Simulating more advanced saturation schemes

[DrSAR]

While you are on the subject of advancing the simulation capabilities, it could be useful to look at saturation schemes where the CW sat pulse is replace by a series (n=100-1000?) of short (ms?), gaussian (or start with block if easier?), separated by short gaps (ms? -> duty cycle 10-100%).

The CERT (chemical exchange rotation transfer or some such) is based around the idea of comparing two spectra with differences in only their sat pulse parameters thereby replacing the questionable Z_asym analysis with something slightly better founded.

[srveale]

I think I'm where I'm supposed to be with the pulsed sequence.

It's possible to determine the pulse repetition time (PTR) from the flip angle (FA), average power (B1avepow), and duty cycle (DC), all as defined and explained in these papers: 'Optimizing Pulsed-Chemical Exchange SaturationTransfer Imaging Sequences.' 'A New Method for Detecting Exchanging Amide Protons Using Chemical Exchange Rotation Transfer'

So I did just that. Using the example of the first paper, I set B1avepow =1.0 uT, FA = 180, and DC = 0.5, to get a PTR of 21.3 ms (ours is slightly different from the example, so our Gaussian pulse must have a slightly different width). It's then a matter of calibrating B1max. Have a look at the notebook "CEST/simulations/Pulsed Saturation- Pre-defined pulse-test spectrum" to see how this is done.

The history of z-magnetization, over the first few pulses looks like this. 180 degrees, followed by relaxation, followed by 180 degrees etc.

There are imperfect bumps; I haven't calibrated to an exact pi pulse - though this might be more realistic anyways.

Here are two spectra: one obtained using a pi pulse and one with a 2pi pulse. The sequence for the 2pi pulse simply doubles the PTR, keeping average power the same.

I know what you're thinking. Why are they so different? Well I forgot that doubling PTR while keeping the number of pulses the same will double the total saturation time, meaning the 2pi spectrum received much more saturation power. Think of this more as proof-of-concept for simulating pulsed CEST rather than a CERT experiment.

[srveale]

And I'd like to point out that in both the papers I mentioned, spoiling is simulated by setting transverse magnetization to zero between each gaussian pulse. Is anyone sure one way or another about if I should do this?

[DrSAR]

Well done. The interesting part will be looking at the small bump in the baseline for preferably two species (with different exchange rates) and seeing how they should be affected differently in the pi v 2pi spectrum. So this requires at least a three-pool model. Or one could do the simulation once with a fast and once with a slow exchanger on a two-pool model.

The spoiling can be done as described. This can be effected (@firasm , this is the correct spelling) by little gradient lobes right after and/or before the pulses. In the simulations those gradients are just described by fully dephasing in-plane magnetization.