A 32-channel transmit/receive radiofrequency head coil for 7T UHF MR
Stefan HG Rietsch1,2, Stephan Orzada1, Sascha Brunheim1,2, Andreas K Bitz3,4, Maximilian N Voelker1,2, Viktor Pfaffenrot1,2, Marcel Gratz1,2, Daniel Leinweber1, Jonathan Weine1, Sarah Handtke1, Oliver Kraff1, Mark E Ladd1,3,5, Peter Koopmans1,2, and Harald H Quick1,2

1Erwin L. Hahn Institute for MR Imaging, University of Duisburg-Essen, Essen, Germany, 2High-Field and Hybrid MR Imaging, University of Duisburg-Essen, Essen, Germany, 3Medical Physics in Radiology, German Cancer Research Center (DKFZ), Heidelberg, Germany, 4Electromagnetic Theory and Applied Mathematics, Faculty of Electrical Engineering and Information Technology, University of Applied Sciences Aachen, Aachen, Germany, 5Faculty of Physics and Astronomy and Faculty of Medicine, University of Heidelberg, Heidelberg, Germany


In this work, a 32-channel transmit/receive (32Tx32Rx) head coil for 7T UHF MRI was developed and evaluated. The performance of this coil is compared to a commercially available 1Tx32Rx head coil regarding noise correlation, signal-to-noise ratio (SNR) and B1 homogeneity. The results indicate that high flexibility concerning RF shimming is provided by the 32Tx32Rx head coil that translates into a more homogeneous flip angle distribution compared to the 1Tx32Rx head coil. First 7T MR imaging results show the increased coverage of both the head and superior parts of the neck.


7T ultrahigh-field (UHF) MRI provides an increase in signal-to-noise ratio (SNR) (1). The increase of the Larmor frequency at 7T, however, requires multi-channel RF technology and methodology for homogeneous signal excitation. In this context, 32 independent RF transmit channels (2) offer the possibility to acquire homogeneous images at 7T by employing parallel transmit techniques (pTx) (3,4). In this work, a 32-channel transmit/receive (32Tx32Rx) head coil to cover both head and neck was developed and evaluated. First results are presented, demonstrating high degrees of freedom for RF shimming as well as promising anatomical coverage in 7T neuro MRI applications.


The 32Tx32Rx head coil consists of an 8-channel anterior part (Figure 1A) and a 24-channel posterior part (Figure 1B) and features 32 Tx/Rx loop elements. The anterior part can be opened to facilitate patient positioning (Figure 1C,D). The coil housing is a combination of 3D-printed components that fit the head geometry and flat polycarbonate sheets (Figure 1D).

Each of the transceiver loop elements of the 32Tx32Rx coil is connected to a custom-built transmit/receive switch (Figure 2A-C). This transmit/receive switch consists of a low-noise preamplifier (Wantcom, Chanhassen, MN, USA), which is (in addition to the internal protection) protected by a PIN-diode and a lambda-over-four transmission line (Figure 2A,B).

Eight transmit/receive switches are housed in an independent polycarbonate box (Figure 2C). Altogether, 4 of these transmit/receive boxes are used to connect the 32-channel Tx/Rx coil to a custom-built transmit chain add-on (2) and to the receive channels of the MR system. The connection on the receive side (Figure 2C) of each board is realized using TIM cables (Total Imaging Matrix, Siemens Healthcare GmbH, Erlangen, Germany). The transmit/receive switches can also be used with any other transmit/receive coil at 7T.

All MR measurements were acquired on a 7T whole-body MRI system (Magnetom 7T, Siemens Healthcare GmbH, Erlangen, Germany). The performance of the 32Tx32Rx head coil was compared to a commercial 1Tx32Rx head coil with a single transmit-only birdcage and 32 receive-only loops (Nova Medical, Wilmington, MA, USA). For the 32Tx32Rx coil, B1+ mapping was accomplished using the B1TIAMO approach by Brunheim et al. (5).

Results and Discussion

The noise correlation coefficients of the two RF head coils loaded with a homogeneous head phantom (Figure 3A,B) indicate that the receive channels of the 1Tx32Rx head coil (Figure 3A) primarily show increased correlation between neighboring channels, while the 32Tx32Rx head coil (Figure 3B) also shows enhanced correlation between non-neighboring elements.

The flip angle distributions (Figure 4A,C) demonstrate the obstacles for a comparison between a 1-channel transmit birdcage (Figure 4A) and a 32-channel TxRx coil (Figure 4C). The transmit voltage was adjusted such that comparable flip angle is achieved in the entire transversal slice (white dotted circle in Figure 4C) for both coils (Figure 4A,C). While the RF shim for the 32Tx32Rx head coil was successfully optimized for homogeneity in the entire transversal slice, the CP+ excitation by the commercial 1Tx32Rx head coil (Figure 4A) is very inhomogeneous (central brightening demonstrated by the black ROI). This can be compared quantitatively by the standard deviation of the flip angle in the black, blue and white dotted ROIs (Figure 4A,C). In the blue ROI (Figure 4A,C) comparable mean flip angles could be achieved, which leads to comparable mean SNR in the white central ROIs (Figure 4B,D). The evaluation of the entire transversal slice (white dotted circles in Figure 4B,D) results in a higher SNR of 104.4 for the 32Tx32Rx coil compared to 83.0 for the 1Tx32Rx coil. This corresponds to an SNR increase of 26% for the 32Tx32Rx coil.

