Scientists and mental health professionals have made major strides in understanding and treating major depression, including the discovery of the role played by imbalances in brain chemicals such as serotonin and noradrenaline. But there is increasing evidence that there is more to the story: depression also involves structural changes in areas of the brain that are involved in mood, memory, and decision-making.

Depression is often precipitated by stressful experiences. The brain interprets our experiences and decides if they are threatening, and then controls our behavioral and physiological responses to them. Data obtained from animal experiments show that harmful physiological changes result from the inability of the brain and body to respond to repeated stress with adaptive modifications in structure and function. Three brain areas - the hippocampus, the prefrontal cortex, and the amygdala - are particularly susceptible to pathological changes in size and function.

These areas are instrumental in interpreting what is stressful and in determining appropriate responses. Many chemical mediators are involved, including cortisol and adrenalin from the adrenal glands, other hormones and neurotransmitters (such as serotonin and noradrenalin), and responses from the autonomic and immune systems. Depression thus leads to changes in the rest of the body, owing to long-term chemical imbalances in the systems that control the heart, the immune system, and metabolism.

The hippocampus, a key structure in the formation of memories of events and contexts, expresses receptors that enable it to respond to stress hormones in the blood. We now know that it atrophies in a number of psychiatric disorders.

Within the hippocampus - a curved elongated ridge located in each of the brain's two temporal lobes - is a structure called the dentate gyrus, which produces new neurons throughout adult life. Chronic stress inhibits neurogenesis and also causes many hippocampal neurons to shrink - a process called "remodeling." Animal studies have shown that the stress hormone cortisol, together with neurotransmitters in the brain, also plays an important role in the remodeling of neurons.

Similarly, there is evidence that the prefrontal cortex - a key structure in emotional regulation as well as in decision-making and working memory - shrinks in major depression. On the other hand, rather than atrophying, the amygdala, the seat of emotional memories, becomes hyperactive both in post-traumatic stress disorder and depressive illness, and there is evidence from animal studies of a hypertrophy of nerve cells after repeated stress.

A hyperactive amygdala, along with abnormal activity in other regions of the brain, leads to disrupted patterns of sleep and physical activity, as well as irregular patterns of secretion of hormones and other chemical mediators that control many of the body's structures. For example, levels of the stress hormone cortisol increase in the evening, when they are normally low.

Over time, if uncorrected, major depression can cause such conditions as progressive mineral loss from bones, accumulation of abdominal fat, increased blood platelet reactivity, and increased risk of cardiovascular disease. Moreover, certain types of antidepressants may contribute to some of these conditions. Nevertheless, one of the most interesting features of these structural changes is that they can be prevented, and potentially reversed, with medication.

Indeed, in key respects depression is comparable to Cushing's disease, a disorder cause by excessive production of cortisol. The elevated cortisol in Cushing's patients is higher than that found in major depression, but the psychiatric and somatic features of Cushing's disease are strikingly similar: melancholia, depression, abdominal obesity, bone mineral loss, and increased risk for cardiovascular disease, along with shrinkage of the hippocampus and memory impairment.

The good news from Cushing's disease is that hippocampal shrinkage and memory impairment is at least partially reversible over several years after correction of the excess cortisol. This encourages the belief that structural changes in the brain that are associated with major depression can be prevented or even reversed with the right medications and other treatments. For example, neurogenesis in the dentate gyrus that is suppressed by acute and chronic stress is elevated by many antidepressant treatments.

Still, one of the problems with many of the current medications is that, even with treatment, the recurrence of major depression is very high - around 70-80%. This suggests that physiological processes underlying depression - the ones that may also lead to structural changes in the brain - are not being completely arrested, and thus that a new generation of anti-depressant medications is urgently needed.

If so, new, the new antidepressants would have to target what appear to be progressive neurodegenerative changes in the brain, which we infer are taking place because recurrence of depression is so high under currently available treatments. Lithium, a mood stabilizer and neuroprotective agent, has the properties we are seeking, although its toxicity makes it less than ideal. As ever, the challenge facing medical researchers is to develop more effective treatments without reducing their safety.