Homogenous 64-channel RF transmit array for brain imaging at 7T, 9.4T, and 10.5T
Tales Santini1, Sossena Wood1, Howard J. Aizenstein1, and Tamer S. Ibrahim1

1University of Pittsburgh, Pittsburgh, PA, United States


In this work, a 64-channel Tic-Tac-Toe (TTT) transmits array is simulated (with a realistic tuning and matching method) for three field strengths. Phase-only RF shimming was performed to access the array performance. In the whole brain excluding the cerebellum, the coefficient of variation (CV)=15.6%/20.9%/25.0%, Max/Min=3.17/6.29/5.84, for an average B1+=7.47/5.34/3.78μT/√kW at 7T/9.4T/10.5T field strengths, respectively. With these shimming cases, the array presented an average SAR=1.67/2.56/2.72 W/kg for 2μT and SAR efficiency=1.55/1.25/1.21μT/√(W/kg) at 7T/9.4T/10.5T, respectively. The peak SAR=5.35/9.49/10.76W/kg for 2μT at 7T/9.4T/10.5T, respectively. The 64-channel TTT array is capable of producing homogenous B1+ and low levels of SAR for UHF human MRI.


As 7T MRI is going towards clinical applications with the FDA clearance of the MAGNETOM Terra, there is a grown interest in higher fields strengths. The advantages of higher signal-to-noise ratio and image contrast is hinder by the RF fields inhomogeneities and elevated SAR. In this work, we explore a 64-channel Tic-Tac-Toe (TTT) transmit array performance at 7T, 9.4T, and 10.5T.


The RF transmit array is composed of 16 Tic-Tac-Toe panels, totaling 64 channels (Figure 1). Each Tic-Tac-Toe coil is composed of 8 square shape transmission lines connected to each other, being 4 of these transmission lines used as excitation ports and the other 4 used to tune the RF coil.

The simulations were conducted using an in-house developed and verified finite difference time domain (FDTD) code, with an accurate transmission line modeling for tuning and matching. The coils were initially tuned to 297MHz (7T frequency) and then retuned to 9.4T and 10.5T by adjusting the dielectric of the coil struts. The Duke model was used with its permittivity/conductivity adjusted to the frequency of the three B0 fields analyzed.

Phase-only optimizations were performed using the MATLAB function “fmincon”. The regions of interest in the optimization were: 1) brain-only excluding the cerebellum for 9.4T and 10.5T fields; 2) whole-head from the cerebellum and excluding the nasal cavities for the 7T fields. The statistics were calculated using the brain only excluding the cerebellum mask for all three fields. The cost function is the standard deviation over the minimum of the signal in the ROI. The SAR levels were also reduced using constraints. The mean B1+ was constrained to guarantee inversion with 8/16/32kW power amplifiers for the fields 7/9.4/10.5T respectively, with 50% of power loss until the coil.


Figures 2, 3, and 4 show the simulations results for 7T, 9.4T, and 10.5T respectively. In terms of B1+, the array presents an efficiency of 7.47/5.34/3.78 μT/√kW at 7T/9.4T/10.5T, respectively, and its homogeneity is CV = 15.6%/20.9%/25.0% and Max/Min = 3.17/6.29/5.84 at 7T/9.4T/10.5T, respectively. For the SAR, the array presents an average SAR of 1.67/2.56/2.72 W/kg for 2 μT and SAR efficiency of 1.55/1.25/1.21 μT/√(W/kg) at 7T/9.4T/10.5T, respectively. The peak SAR is 5.35/9.49/10.76 W/kg for 2 μT at 7T/9.4T/10.5T, respectively.


Previous work has shown the homogeneity of the Tic-Tac-Toe design for the 7T MRI1,2,3,4. Although the shorter wavelength at UHF presents significant challenges due to the RF inhomogeneities, the presented array was able to deliver homogeneous B1+ at all 3 field strengths simulated. Considering the losses between the power amplifiers and the coil plugs, the array presented in this work needs 8/16/32kW power amplifiers at the fields 7T/9.4T/10.5T, respectively to have inversion with 1ms square pulses. Although relatively inefficient, the array can produce inversion with short pulses and presents low levels of SAR for 2 μT and high SAR efficiency in μT/√(W/kg).


This work was supported by the National Institutes of Health under award number R01MH111265 and by the CAPES Foundation, Ministry of Education of Brazil, 13385/13-5


[1] Santini, T., et al. (2017). “64-channel Double-Octagon Tx Head Coil for 7T Imaging.” In Proc. of the 25th International Society of Magnetic Resonance in Medicine Annual Meeting, Honolulu, Hawaii, USA.

[2] Ibrahim, T., et al. (2017). “Towards Homogenous 7T Neuro Imaging: Findings and Comparisons between 7T TTT and NOVA RF Coil Systems.” In Proc. of the 25th International Society of Magnetic Resonance in Medicine Annual Meeting, Honolulu, Hawaii, USA.

[3] Kim, J., et al. (2016). "Experimental and numerical analysis of B1+ field and SAR with a new transmit array design for 7T breast MRI." J Magn Reson 269: 55-64.

[4] Santini, T.,et al. (2018). "In-vivo and numerical analysis of the eigenmodes produced by a multi-level Tic-Tac-Toe head transmit array for 7 Tesla MRI." In press. PloS one.


Figure 1: 64-channel TTT RF transmit array design. In 1), the FDTD model of the coil loaded with the Duke model. In b), on TTT coil assembled, showing it's 4 channels

Figure 2: RF shimming results for the RF array operating at 7T proton frequency

Figure 3: RF shimming results for the RF array operating at 9.4T proton frequency

Figure 4: RF shimming results for the RF array operating at 10.5T proton frequency

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)