Quiz: Physics & Engineering: MR Fingerprinting

Magnetic Resonance Fingerprinting is based on the principle that all images generated during a conventional clinical exam are related in what way?

The images are all of the same person
All images are just different mathematical combinations of a few underlying physical parameters (e.g., T1, T2)
Many MR images exhibit structural similarities with photographs of human fingerprint

Jump to the relevant part of the video for the above question:

What data acquisition requirement does MR Fingerprinting share with Compressed Sensing?

Data must be collected with a non-Cartesian sequence
K-space must be sampled symmetrically about the origin
Sampling must be incoherent, which for MRF is with respect to the signal evolution domain

Jump to the relevant part of the video for the above question:

How is the dictionary which underlies MRF image reconstruction generated?

Using a sequence-specific MR physics model (e.g., Bloch Simulation), evaluated under a large set of different tissue parameter combinations
Randomly
By linearly combining several common compressed sensing sparsity transforms

Jump to the relevant part of the video for the above question:

How is the dictionary used to generate images from highly incoherent data?

Pixel values are assigned as specific combinations of randomly chosen dictionary elements
Patterns in the acquired data are matched to specific signal evolutions in the dictionary, which are in turn associated with specific tissue parameter values
The dictionary is only used during the design of the data acquisition sequence

Jump to the relevant part of the video for the above question:

MRF techniques are capable of reconstructing images representing:

Both physiological and non-physiological (e.g., B0 inhomogeneities) signal sources
Fundamental tissue properties only
At most 4 different signal components

Jump to the relevant part of the video for the above question: