This paper addresses the question whether negative cognitive style represents a state or trait variable of depressive patients. For this reason, it studies the influence of sleep deprivation on negative self-schemes of those patients. 10 patients suffering from DSM-III-R major depression were compared with 10 age- and sex-matched controls on a task for rating the self-descriptiveness of positive and negative adjectives as well as a subsequent word recognition task. Three sessions were involved: an initial session (baseline), the second following a night of sleep deprivation, and the third after a successive full night's sleep. During the baseline examination, depressives showed a relatively negative cognitive bias; that is, the same number of positive and negative self-scheme elements. In comparison to controls, they showed significantly more negative and significantly less positive self-scheme elements. The same pattern emerged in a word recognition task for the number of recognized self-scheme elements. These variables indicated no change in the depressive group following sleep deprivation. Depressive subjects' reaction times on self-descriptiveness rating were significantly longer for positive than for negative self-scheme elements at the baseline session. The opposite was true for controls. Here, a sleep deprivation effect was evident. There was no longer a difference in the speed of information processing for positive as compared to negative self-scheme elements. This applied to both depressive and control groups.