T-1: Modern MRI – beyond Fourier Encoding

Date: Wednesday, May 14

Presented by

Prof. Klaas Pruessmann, Institute for Biomedical Engineering, ETH and University of Zurich, Zurich, Switzerland

Abstract

Magnetic resonance imaging (MRI) is a widely used and highly versatile means of studying live organisms in a non-invasive fashion. For achieving spatial resolution traditional MRI relies exclusively on so-called Fourier encoding with magnetic gradient fields. Correspondingly, traditional MR image reconstruction is mostly done by mere Fourier transform. This situation is increasingly changing. One major strand of development beyond the Fourier paradigm is so-called parallel MRI with arrays of signal detectors. Parallel MRI relies on the distinctness of the spatial reception characteristics of detector elements, which effectively act as a complementary encoding mechanism. As a result, parallel MRI poses a more complex image reconstruction problem and has thus prompted a range of developments in this area.

The goal of this tutorial is to introduce the audience to parallel MRI and the related developments in MR image reconstruction. The first part will cover the principles of both traditional and parallel MRI, along with hardware aspects, basic reconstruction approaches, and a selection of applications. The second part will offer a more formal approach to the general image reconstruction problem and discuss advanced reconstruction tasks. It will also address remaining challenges of generalized MRI reconstruction and some potential solutions.

Part 1: Traditional and parallel MRI

• A brief primer of magnetic resonance and traditional MRI
• Parallel MRI: principle, instrumentation, data acquisition
• Basic reconstruction algorithms for parallel MRI
• Clinical and research applications of parallel MRI

Part 2: The general reconstruction problem, its solutions, and remaining challenges

• A formal framework for general parallel MRI reconstruction
• The special case of lattice sampling
• Iterative solutions: non-lattice sampling and inhomogeneous fields
• Remaining challenges: the fidelity of the forward model

Speaker Biography

Klaas Pruessmann studied Physics and Medicine at the University of Bonn, Germany, and received an M.A. in Physics in 1995. He then undertook doctoral studies at the ETH Zurich, Switzerland, and received a Ph.D. in Physics in the year 2000. In 2002 he was appointed as an Assistant Professor at the ETH’s Department of Information Technology and Electrical Engineering. Since 2005 he serves as an Associate Professor of Bioimaging at the same institution. Between 2003 and 2005 he also held an adjunct professorship at the University of Minnesota, Minneapolis, USA.

The work of Dr. Pruessmann and his group is dedicated to the development of medical imaging techniques on the basis of nuclear magnetic resonance. Its current focus is on MRI at high field strengths, data acquisition strategies, image reconstruction, and MR instrumentation. Dr. Pruessmann is a recipient of the ENC Gunther-Laukien-Prize (2001), the Philips Medical Systems Innovation Award (2001), the European Magnetic Resonance Award (2004), and the ISMRM Gold Medal (2006).