Purpose

To evaluate usage of short dipole antennas as receive (Rx) elements of a human head array at 9.4T, a 16-element prototype of a novel array combining transceiver (TxRx) surface loops and Rx-dipoles was developed, constructed and tested.

Introduction

Increasing the number of surface loops in a human head Rx-array improves the peripheral SNR, while SNR near the brain center doesn’t substantially change (1-3). Recent work on Ultimate Intrinsic SNR (UISNR) demonstrated that an optimal central SNR at UHF requires contribution of two current patterns associated with a combination of surface loops and dipole antennas (2-6). However, to incorporate multiple dipoles into a multi-row human head loop Rx-array, the dipole length has to be substantially reduced, which may compromise its performance. At the moment use of short dipoles for human head Rx-arrays has not been well studied. Another issue of using short Rx-dipoles is a sensitivity of their resonance frequency to loading. A large conservative electrical field near the dipole substantially shifts the resonance frequency with changing the distance to the load due to head size variation. This in turn compromises SNR. In this work, first, we optimized the design of a short Rx-only dipole element. Then we constructed and characterized a novel human head phased array consisted of 8 TxRx surface loops and 8 Rx-dipoles.

Methods

The load dependent frequency shift can be minimized by bending the dipole (Fig.1A) and moving its ends away from the object. This modification shouldn’t substantially affect the RF magnetic field, B₁, because ends of the antenna carry a small current. To evaluate the frequency shift due to sample size variation, we used two different in size cylindrical phantoms. We also constructed a straight 100-mm dipole (Fig.1B). The frequency shift measured for the folded dipole was about 2 times smaller than that measured for the straight dipole (Figs.1C and 1D). We numerically evaluated several folded dipoles as well as a vertical loop and a surface loop. We also evaluated SNR of single elements (Fig.2C) and a coupling between a pair of elements placed at 45º from each other. SNR was evaluated as B₁^-/√P, where P is an RF input power. Finally, we simulated SNR maps (Fig.2D) of several 8-element arrays placed on an elliptical holder (Fig.2B). Table 1 shows results of the entire evaluation. Optimal dipole geometry, i.e. 30-mm folded dipole, was chosen using following criteria: a small frequency shift, good decoupling, and high SNR at the phantom center. Based on optimization we constructed the final array (Figs.3A-D). We also compared the new array to the 16-channel array of similar size with 8 surface loops and 8 Rx-only vertical loops described previously (7). Electromagnetic (EM) simulations of the B₁ distribution and local specific absorption rate (SAR) were performed using CST Studio Suite 2015 (CST, Darmstadt, Germany) and the time-domain solver based on the finite-integration technique. Three voxel models were used, i.e. a head/shoulder (HS) phantom (ε=58.6, σ=0.64 S/m at 400 MHz), and two virtual family multi-tissue models, “Duke” and “Ella”. Experimental B₁⁺ maps were obtained using the AFI sequence (8). All data were acquired on a Siemens Magnetom 9.4T human imaging system.

Results and Discussion

While B₁ distribution of the straight dipole is symmetrical, the vertical loop B₁ map shows a substantial asymmetry (Fig.2C). Folded dipoles have small asymmetrical contribution increasing with an increase of the folded portion. Thus, folded dipoles provides a better approximation of “z-directed”, i.e. dipole-like, currents than vertical loops. As seen from Fig.3E, presence of actively detuned dipoles did not compromise the decoupling of TxRx surface loops. Also both arrays qualitatively demonstrated very similar B₁⁺ distributions (Fig.4). Quantitatively, experimental B₁⁺ averaged over the central 20-mm transversal slab measured 11.46±2.75 μT/√kW and 12.01±2.92 μT/√kW for the dipole and vertical loop arrays, respectively. Thus, our experimental and numerical evaluations demonstrated that the Tx-performance and SAR of the new array was not substantially altered by the presence of Rx-dipoles. Maximum increase of the SAR measured 7.1% for the Duke voxel model. At the same time the dipole array shows substantially higher (up to 2 times) SNR at the periphery and ~1.17 times higher SNR at the center (Fig.5C).

Conclusion

We evaluated usage of optimized 30-mm folded short dipole antennas as elements of a human head Rx-array. Folding the dipole allows decreasing the frequency shift due the head size variation. Addition of Rx-only dipoles doesn’t substantially alter B₁⁺ transmit field and the maximum local SAR. At the same time the new design improves both peripheral and central SNR as compared to the similar 16-element array with Rx-only vertical loops.

Acknowledgements

No acknowledgement found.

References

1) Wiggins GC et al. Magn Reson Med 2009;62:754–762. 2) Vaidya MV et al. Conc Magn Reson Part B 2014; 44B(3):53-65. 3) Lattanzi R, Wiggins GC, Zhang B, Duan Q et al. Magn Reson Med 2018;79(3):1789-1803. 4) Pfrommer A and Henning A. Magn Reson Med 2018;80(5):2122-2138. 5) Chen G, Lakshmanan K, Sodickson D, and Wiggins GC. Proc. ISMRM 2015, 3133. 6) Zhang B et al. Proc. ISMRM 2017, 4314. 7) Avdievich NI, Giapitzakis IA, and Henning A. NMR in BioMed 2018, 31(2):1-13. 8) Yarnykh VL. MRM 2007;57:192-200.