There is some controversy in the literature concerning whether chronic alcohol consumption damages the cerebral cortex. While decreased neuronal density in specific cortical regions is well described in chronic alcoholics, a recent study by Badsberg Jensen and Pakkenberg using unbiased stereological methods questions whether neurodegeneration occurs. In order to assess selective neurodegeneration in the cerebral cortex of chronic alcoholics, regional volumes and unbiased estimates of regional neuronal number (including neuronal identification with calcium-binding proteins) were calculated for 14 chronic alcoholics and 21 controls. Cases were carefully screened to exclude any interfering pathologies. Lifetime and maximum daily alcohol consumption was determined, and homogeneous groups were identified (four chronic alcoholics with Wernicke's encephalopathy and Korsakoff's psychosis, four chronic alcoholics with Wernicke's encephalopathy alone, six chronic alcoholics without Wernicke's encephalopathy or Korsakoff's psychosis, and 21 controls). Brain volume analysis revealed that discrete regions were significantly smaller in the chronic alcoholics compared to controls. As previously shown, white matter regions (particularly in the frontal lobe) were the most significantly reduced in volume. Alcoholics with Wernicke's encephalopathy (either alone or in combination with Korsakoff's psychosis) had significantly smaller white matter volumes than controls or alcoholics without these complications. Medial temporal lobe regions and the thalamus were also reduced in volume. Regression analyses revealed that the volume of both the white matter and thalamus negatively correlated with alcohol consumption. Consistent with the interpretation of previous neuronal density studies, selective neuronal loss was found in the superior frontal association cortex of chronic alcoholics, while no loss occurred from the motor cortex. The number of parvalbumin-, calbindin- and calretinin-immunoreactive neurons was found to be unaltered in chronic alcoholics, suggesting that the neurodegeneration is confined to the non-GABAergic pyramidal neurons. As neurodegeneration was observed in all alcoholic groups, damage to the frontal association cortex is not restricted to alcoholics with the amnesia of Korsakoff's psychosis. These results are consistent with the notion that chronic alcohol consumption is associated with selective neuronal vulnerability. The selective frontal neurodegeneration and the frontal focus of white matter atrophy are supported by neuropsychological, regional blood flow, and magnetic resonance imaging studies of frontal lobe dysfunction in chronic alcoholics and may correlate with abnormalities in working memory.