6 Image contrast and noise
Timing is everything: decisions about excitation, k-space coverage, and acquisition determine image contrast and signal-to-noise ratio. We’ll start by talking about image contrast.
For a full derivation of the Ernst angle, check out the last few pages of these notes on the Bloch equations: MRImagingBasics_BlochEquations
Exercises
1. What is the Ernst angle for a tissue with a T1 of 1100 ms and a TR of 2 seconds?
2. What is the Ernst angle for a tissue with a T1 of 1600 ms and a TR of 2 seconds?
The reason we care about optimizing signal is because we never have enough to work with — we’re always trying to increase our signal because fMRI contrast is relative (it’s measured as percent signal change). So the more signal we have, the more BOLD response we have … and we need it because there are many noise sources, and the noise can be as large as the task response that we’re looking for.
Exercises
3. Imagine 2 datasets, both acquired with the same matrix size, but one had a smaller field of view than the other so the resolution is higher. Which one has better SNR?
4. If I tell you that one of those images is “thermal noise limited”, which one do you think it will be?
5. If you acquire 2 datasets that have the same resolution, and both have a big enough FOV to cover your entire sample … but one has a larger FOV in the read-out direction, which dataset will have higher SNR?
