It has been suggested that the Mini-Mental State examination can be used to examine a patient's cognitive profile. We therefore examined the validity of Mini-Mental State subtests and individual items. The memory item, attention-concentration items, and constructional item had satisfactory sensitivity-specificity and correlated significantly with scores on neuropsychological tests. In contrast, four of the five Mini-Mental State language items had very low sensitivity, and three of five failed to correlate with neuropsy-chological test scores. These findings establish limits with regard to the ability of the Mini-Mental State to generate a cognitive profile. Our data also provide information regarding validity, difficulty level, and optimal cutoff scores for widely used mental status tasks.