BACKGROUND Until recently, little account had been taken of the confounding effects of caffeine withdrawal and withdrawal reversal when examining the net effects of dietary caffeine. OBJECTIVE By including a manipulation involving sleep restriction, the present study aimed to extend recent findings from research in which caffeine withdrawal and withdrawal reversal were controlled. The main aims of the study were to examine the net effects of caffeine, as well as its potential restorative effects following sleep restriction, on EEG, performance and mood. METHODS A randomised cross-over design was used in which 15 participants alternated weekly between ingesting placebo and caffeine (1.75 mg/kg) three times daily for four consecutive weeks following either usual sleep or sleep restriction. EEG activity was measured at 32 sites during eyes closed, eyes open and performance of a vigilance task. RESULTS Modest effects of caffeine were found in the delta and beta bandwidths, but no main effects of caffeine were observed in the theta or alpha bandwidths. Overall, the effects of caffeine on EEG activity were relatively few, weak and inconsistent, and no evidence was found of net restorative effects of caffeine for any outcome variables. CONCLUSIONS The findings do not support the use of caffeine as a means for enhancing human function or as an antidote to the negative effects of sleep loss.