The first half of this paper briefly reviews the evidence that (i) stress precipitates depression by damaging the hippocampus, leading to changes in the activity of a distributed neural system involving, inter alia, the amygdala, the ventromedial and dorsolateral prefrontal cortex, the lateral habenula and ascending monoamine pathways, and (ii) antidepressants work by repairing the damaged hippocampus, thus restoring the normal balance of activity within that circuitry. In the second half of the paper we review the evidence that heightened vulnerability to depression, either because of a clinical history of depression or because of the presence of genetic, personality or developmental risk factors, also confers resistance to antidepressant drug treatment. Thus, although antidepressants provide an efficient means of reversing the neurotoxic effects of stress, they are much less effective in conditions where vulnerability to depression is elevated and the role of stress in precipitating depression is correspondingly lower. Consequently, the issue of vulnerability should feature much more prominently in antidepressant research. Most of the current animal models of depression are based on the induction of a depressive-like phenotype by stress, and pay scant attention to vulnerability. As antidepressants are relatively ineffective in vulnerable individuals, this in turn implies a need for the development of different clinical and preclinical methodologies, and a shift of focus away from the current preoccupation with the hippocampus as a target for antidepressant action in vulnerable patients.