Jens Hoffmann

Radiofrequency Transmission Techniques for Human Brain Imaging at 9.4 Tesla

The Max Planck Institute for Biological Cybernetics is equipped with one of the strongest magnets (9.4 Tesla) for human Magnetic Resonance Imaging worldwide. These high-field magnets facilitate structural and functional neuroimaging at unprecedented spatial and temporal resolutions.

However, the high field strength comes at the cost of a short wavelength of the radiofrequency (RF) field that is generated by the RF coils. As a result, interference effects cause severe inhomogeneities in the RF field across the human brain. This has two major implications: First, inhomogeneities in the magnetic field component (B₁⁺) lead to an impaired image quality if conventional RF coils and pulses are used. Second, the non-uniform electric field across the brain can cause localized energy dissipation and tissue heating, respectively, and hence poses a potential risk to the subject.

This project aims at the development of RF transmission techniques and hardware in order to overcome those problems. In particular, I work on:

• Development of multichannel traveling-wave antennas that serve as simple, remotely-placed RF coils capable of imaging the entire head at 9.4 Tesla.

• RF shimming: The RF fields emitted by an array of independent coils are superimposed in a way that a more uniform B₁⁺ field in user-defined regions in the brain is achieved.

• Dynamic parallel transmission (pTx): The single RF fields emitted by a coil array are modulated dynamically in concert with the spatial encoding gradients to actively control the tissue magnetization during an RF pulse and to achieve a more homogeneous excitation.

• Safety evaluation of self-built RF coils, particularly related to potential tissue heating caused by the RF field during conventional transmission, RF shimming and pTx.

All subprojects involve the precise numerical simulation of the RF coils and their electromagnetic field distribution for the purpose of coil design, prototyping and safety evaluation.

Simulation and safety evaluation of self-developed RF coils

Any self-built RF coil must meet requirements for safe application on humans with respect to mechanical and electrical safety and it must be ensured that RF transmission is in compliance with specific absorption rate (SAR) limits in order not to raise tissue temperature above critical levels.

In this project, we developed protocols for safety testing and operational procedures including mechanical and electrical safety, SAR, risk analysis, coil documentation, quality assurance and user training. The local SAR distribution is assessed using numerical simulations that include a precise representation of the coil setup, the scanner environment and the input- and decoupling circuits (Fig. 1A). Simulations are thoroughly validated via comparison to experiment (Fig. 1B) before SAR is assessed in human voxel models. Moreover, these simulations support the design of novel coils and provide insight into RF field propagation phenomena at ultra-high fields.

RF Shimming

RF shimming  or static parallel transmission uses multiple independent RF coils to control the RF field pattern across the human body. In order to improve the B₁⁺ pattern in a pre-defined body region, for example to achieve a more homogeneous B₁⁺ field, a fixed set of RF voltages with optimized amplitudes and phases is calculated and applied to the single coil elements. An optimized excitation mode can readily be used with any standard MR sequence in order to improve image quality.

In this subproject, we developed techniques for rapid, subject-specific single-channel B₁⁺ mapping and for the numerical calculation of optimized RF shim configurations. The performance and limitations of RF shimming at 9.4 T using a state-of-the-art dual-row transmit array were investigated with respect to B₁⁺ homogeneity, efficiency and SAR.

Multimode traveling-wave imaging

At a frequency of 400 MHz, the bore of a human 9.4 Tesla scanner acts as a waveguide that conveys the RF field from a remotely placed antenna towards the subject. The propagation of the RF field can be interpreted as a superposition of three basic waveguide modes.

In this project, we designed and evaluated a simple three-port antenna (Fig. 3A) that can be positioned distant to the subject and that is capable of controlling all three waveguide modes simultaneously (Fig. 3B). Two time-interleaved acquisitions using complementary RF shims (TIAMO) and their subsequent combination facilitate 9.4 T images of the entire brain without signal dropouts (Fig. 3C). This makes the antenna suitable as a remote ¹H volume coil for anatomical localization and B₀ shimming. The compatibility of similar designs with dedicated receive arrays for ¹H, ³¹P or ²³Na applications was demonstrated.

Dynamic parallel transmission

The dynamic parallel transmission (pTx) method persistently and independently varies the RF amplitudes and phases on multiple coil elements using dedicated hardware and hence dynamically changes the B₁⁺ pattern during the RF pulse. This is performed in concert with the spatial encoding gradients in order to directly control the buildup of the magnetization, for example to achieve a homogeneous flip angle (Fig. 4A) or a desired flip angle distribution (Fig. 4B) in a selected region in the human brain.

This subproject aims at the improvement of different aspects of pTx methodology to take the method towards routine application at 9.4 T. It involves the development of optimized coil arrays and associated hardware, comprehensive numerical simulations to ensure compliance with SAR limits as well as rapid volumetric B₁⁺ and B₀ calibration scans that constitute the basis for subsequent, subject-specific pTx pulse design.