⁺ enhancement in the temporal lobes than the 3^rd-order MDs (green rectangles), visible in the coronal and axial slices. However, the B₁⁺ gain in this region was paired with strong losses in other regions. A comparable field distribution was confirmed by the B₁⁺ maps measured in vivo (Figure 4). The third row shows the relative B₁⁺ change when compared with the reference. The field was enhanced by about 60% in the temporal region close to the MDs (white rectangles), unchanged in the center and reduced by about 60% in frontal, posterior and top regions.The in vivo GRE images show a similar distribution (Figure 5) with an almost complete signal loss in certain regions.

Discussion

The 3^rd-order Hilbert fractal design proposed earlier allowed us to produce compact structures, which, however, compromised the effect of the MD as a secondary B₁⁺ field source. To regain the effect, we required a more elongated structure which was obtained with the 2^nd-order Hilbert fractals. The resulting structures had the double length (18.2 cm) but were much shorter than a straight electric dipole used for 7T applications (approx. 40 cm). The chosen substrate made it possible to obtain thin, compact, and flexible resonant structures. By adding the cross pattern, we were able to decrease the quality factor. The obtained meander dipoles showed a 60% higher B₁⁺ in the area between them, (Figure 4), doubling the gain achieved with the earlier-proposed 3^rd-order meander dipoles⁸. Nevertheless, the new structures also led to strong B₁⁺ reductions in other regions, which was not the case for the former (Figure 3). Complete loss of GRE image signal was observed in anterior, posterior, and top regions of the brain (Figure 5), similar to other RF shimming solutions^6,7. The 2^nd-order meander dipoles could therefore be seen as spatial lenses, focusing the field on a specific region.

Conclusion

We proposed a new design of meander dipoles based on composite 2^nd-order Hilbert fractals with the aim of reducing B₁⁺ inhomogeneities present in a 7T birdcage coil. Thanks to their compact and flexible design they were easy to handle and could be placed between the head and the RF coil. These passive elements were able to increase the B₁⁺ field by up to 60% in the temporal lobes. However, strong signal loss was observed in other regions. Further studies and optimization of these structures are required to reach the initial goal of enhancing multiple regions without losing efficiency of the coil in other areas.

Acknowledgements

This project has received funding from the European Union's Horizon 2020 research and innovation programme under grant agreement No 736937 first (M-CUBE project), and then under grant agreement No 952106 (M-ONE project), and from the Excellence Initiative of Aix-Marseille University - A*MIDEX, a french "Investissements d'Avenir" programme under Multiwave chair of Medical Imaging.