Although one of the first biological treatments of a major psychiatric disorder was the dietary treatment of pellagra, the use of diet and dietary components in the study of psychopathology has not aroused much interest. This article reviews three areas in which the dietary approach has provided interesting information. The tryptophan depletion strategy uses a mixture of amino acids devoid of tryptophan to lower brain tryptophan in order to study the symptoms that can be elicited. One effect of tryptophan depletion is a lowering of mood, the magnitude of which seems to depend on the baseline state of the subject. Therefore, recovered depressed patients often undergo an acute relapse, while normal subjects show more moderate changes of mood. Totally euthymic subjects show no lowering of mood, but subjects with high normal depression scale scores or subjects with a family history of depression show a moderate lowering of mood. These data indicate that low serotonin levels alone cannot cause depression. However, serotonin does have a direct effect on mood, and low levels of serotonin contribute to the etiology of depression in some depressed patients. Folic acid deficiency causes a lowering of brain serotonin in rats, and of cerebrospinal fluid 5-hydroxyindoleacetic acid in humans. There is a high incidence of folate deficiency in depression, and there are indications in the literature that some depressed patients who are folate deficient respond to folate administration. Folate deficiency is known to lower levels of S-adenosylmethionine, and S-adenosylmethionine is an antidepressant that raises brain serotonin levels. These data suggest that low levels of serotonin in some depressed patients may be a secondary consequence of low levels of S-adenosylmethionine. They also suggest that the dietary intake and psychopharmacological action of methionine, the precursor of S-adenosylmethionine, should be studied in patients with depression. Normal meals have definite effects on mood and performance in humans. The composition of the meal, in terms of protein and carbohydrate content, can influence these behaviors. Because protein and carbohydrate meals can influence brain serotonin in rats, these effects in humans have usually been interpreted in terms of altered serotonin functioning. However, the current balance of evidence is against the involvement of serotonin in the acute effects of protein and carbohydrate meals in humans. The underlying mechanisms involved are unknown, but there are a variety of possibilities.(ABSTRACT TRUNCATED AT 400 WORDS)