The after-effects of fatigue or stress on the performance of cognitive tests have been particularly difficult to demonstrate. In this study we examined whether salivary cortisol, used as an index of stress evoked by the continuous performance of mental tasks, reflected individual differences in cognitive performance. In a within-subject experiment in which 24 subjects were exposed to 4 hours of continuous mental activity and to a control session, significantly higher cortisol levels were found during the continuous task session. Cognitive performance was assessed before and after each session. The relevant test parameters involved aspects of verbal memory, concept shifting and divided attention. When subjects were divided into two groups based on the magnitude of individual cortisol responses to the continuous tasks, it was found that the subgroup with higher cortisol responses decreased in attention compared with their attention after the control session. In contrast, the performance of the subgroup with no or lower cortisol responses did not differ between the two sessions. There was no evidence of similar effects on verbal memory or concept shifting.