They definitely have a physical effect, but does that effect 'treat' depression any better than, say, atropine?

(http://www.medicalobserver.com.au/news/antidepressants-vs-ac...)

The current understanding is that the difference between SSRIs and active placebos is very small. When you take into account how many negative results have gone unpublished, it starts to look plausible that SSRIs have no psychiatric benefit beyond patients knowing 'I took a pill' and 'it has side effects'.