I'm not sure whether sliced or whole brain comes first in the list, but essentially I agree that both are somewhere between whole-body and severed-head. This is curious since a brain is less humanlike than a severed-head, whereas whole-body is actually more humanlike.
The objective question is, of all the problems to be overcome with a given method, which are the least
likely to be solved by technology? These are the ones we must be the most anxious to avoid.
Cloning new organs and limbs is a problem that is about 100% certain to be solved via technology at some point. Stem cell research is already making huge progress in this area. It is a form of damage we can afford to incur. On the other hand, reversing direct freezing damage (where no cryoprotectant is involved) may never be solved by technology. Thus incomplete perfusion is a problem we want to be as sure of avoiding as possible.
Matching and reconnecting neurons based on their position in the tissue and various structural features is something I think quite solvable by technology -- I can certainly think of no reason it would not be. It is a matter of matching patterns and basically reconnecting some disconnected wires. The more the dendrites and cells remain in the same position as when they were seperated, the better. However, a bit of macroscopic warping would not make it impossible to match the proper neuroconnections, since the surrounding terrain would be tend to be very distinctive. If OCR is possible, I can't see why this would not be.
The biggest concern I can think of would be information loss caused by missing material. 250 nanometers is very narrow, but narrower or zero-width would be preferable. Neuron soma are typically between 10000-25000 nanometers across, ranging 4000 to 100000 (ref. wikipedia
), so they could probably be stitched back together wherever they are intersected. Dendrites, on the other hand, are a lot narrower, and some of them are going to be destroyed in very intricate areas. But a good bump on the head probably destroys lots of dendrites in very intricate ways too. The question is whether it would be enough information loss to notice.