d t b j l p d p a a T wo manuscripts appear in this issue of Biological Psychiatry involving suicide. Ernst et al. (1) found reduced expression of connexin 30 and connexin 43, which regulate calcium transients in the lateral prefrontal cortex of suicide completers. It is thought that astrocytes express connexins and it is argued that altered connexins affect astrocyte function, which could contribute to suicide. Cyprien et al. (2) found that the size of the caudal third of the corpus callosum was reduced in a group of suicide attempters. Taken together, the findings support the notion that there may be a neurobiology of suicide distinct from the psychiatric disorders commonly associated with suicide, involving alterations in glia, white matter, and by extension, the regions where the astrocytes are derived from (1) and where the regions contribute fiber tracts to the corpus callosum (2). Given the markedly different substrates (astrocytes vs. white matter; prefrontal cortex vs. corpus callosum), there is the further likelihood that suicide neurobiology has multiple phenotypes involving not only neurodegeneration but also neurodevelopment and neuroplasticity. The possibility that there is a neurobiology of suicide has been raised before (3). Although suicide as a cause of death was originally viewed as an extreme outcome of major depression and then later as a potential confound in studies of the neurobiology of mood disorder, suicide is increasingly thought to have its own unique biological phenotype (4). Psychological autopsy studies have found that major depressive disorder is present in approximately 60% of suicides, while there is an Axis I diagnosis in more than 90% of all suicides (e.g., see [5]). The conclusion most often reached is that suicide is not an extreme response to stress, adverse events, or disease in normal people, but rather that suicide is an extreme act of someone with a psychiatric illness compounded by one or more adverse life events and a biological predisposition or vulnerability. Biological studies of suicides have found differences in neurotransmitter receptors, neurons, glia, and/or white matter in various anatomical regions (6). In those studies reporting differences, the brain regions most commonly involved were in the prefrontal cerebral cortex. Other studies report differences in subcortical regions, adding further biological complexity. Interestingly, not all studies report differences, and even in those studies where differences are found, the magnitude of the differences is small. There are multiple confounds that can compromise human postmortem studies (postmortem delay, medication, psychiatric diagnosis, method of death, anatomy, methodology, age, sex, race, religious affiliation, socioeconomic status, tissue storage duration, tissue storage method, time of death, season of death, family history, agonal state, and others [7]), making it almost remarkable that any findings are detectable at all. This view is overly pessimistic, as the majority of human postmortem studies, including the report in this issue of Biological Psychiatry, go to lengths to control and minimize potential confounding variables. Yet, even with carefully controlled