Someone with a better grasp of the scientific method perhaps you can answer this. I am under the impression we (science) hasn't definitively determined causation for depression. How can you definitively determine a link without knowing the underlying causes. It would imply you know how it works, doesn't it?
As a (recovered, former) neurobio phd it seems the "grasp" is mostly there + the clarity you want just doesn't exist here
I.e., the concept of "depression" is pretty hard to nail down quantitatively in the first place: The best model we've got is a kind of self-contained proven-by-internal-self-consistency thing bootstrapped from questionnaires, which is of course further abstracted away onto those model systems we're allowed to kill: For example, calling "how long do mice swim before giving up if you chuck them into a tank of water" is roughly equated to "amount of depression" (this is oversimplified a bit; you correlate a bunch of those measures etc. but still a lot of very weak signals). Even "relevant quantitative data" like fMRI / ECoG is bootstrapped from questionnaires when it's something undefinable like "depression"
Which isn't the end of the world, everything started out like this, it just sucks when people take these results too seriously or infer too much from them.
And even the human black-box stuff in the DSM manuals admits a lot of different subtypes, lots of different and sometimes opposite symptom clusters, etc. It is entirely possible that mechanistically "depression" is a diverse group of mechanisms all boxed up into the most familiar mainstream concept/word.