Considerable evidence implicates the neurotransmitter gamma-aminobutyric acid (GABA) in the biochemical pathophysiology of mood disorders. Animal models of depression show regional brain GABA deficits and GABA agonists have antidepressant activity in these models. Somatic treatments for depression and mania upregulate the GABAB receptor, similar to the effect of GABA agonists. Clinical data indicate that decreased GABA function accompanies depressed or manic mood states. GABA agonists are effective antidepressant and antimanic agents. Low GABA levels are found in brain, cerebrospinal fluid and plasma of patients with depression and in plasma of patients with mania. Plasma GABA levels, which reflect brain GABA, are not normalized with treatment and clinical remission in depression, suggesting low GABA is not a marker for mood state. Some somatic treatments, including valproic acid and electroconvulsive shock, reduced plasma GABA and response to these correlates with higher levels of baseline plasma GABA. From these data, a GABA hypothesis for mood disorders is formulated. Low GABA function is proposed to be an inherited biological marker of vulnerability for development of mood disorders. Environmental factors, including stress and excessive alcohol use, may increase GABA, causing symptoms of depression or mania. Treatment, or the passage of time, then returns GABA to its presymptomatic baseline as the symptoms remit. This hypothesis, applicable to a subset of mood disordered persons, is testable.