Purpose

The last decade has seen the widespread application of phased array coils. This is largely due to their superior combined SNR performance, compared to single-channel implementations¹, and due to the fact that their independent elements are sufficiently distinct for spatial encoding, which addresses parallel imaging applications. However, the use of these coils presents a number of drawbacks, including the handling of receive antennas with bulky cable bundles dangling from one side, and safety issues arising from shield currents. One possible way to address these problems is to transmit the receive signals wirelessly, and recently, a number of designs which implement wireless receive data transmission for MRI have been published^2,3,4. Here we investigate the feasibility of using existing high data rate wireless transmission protocols - a wireless gigabit (WiGig) - suitable for transmitting data from high channel count array antennas. Moreover, the study aims to use off-the-shelf commercial components that, with only minor modifications, allow for maximum flexibility of the implementation. The system presented here overcomes these shortcomings by increasing the antenna gain and provides a practical solution for the wireless transmission of receive signals.

Methods

Two WiGig USB dongles (Mr.Loop, Taiwan) were used for data transmission inside the MR scanner room. The WiGig receive dongle was connected to a laptop computer located at the outer circumference of the MRI scanner room. An ODROID XU4 single PCB computer (Hardkernel, Korea) was used as the host PC and transmitted via the on-board USB 3.0 interface. For shielding, the single PCB computer was placed inside a custom built box (Fig. 1(a)). Fig. 1(b) shows the location of the WiGig dongles inside the MR room. To evaluate the achievable data rate, a large file (840 MB) was transferred via WiGig, and the transmission duration was timed. The procedure was repeated for different distances (d) between the transmitter and the receiver in a line of sight configuration. The effect of the WiGig transmission on the SNR of the MRI scanner was investigated when operating/not operating the WiGig dongle. All scans were performed on a Magnetom Tim Trio 3T scanner (Siemens Healthcare, Germany) using a 32 channel receive array coil. SNR measurements used a 170 mm diameter spherical water phantom doped with NiSO4 x 6H2O, and a gradient echo (GRE) sequence. The imaging parameters were: TR = 40 ms, TE = 3.84 ms, number of slices = 1, slice thickness = 5 mm, matrix size = 128 x 128, BW = 260 Hz/Px, FA = 25 deg. Initial investigations showed that the raw data rate achieved with the system did not reach the desired data rate. Thus, a simple solution was sought that enabled transmit capabilities to be increased. The WiGig dongle used has four antennas (Fig. 2(a)) and the antenna gain with beamforming is about 10 dBi. This can be increased with the self-shielded method⁵, in which two metal plates are located on the top and bottom side of the beamforming antenna array, respectively (Fig. 2(b)). By changing the gap between the antenna and the metal plates (Fig. 2(c)), the antenna gain can be changed. We used CST Microwave Studio (Darmstadt, Germany), to analyse antenna gain and to optimise shielding distance.

Results

The simulated array antenna gain of the unmodified WiGig dongle is approximately 10 dBi. With the shields in place, the antenna gain varies with the distance of the gap g. With g=4.5 mm the antenna gain is increased to around 14 dBi. Table. 1 shows measured data throughput in the MRI room as a function of communication distance for the unshielded and shielded antenna. The maximum transmission rate for a maximum distance of 3 m was above 650 Mbps in both cases. At a distance greater than 3.5 m, the transmission link broke down in the unshielded configuration. However, when using the shielded antenna array, the data rate required to sustain high channel MR receive antennas was sustained up to a distance of 4 m. Fig. 3 shows the acquired MR image when the WiGig system was transmitting inside the scanner room. Also, the transmission of RF pulses did not interfere with the wireless data transfer and no reduction in data rate during MR operation could be observed.

Conclusions

We have successfully implemented and tested an off-the-shelf wireless transmission system capable of sustaining the data rate required for high channel-count MRI receive arrays. There were no interferences when operating either system with the performance of the other. The system presented could be a 'drop-in solution' to reduce cable handling requirements and potential safety hazards in standard MRI scanners.

Acknowledgements

We would like to thank Claire Rick for proofreading, and Annette Weber for her assistance in 3D printing used in this study.