HiWi position / Master-, or Bachelor thesis project:
Optimization of slab-selective pulses for MRI with deep learning

Stellenangebot vom 24. April 2023

The Max Planck Institute for Biological Cybernetics Tuebingen has one of the most powerful magnetic resonance tomographs in the world, with a field strength of 9.4 Tesla. We are developing the technologies to use this device in basic research for non-invasive imaging of the central nervous system with very high spatial and temporal resolution. A major difficulty here is homogeneous signal excitation, as RF pulse interference can occur due to the higher sense frequency. To minimize resulting image artifacts, the 9.4 T MRI is equipped with 16 independent transmit channels (parallel transmit, pTX).

For the optimization of spatially selective pTX excitation pulses for high-resolution imaging of vessels in the human brain we are looking for a motivated HiWi, Bachelor or Master student.

Project overview:
Problem: Transmit field inhomogeneities (Figure 1) can lead to different inflow signal weightings in time of flight images + spatially varying signal to noise ratio (Figure 2).

Solution: pTX 'spoke' pulses for homogeneous slab selective excitation
Instead of one RF-pulse, use multiple sub-pulses with gradient blips in between (Figure 3)
Instead of a single scaling factor (conventional MRI), each RF pulse has a set of complex scaling weights (16 transmit channels, "parallel Transmit")

Project organization:
Step 1:

  • optimize spokes pulses
  • Existing simulation framework
  • Based on slice orientation and position, calculate RF-magnitudes, RF-phases and gradient moments

Step 2:

  • train neural network
  • Input: slice position (fixed orientation and thickness)
  • Self-supervised learning?

Step 3:

  • Integrate into MR-sequence for online spokes design

Step 4:

  • Implement for arbitrary slice orientation, thickness and positioning

Required Skills:

  • Python programming
  • basic understanding of machine learning techniques
  • interest in medical imaging technology and physics
  • reliability

Do not hesitate to contact us if you are interested:
Max-Planck-Ring 11
Dr. Jonas Bause

Zur Redakteursansicht