Depression is a major contributor to the global burden of disease and disability, yet it is poorly understood. Here we review data supporting a novel theoretical model for the biology of depression. In this model, a stressful life event leads to microdamage in the brain. This damage triggers an injury repair response consisting of a neuroinflammatory phase to clear cellular debris, and a spontaneous tissue regeneration phase involving neurotrophins and neurogenesis. During healing, released inflammatory mediators trigger sickness behavior and psychological pain via mechanisms similar to those that produce physical pain during wound healing. The depression remits if the neuronal injury repair process resolves successfully. Importantly, however, the acute psychological pain and neuroinflammation often transition to chronicity and develop into pathological depressive states. This hypothesis for depression explains substantially more data than alternative models, including why emerging data show that analgesic, anti-inflammatory, pro-neurogenic and pro-neurotrophic treatments have antidepressant effects. Thus, an acute depressive episode can be conceptualized as a normally self-limiting but highly error-prone process of recuperation from stress-triggered neuronal microdamage.