First MR imaging results in a melon (head) and stack of pineapple slices (neck) at 7T in transversal (Figure 5A), sagittal (Figure 5B), and coronal (Figure 5C) orientation were acquired using a 3D FLASH gradient echo sequence with 1 mm isotropic resolution, a TR/TE of 12/4.08 ms, and a total acquisition time of TA = 3 min 22 sec. These first experiments demonstrate the large field-of-view the 32Tx32Rx coil can cover.


The presented 32Tx32Rx head coil shows quite homogeneous excitation in preliminary investigations with static RF shimming and, depending on the ROI used for the evaluation, increased SNR in comparison to a commercial 1Tx32Rx coil. First MR imaging results in phantoms and fruits indicate the coverage of the human head as well as the neck. In-vivo brain/neck imaging and evaluation will start as soon as the ongoing safety evaluation of the RF coil is completed. Further steps will include the evaluation of more advanced pTx techniques like selective volume excitation.


The research leading to these results has received funding from the European Research Council under the European Union's Seventh Framework Programme (FP/2007-2013) / ERC Grant Agreement n. 291903 MRexcite, and DFG grant KO5341/1-1


1. Pohmann R, Speck O, Scheffler K. Signal-to-Noise Ratio and MR Tissue Parameters in Human Brain Imaging at 3, 7, and 9.4 Tesla Using Current Receive Coil Arrays. Magn Reson Med 2016;75(2):1–9.

2. Orzada S, Bitz AK, Gratz M, Johst S, Shooshtray S, Völker MN, Rietsch SH, Flöser M, Abuelhaija A, Oehmigen M, et al. A 32-channel transmit system add-on for 7 Tesla body imaging. Proc Intl Soc MRM 25 2017. p. 1219.

3. Mao W, Smith MB, Collins CM. Exploring the limits of RF shimming for high-field MRI of the human head. Magn Reson Med 2006;56(4):918–22.

4. Katscher U, Börnert P, Leussler C, van den Brink JS. Transmit SENSE. Magn Reson Med 2003;49(1):144–50.

5. Brunheim S, Gratz M, Johst S, Bitz AK, Fiedler TM, Ladd ME, Quick HH, Orzada S. Fast and accurate multi-channel B1+ mapping based on the TIAMO technique for 7T UHF body MRI. Magn Reson Med 2018;79(5):2652–64.


Figure 1: The presented 32-channel transmit/receive coil consists of an anterior part with eight loop elements (A) and a posterior part with 24 loop elements (B). For easy patient positioning the anterior part can be moved along z-direction. The connection between transmit/receive switches and loops is accomplished by semi-rigid coaxial cables and cable traps (B). The coil housing (C,D) is a combination of 3D-printed parts (white) and polycarbonate sheets (transparent).

Figure 2: Photograph of a single custom-built transmit/receive switch (A) with preamplifier and circuitry. The schematic (B) indicates the separation of transmit and receive path. Eight transmit/receive switches are grouped in a polycarbonate box (C). The connection to the transmit chain and to the coil is accomplished via BNC connectors, while the receive side is connected to the MR system using MR system-specific TIM cables (C).

Figure 3: Noise correlation coefficients of the 1Tx32Rx head coil (A) shows that neighboring receive-only channels show relatively high correlation. For the 32Tx32Rx head coil (B) there is enhanced correlation between non-neighboring channels. The anterior (channels 1-8) and the posterior (channels 9-32) parts show very low noise correlation (B). Please note that channel #7 of the anterior part was not working in the presented results (B).

Figure 4: Transversal phantom measurements comparing the 1Tx32Rx coil (A,B) and the 32Tx32Rx coil (C,D). Comparable mean (ø) flip angles (A,C) are achieved in the entire slice (white dotted circle) and in the blue ROI. Consequently, the mean SNR in the white central ROI (B,D) for both coils is also approximately comparable. Higher SNR is achieved with the 32Tx32Rx coil when the entire slice is evaluated (B,D). Please note the clearly improved homogeneity of the dedicated RF shim (C) provided by the 32Tx32Rx head coil compared to the very inhomogeneous excitation of the CP+ mode generated by the birdcage of the 1Tx32Rx coil (A).

Figure 5: Demonstration of field-of-view capabilities. First MR imaging results using the 32Tx32Rx head coil loaded with a melon (head) and a stack of pineapple slices (neck) in transversal (A), sagittal (B) and coronal (C) orientation. The coil allows to cover 26 cm in z-direction (C). Consequently, with this RF coil, the neck region can be covered as indicated by the pineapple slices in sagittal (B) and coronal (C) orientation.

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