Despite a range of etiological theories since the introduction of electroconvulsive therapy (ECT) more than 75 years ago, its mechanism of action remains poorly understood. The neuroendocrine hypothesis is based on the seizure-related release of hypothalamic hormones into the blood and cerebrospinal fluid and evidence of endocrine dysfunction in many patients with severe mood disorder. The specific effect of ECT was hypothesized to result from the transverse passage of current through the brain with direct stimulation of axial structures including the diencephalon. The prompt release of adrenocorticotropic hormone, cortisol, and prolactin into blood followed ECT with a return to pretreatment baseline levels in several hours. The elevated levels of hormones were absorbed by the cerebrospinal fluid, providing contact with brain cells and central nervous system structures. An apparently specific pattern of ECT-induced hormone changes, limited to prolactin and cortisol, suggested that ECT released a substance with dopaminergic antagonist and antipsychotic properties. As hypothalamic dysfunction is a key finding in endogenomorphic depression and the abnormal endocrine and physiological functions usually normalize with recovery, this led to a search for biological markers that would supplement clinical assessment of diagnosis and treatment response. One of these, the overnight dexamethasone suppression test found that 40% to 50% of melancholic depressed patients had abnormal results, whereas 90% of control patients suppressed normally. This was followed by a period of uncritical overenthusiasm followed by wholesale rejection of the clinical neuroendocrine strategies. Several key methodological issues received inadequate attention, and there have been calls to revisit this topic.