Designing a multichannel TMS/MRI system for 3 T: a 7-channel RF receive-only coil array prototype
Lucia I. Navarro de Lara1,2, Anthony Mascarenas3, Douglas Paulson3, Sergey Makarov4, Jason P. Stockmann1,2, Lawrence L. Wald1,2, and Aapo Nummenmaa1,2

1Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Charlestown, MA, United States, 2Radiology, Harvard Medical School, Boston, MA, United States, 3Tristan Technologies, San Diego, CA, United States, 4Department of Electrical and Computer Engineering, Worcester Polytechnic Institute, Worcester, MA, United States


An integrated multichannel TMS/MRI head coil array for 3T is currently under development to enable electronically controlled multifocal TMS with concurrent whole-head fMRI. To test the feasibility of the proposed RF hardware design, a 7-channel RF receive-only coil array prototype was built. Calculated B+1 maps showed the attenuating effect of the TMS coil on the transmit field. The improvement by retuning the affected RF loops was demonstrated using MPRAGE images. Functional images showed no additional artefacts when TMS pulses were interleaved between the EPI volumes. The prototype results support our RF hardware design approach for the TMS/MRI system.


We built a 7-channel RF receive-only coil array to assess the feasibility of the RF hardware design for an integrated multichannel TMS/MRI system for 3T.


The combination of Transcranial Magnetic Stimulation (TMS) with functional magnetic resonance imaging (fMRI) was demonstrated already in 1999(1). The most important challenge of this technique is accurately placing the stimulation coil and ensuring the positioning over the entire experiment. Multichannel TMS(2) is an emerging technology that allows multiple sites to be stimulated simultaneously or sequentially under electronic control without physical movement of the coils. This approach combined with whole-head MRI acquisition could make this challenging method to become practically usable for a wide range of neuroscience and clinical applications. To achieve this goal, a multi-TMS/MRI head coil array for 3T is currently being designed (see Figure1A).

To develop this new tool for concurrent brain imaging and stimulation, the influence of the multichannel TMS system on the MR image quality has to be carefully investigated. Effects on the transmit field of the 3T volume coil were previously assessed based on EM simulations(4). On the MRI hardware side, the RF design planned to be implemented must be tested. In this study, a 7-channel RF receive-only coil array was built as a proof of concept for the planned system. A single TMS element prototype fabricated by the manufacturer (Tristan, US) was employed for the study (see Figure1C).


Copper loops (6 cm diameter) were distributed on a hexagonal layout over a Plexiglas cylindrical form without overlapping. The support had 5cm diameter holes inside the RF loops to allow TMS element placement through the RF coil array as planned for the final design. Loops were matched (50 Ω) and tuned. Transformer decoupling was implemented between each element. Preamplifier decoupling was additionally implemented to reduce non-neighbor mutual induction. To guarantee decoupling of the loop elements to the body coil during the excitation, active detuning was implemented. The TMS coil prototype was placed through the central RF loop, shifting the resonance peak.

B+1 measurements using the body coil of the Skyra Connectome (Siemens, Germany) were done to assess B+1 artefacts from TMS or MR receive coil array. All four setups combinations were measured, two of these are shown in the bottom of Figure3. Gradient Echo (GRE) images at different flip angles (FA) (TR=3000ms, TE=10 ms, FA=20, 50, 70 and 90,80 slices, 3mm isotropic-voxels, MA=64x64) were acquired for each setup to calculate B+1 maps.B+1 map calculation and ROI-based artefact evaluation were done using in-house written MATLAB scripts.

To evaluate the detuning effect of the TMS coil on the RF loop for anatomical imaging, MPRAGE images were acquired (TR=2530ms, TI=1100ms, TE=1.6ms, FOV=256mm, MA=256x256, SL=208). Three setups were investigated: (i) noTMS, (ii) TMS and no retune and (iii) TMS and retune. ROI-based sensitivity decay was evaluated using in-house MATLAB scripts.

To investigate effects on the functional imaging, echo planar images (EPI) (TR=2800ms, TE=20ms, 27 slices, 3 mm isotropic voxels) were acquired and stimulation pulses were given in the imaging gap as shown in Figure5. For all experiments, a cylindrical Siemens phantom shown in Figure2 was used.


Figure2 shows the constructed 7-channel RF receive-only coil array. The calculated B+1 map through the main sagittal and coronal slices are shown in Figure3. Mean effects of the TMS coil on the B over the ROIs drawn in the figures are listed in the table of Figure3. Reference image was always the no TMS case. Figure4 shows MPRAGE images acquired for the three setups. The table in the figure summarizes the mean signal loss on the ROIs depicted in each case. The reference image was always the no TMS image. The EPIs images acquired are shown in Figure5. There were no artefacts visible due to the stimulation pulse interleaved with the EPI acquisition. Susceptibility effects produced mainly by the coil housing were stable over the stimulation protocol.


Artefacts seen on MPRAGE images are produced mainly by interactions between TMS coil and the B+1, as can be seen from the corresponding maps. These effects decay fast with depth and become subtle at depths of 1.5cm or greater. Retuning the RF coil elements that will contain TMS coils will be critical for optimizing the performance of the final coil array. Stimulation pulses do not significantly deteriorate the quality of the acquired EPIs as long as they are applied in an interleaved fashion between two consecutive volumes. We confirmed that the proposed hardware design is well suited to be used for the construction of the multi-TMS/MR head coil array for 3T.


This work was funded by NIH R00EB015445, R01MH111829, NIH R00EB021349.


(1) Bohning et al.,Invest Radiol,33(6):336-340,1998

(2) Ruohonen J and Ilmoniemi R. Medical and Biological Engineering and Computing, Vol. 36 p297-301,1998

(3) Makarov et al., IEEE Trans Biomed Eng., ahead of print, 2018

(4) Navarro de Lara et al., Proceedings EMBC 2018, Honolulu, HI, US


Figure 1. A) 3D-CAD model of the multichannel TMS MR coil to be designed. B) Top. 3-axis TMS coil alone and in array configuration. The geometry was selected to increase the degrees of freedoms to synthesize the E-field using a multichannel approach – the basic idea is illustrated using high-resolution anatomically realistic TMS modeling (3). Red coils show which coils are activated in each case. C) First 3-axis TMS coil prototype.

Figure 2. Photograph of the 7-channel receive-only coil array with a TMS coil prototype. Coil support has holes to position the TMS prototype through the RF coil in close proximity of the phantom. RF coil loops are decoupled using transformer decoupling. Preamplifiers are placed far from the RF loops due to space constraints and to minimize the possible TMS/preamplifier interactions.

Figure 3. Measured B1+ fields based on GRE images. Left) Coronal slices through the middle of the TMS coil. Top maps are based on data acquired without placing the RF coil array over the phantom. Bottom maps are based on data acquired placing the RF coil array over the phantom. On the top of the phantom, the TMS coil artefact can be seen. The ROI is drawn in blue. Right) Sagittal slice through the TMS coil.

Figure 4. MPRAGE images. Top) Images acquired without placing the TMS coil through the loop. Center) Images acquired with the TMS coil in the center through the RF coil but without retuning the affected RF loop. Bottom) Images acquired with the TMS coil in the center through the RF coil and retuning the affected RF loop. Red arrows show TMS artefacts due to B+1 interactions. Blue arrows show improvement in the image due to retuning of the RF loop.

Figure 5. EPI data before and after TMS stimulation. The images show only the susceptibility effect produced by the 3-axis TMS coil housing which remains the constant during the stimulation protocol. Images do not show any prominent artifact due to the stimulation pulse as long as the MRI acquisition is interleaved with the TMS administration.

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