MRI simulation for sequence development, protocol optimisation and education
June 28-30, 2017
Eindhoven University of Technology
German Center for Neurodegenerative Diseases (DZNE)
Medical Image Analysis Department
Faculty of Biomedical Engineering
Eindhoven University of Technology
M. Breeuwer, P. Ehses, L. G. Hanson, D. Pflugfelder, T. Stöcker,
Please click here to view the programme.
Goals of the course:
The course MRI simulation for sequence development, protocol optimisation and education provides insight into practical implementation of MRI computer simulations. It covers the theory of classical MR physics based on the Bloch equations and the necessary steps to simulate and visualize basic MR phenomena as well as basic and advanced imaging sequences. Model-based simulations are helpful for the basic understanding and education of MR physics, and are necessary for MRI method development and sequence design. The lectures introduce the most important aspects, physical models and computational techniques in order to simulate realistic MR experiments. Special emphasis is given to pictorial design and hands-on-the-keyboard lectures, enabling the attendees to use MR simulations in future research projects.
This course will focus on
•Theory of classical MR: Applicability and limitations of the physical models
•Computer simulation of classical NMR and MRI experiments
•MR simulations as an educational tool and for visualisation
•Design and simulation of basic and advanced sequences
•Simulation pitfalls: avoiding beginners mistakes and the generation of “simulation artifacts”
•Simulation of MRI contrasts and artifacts related to basic physics, off-resonance effects, motion, diffusion and flow
•MR simulations in research: Protocol optimisation, pulse design, model verification
•Interfaces to feed simulated MR signals into existing image reconstruction libraries
•Methods to generate ground-truth in silico MR phantoms for image analysis and post-processing
•Differences between simulations and real life
This course in intended for MR physicists and scientists with basic knowledge of classical MRI spin physics and sequence design who are interested in improving their knowledge on computer-based modeling of MRI. Attendees should have a working knowledge in Matlab, Python and/or C/C++ in order to work through the hands-on tutorials.
The course covers theory, application and practical implementation of MRI computer simulations, i.e. the simulation of spin dynamics and their coherent summation in large-scale spin systems under influence of the driving magnetic fields (RF pulses and field gradients). On the one hand, MR simulations based on the Bloch equations are of high educational value. Further, they serve as essential tools in basic MRI method development, sequence design and protocol optimization, and generating ground truth data for image reconstruction and post-processing algorithms.
In this three-day course, the students learn the basics of NMR and MRI simulations. The underlying physical models, their field of application and limitations are discussed. Several tools will be introduced and used in accompanying hands-on sessions. The tutorials will concentrate on exercises of educational value and practical relevance in order to improve the understanding of principal effects, such as echo generation and image artifacts. The students learn to simulate basic acquisition schemes, such as gradient echo, spin echo and EPI, as well as non-Cartesian sampling. Magnetization preparation will be simulated and used to tailor measurement protocols for specific applications. By means of pictorial examples, MRI simulations are shown to serve as a valuable tool for MRI methods development and research. It is shown how image analysis and reconstruction algorithms benefit from MR simulations by knowledge of the ground truth. Finally, the differences between MR simulations and real experiments are outlined by a closer look at the hardware components of a MR scanner.
The course is aimed at post-graduate and post-doctoral MR scientists interested in learning the simulation of general NMR and MRI spin-physics and signal formation based on the Bloch Equations (excluding effects of intra-molecular interactions). A basic background in classical MR spin physics as well as computer programming is required. A working knowledge in MATLAB, Python and/or C/C++ will be helpful. Exercises will primarily be based on the JEMRIS open source MRI simulator distributed for a range of operating systems. Access to a computer cluster will be available. Students are, however, encouraged to bring their own laptop with a pre-installed MATLAB, if possible.
At the end of the course the student will be able to
•Use several basic tools for classical NMR and MRI simulation
•Visualise basic MR phenomena such as excitation, resonance and relaxation
•Explain the advantages, limitations and differences of the physical models on which the MR simulation is based
•Use MR simulation as an educational tool
•Avoid beginners mistakes in MR simulations such as numerical artifacts
•Simulate fundamental MR sequences for image acquisition and contrast manipulation
•Explain basic MRI artifacts and the procedures to avoid them
•Simulate the influence of eddy currents, nonlinear gradient fields, concomitant fields, chemical shift, susceptibility and any other off-resonance sources
•Simulate the influence of motion, flow and diffusion
•Use optimised code for MR physics simulation for integration in own projects
•Perform rapid prototyping and testing of new acquisition schemes
•Perform multi-channel excitation simulations to test new RF pulses
•Use MR simulations in research, e.g. to validate new diffusion or perfusion models
•Feed simulated MR signals into existing image reconstruction libraries
•Generate ground-truth in silico MR phantoms for image analysis and post-processing
•Understand the differences between MR simulations and real life MR experiments