Thank you for concise explanation :). I guess that I must have overestimated the macroscopic change following denaturation.

Np, this all falls under the category of connectomics / next-gen proteomics, so I'm always happy to discuss it.

is there anything useful that we could learn resulting from using MRI on the enlarged sample?

I doubt it, MRI's advantage is as a non-invasive imaging tool. You can look at changes in blood flow, inflammation, structural damage, etc, without having to open up the organism or stick any probes (other than maybe contrast agents in the blood) inside. Taking out an organism's brain just to stick it back in an MRI, instead of using a finer microscope, renders those advantages invalid.

At best, you could create a full 3D image of a tissue without having to cut it into slices (even expanded, the tissue tends to scatter light after a few cm). But iirc, even the strongest MRI magnets have a resolution limit around a millimeter in scale, compared to light microscopes, which can get micrometer resolution without much difficulty.

For connectomics, you want a method that can image down to the synapse (Resolution ~= 100s of nm), but also do so at a wide enough scale so as to capture information on the whole brain / brain region (1 cm - 10 cm). But that's still not a solved problem outside of brute-forcing it with millions of knife slices and decades of microscope time.