I am very interested in how the actual PSF of a given MRSI acquisition is calculated. Could you, please, recommend me good literature sources, where is this described?
In my reading so far, I understood that the PSF depends on the sampling trajectory (phase-encoded Cartesian/ spatial-spectral-encoded rings/ spirals/ …), truncation of k-space, spatial resolution, but that it also depends on the specifications of the excitation pulse used.
I would very much appreciate relevant literature recommendations, where all things about the PSF are explained from the very basics to advanced considerations. Perfect would be a list of all factor, which influence the PSF.
I’m not aware of a good tutorial or paper that is out there. I’ve asked some people who might know, and will pass them on if I hear back.
Many of the factors you mention (sampling trajectory, truncation of k-space, spatial resolution) are all intimately linked: the PSF is defined by the sampling of k-space in any acquisition. One simple way to calculate it is to pass a uniform signal (i.e. an array of ones) through the reconstruction chain of a sequence. You may need to add some padding to get a decent representation of the continuous PSF.
Regarding excitation pulses, this wouldn’t normally be expected to affect the PSF in MR(S)I. If there is a unresolved dimension in the read-out e.g. the through-plane direction in a 2D MRSI slab, then one could consider the slice profile of any slab-selective pulse applied in that unresolved dimension to be akin to a PSF.
Hope that helps.