MRI Engineering Core

National Institute Of Neurological Disorders And Stroke

Investigators

Linked publications, trials & patents

Abstract

NIH COVID policy has severely restricted experimental progress by the MRI Engineering Core. The MRI EC supports new hardware developments for the highest field MRI systems at NIH, and supports specialty projects for other systems. Currently, our main goal is to develop 11.7T human MRI and 17.6T animal MRI in order to perform neuroimaging with superior contrast and resolution. The 11.7T system is awaiting (re-)installation to be scheduled once COVID-19 restrictions are lifted, while a feasibility study will be performed for possible installation of the 17.6T animal MRI. In particular for the 11.7T system, major technical developments are needed to allow transmission and detection of the required RF fields, a difficult and yet unresolved problem. For this reason, the EC has been working on three RF projects in parallel. For the initial human studies at 11.7T, we plan to use a general-purpose whole head coil for both transmission and reception. As we plan to use this coil for the initial safety testing required by the FDA, we opted for a relatively simple design that allows for accurate prediction of transmitted electro-magnetic (EM) fields and associated tissue heating from simulations and measurements. We settled on a transmit-receive type 500 MHz inductive birdcage resonator. EM field simulations were performed using commercial software and used to compare measured and simulated frequency response when transmitting and receiving at different combinations of teo coupling points. Excellent correspondence between simulations and measurements was observed under various conditions, improving our confidence in our ability to predict coil performance. After additional measurements and simulations with a spherical test object (phantom), we will repeat these measurements and simulations on our brain phantom, which more closely matches the human head in terms of shape and dielectric and conductive properties. We constructed this phantom by filling a head-shaped glass container with polyvinylpyrrolidone (https://amri.ninds.nih.gov/cgi-bin/phantomrecipe). Temperature measurements (using fiber optic probes) with this phantom during RF transmission will be compared to simulations of the deposition of RF power and heat conduction as a steps towards setting safe RF transmission limits in human. This assessment will be performed without the 11.7T magnetic field, after which further evaluations will be performed (e.g. comparing simulated and MRI-measures B1 field distributions) when the 11.7T scanner becomes operational. For several years, on-coil RF amplification technology has been developed for multi-channel transmission (also called pTx). At 3T and below, pTx has been successfully commercialized as it reduces cabling issues and amplifier cost. At higher field, pTx is more challenging to implement but has the important advantage of improved control of the RF transmit field. At 11.7T, this is a critical issue, as it affects not only contrast and sensitivity, but also safety for operation in-vivo. Compared to conventional remote voltage-mode RF power amplifiers, on-coil amplification allows better B1 control, reduced load sensitivity, and reduced power losses at a lower implementation cost. In addition, this technology also allows direct sensing of coil current, information that can be used for safety monitoring and feedback. Accurate RF transmit control allows better estimation of increased tissue heating associated with high field MRI. In order to evaluate feasibility and identify potential roadblocks for on-coil pTx at 11.7T, we built an 8-channel 7T version using optically controlled current-source RF power amplifiers. This version incorporated an optimized current monitoring circuitry for higher power efficiency and bandwidth and safer operation. RF current was monitored at the amplifier output, digitally encoded on the amplifier board, and optical transferred to a remote controller (Linux PC). Each amplifier generated at least 100W of peak power with power efficiency around 70%. This power could be used at MRI-relevant duty cycles of 10% without causing excessive heating of the electronics To control eight of these amplifiers in parallel, a broadband eight-channel pTx interface was designed and built in-house. This interface can generate, from a single small RF signal (scanner RF power amplifier input), 8 optical carriers and 8 optical envelope signals whose phase and amplitude can be independently controlled through vector modulation. This interface also optically receives the eight monitored RF currents. Based on this information the controller can update the 16 baseband signals that control the vector modulators and set amplitudes and phases for each RF pulse. All electronics were designed to allow interfacing with a wide range of MRI systems (covering at least the 1.5T-11.7T frequency range). For initial testing, we operated the amplifiers in combined transmit/receive (T/R) mode. This allowed imaging, and initial assessment of ability to control the transmit field distribution and identify potential issues. For this purpose, each amplifier was connected to a 6 cm diameter loop through a small printed circuit board that contained a T/R switch and a low noise amplifier (LNA). Coils and electronics were assembled on a 265 mm inner diameter cylindrical former with a sliding RF shield to reduce interference with the surrounding hardware in the scanner bore. With this setup, good quality images of a spherical oil phantom (24 cm diameter) and our brain phantom were acquired while the RF transmit current in each loop was monitored in real time. It was observed that on coil power amplifier and near coil LNA implementations allowed well decoupled T/R elements by impedance-based methods only. Bench measurements confirmed that inter-channel transmit power leakage was less than -15dB (3%), owing to the high transmit amplifier output impedance. A preliminary human study was performed with the new T/R array under restricted power under IRB approval. Optimal, whole-brain receive sensitivity will require detectors with at least 128 receive elements. While the electrical design issues of such high-element count detector arrays are well understood, the limited space inside head-only scanner makes actual construction a challenge. In part, this is due to the requirement that at 500 MHz, the receive amplifiers need to be close to the receive elements. Further challenges are due to the fact that the inner diameter for the transmit coil we plan to use outside the receive array is only 27cm due to the gradient coil. To expeditiously start with human experiments when the scanner becomes available, we opted to start with a relatively modest element count of 32, to first get experience with the practical issues at 500 MHz before increasing channel count. The 32 channel receive-only head coil was designed to be used in combination with either a detuneable version of the volume resonator, or with a multi-channel transmitter currently under development in the lab. The array has 24 posterior loop elements and 8 anterior loop elements, all with identical surface area. Using 3D modeler software, the coil conductor layout was designed to allow placement on a former that was constructed to fit a large human head. Inter-element decoupling, important for optimal sensitivity, was performed in part by geometric overlap of elements, and in part by proper impedance matching to in-house built low input impedance LNAs previously tested in our 11.7 T animal system. With four of the elements connected to LNAs, coupling (leakage) between neighboring row elements was less than -18.4 dB (1.5% coupled power). The EC continues to help a number of other investigators at NIH who have hardware needs. Personnel continue to support human subjects research and human protocols.