Looks like they are using molecular methods on post-mortem brains - all sorts of bugs can grow post-mortem and molecular methods can give false-positives as well as detect non-viable microbes, etc. But that's also based on a quick skim / haven't read the full paper yet.
Culturing from CSF in general will depend on the concentration of the microbe, whether they are viable, if antibiotics/antivirals were already initiated pre-collection, whether it's plated on the appropriate media (e.g. a rare microbe that only grows on one specific type of agar plate), etc. Culturing viruses is also hard/many hospital micro labs have moved away from that.
I think this study may suggest that we are failing to detect certain brain infections (and many are notoriously hard to diagnose if you don't catch them in the right window of time). But a brain microbiome sounds far-fetched. We even plate from brain tissue directly at times and aren't growing a bunch of organisms. I'd approach that claim with a healthy dose of skepticism.
plus we have the fact that the brain at night shifts the fluid to the periphery of the blood vessels to clear out the lactate bound NH4 which would impact any microbiome if it exists...
