> Indeed, researchers recently reported in Frontiers in Psychiatry that fecal transplants improved symptoms in two depressed patients.
Depression research is challenging because even placebo groups tend to have dramatic improvements.
That is a frequentist argument.
If the response to treatment is significant enough, then that is a worthwhile signal (cliché: parachutes save lives without a control group).
Placebo is an obvious contender. However presuming the patients had a chronic condition with plenty of previous interventions, then the placebo effect has had plenty of chances in the past. If it were due to placebo then there would be something to learn about why the placebo effect were so effective in that situation!
But you can imagine looking at a picture and feeling hungry.
In addition to our fast new electrical nervous system for rapid response there is the older slower endocrine system from when we were just bags of chemicals coordinating with concentration gradients (still are, but were).
visual -> brain -> signals gut -> gut makes molecules to get ready -> molecules detected report to brain -> you drool
Moods depend on the chemical balance of small molecules bacteria eat sleep and poop small small molecules.
To consider that after billions of years of coexisting we, both multi cell & single cell, have not adapted to responding each others's small molecules is unthinkable
The sibling point that treatments targeting the biome would be expected to fail, if that were the case, is complication.