Implementation of a Novel 8-Ch Phase-Array Transmit/Receive Head Coil with RF Interface for Parallel Transmission on 3T

INTRODUCTION High and/or ultra-high field (UHF) whole-body MR spectrometer emerges as a cutting-edge new millennium technology [1-2]. With significantly higher signal to noise ratio (SNR) and higher resolution, it obtains increased abilities in detecting human microscopic structures comparable with conventional histology, which were previously unseen in lower field. It also enables the observation and analysis of metabolic pathways and functions in living tissues with dramatically abundant details. However, UHF technology also prompts problems related with high frequency RF signals. At ultra high field, due to the sample’s “dielectric resonance effect”, the transmit RF signal generates inhomogenous B1 distribution [3]. High resonance frequency also renders high energy deposition (high SAR values) in human subjects. One feasible solution for these issues is to implement multi-channel transmit/receive (Tx/Rx) RF coils in ultra high field, incorporated with parallel imaging techniques [1-4]. In this project, we have developed a novel 8-channel phase-array Tx/Rx head coil on a Siemens 3T Tim Trio system for parallel transmission, and examined the property of this Tx/Rx coil to expect higher SNR for visual stimulation fMRI experiments. METHODS A novel 8-channel phase-array Tx/Rx head coil with related RF interface was constructed as shown in Fig.1 (a) and (b). The coil was made of polycarbonate and took a cylindrical mechanical structure with a slight bump in the nasal region. The coil inner diameter was 21cm with 26.5 cm in length and 5mm in thickness. The diameter of the coil was slightly less than the standard 12-channel head coil in order to achieve better SNR. A rectangular window of size 116mmx74mm was opened in the upper part of the coil to ensure an unblocked visual field for cognitive science studies. Eight relatively even-spaced rectangular surface loops were divided into 2 clusters, each consisting of 4 loops. The adjacent phase-array loops were decoupled via properly overlapped regions and/or shared common capacitors [5]. For the opposite surface loops, the coupling effect was neglected due to the geometric distance and low-impedance preamplifier decoupling [6]. All the elements were tuned to 123.2MHz and matched to 45-57 ohms with a 7300ml water phantom loading. Network analyzer output was as follows: S11 less than -16db for each channel, S21 decoupling less than -12db between adjacent and opposite loops. Parallel transmission pathway was implemented via a 1-to-8 power splitter and 8 T/R switches, as shown in Fig.1(b) and (c). Phase shifting was achieved using Belden RF coaxial cables with distinct length, and signal losses along the cables between the coil and preamplifiers were less than 0.3 dB. In this setup, RF shimming was achieved by applying linear phase shifters to upper and lower clusters of loops and there was a 180o phase shift of adjacent loops in two clusters. Using special mode-scanning excitation patterns via cable phase shifters and/or RF attenuators, we expect to get region-specific B1 distribution, which may yield higher image SNR in regions of interest [7,8]. In the case of homogenous RF shimming, the global and local SAR distribution of this Tx/Rx coil seemed to function similarly as a TEM birdcage coil. We had set the global SAR control to a conservative value in conformity with the existing FDA standard. The local SAR distribution was monitored via the temperature measurment taken by a Fluke T125 thermal imager through long-time, high power-level, parallel transmission experiments on both a water phantom and a piece of 5kg pork meat. The standard GRE, T1-MPRAGE, T2-TSE and BOLD-EPI sequences were employed to examine properties of the Tx/Rx coil, in compliance with the SAR limitation.