This is fantastic. I have been working with a microfluidic guy to sort neurons based on their response to stimuli. It is always a given that you lose many of the cells due to the described issues (even when they are in droplets. This is huge for them! Thanks for the share!
Well, we do harm them because we have to cut off their axons and dendrites to be able to use microfluidics. But this is a problem accounted for. For example, by blocking their protein synthesis so they can't really react to losing their axons. And you can separate them from the cells around them by using enzymes such as trypsin (found on pancreas juice) or collagenase (which digests collagens). Basically loosening up the cells around the neurons.
I'm fine with magnetohydrodynamics (prefer collisionless, but oh well), and TIL "fouling". Don't usually think about medical applications.
Oh please don't make me think about the geodynamo. It's a mess down there. :)