Great question. I think this is why organ systems with (arguably) relatively fewer interconnecting parts and more accessible parts than the brain, for example blood, are easier to research. Not all of the parts will be able to be controlled, and the closest approximations are probably animal models with a few cloned or inducible genetic changes where significant results are not easily generalizable to people. Biology will probably not be like physics or chemistry, where underlying principles are often generalizable to matter under specified conditions, because we can't defineand control the initial context within and around a human, and ethical guidelines ensure that studies are limited to a morally acceptable degree. So we often have to see what works/fails in practice and explain it in terms of what we know, try to fill in the gaps with more research, and see where it takes us in terms of new treatment possibilities. It's a very iterative ordeal.
