Mental effects of caffeine in fatigued and non-fatigued female and male subjects.

In experiment 1 eight male and eight female subjects were randomly assigned to either a caffeine or a placebo condition. Caffeine (150 mg) was given at midnight and at 4 a.m. Oral temperature, subjective ratings of fatigue and mood, and performance in two cognitive tasks (an auditive attention task and a visual coding task) were assessed. Subjective 'drowsiness' and 'tiredness' increased significantly more in subjects given placebo than in subjects given caffeine treatment. The effects of drug treatment in performance and temperature were non-significant. However, the temperature of female subjects increased between midnight and 4 a.m. and the temperature of male subjects decreased during the same period of time. On the other hand, at 5 a.m. female subjects rated themselves as more sleepy, tired and 'disorganized' than the male subjects. In experiment 2 nine female and nine male subjects were assigned randomly to either placebo or caffeine treatment. Caffeine (200 mg) was given at 5 a.m. Oral temperature, subjective ratings of fatigue and mood, and level of performance in three cognitive tasks (the same as above plus Raven's progressive matrices) were assessed. Moreover, the subjects rated the effort of performing each task. The effects of drug treatment in level of performance were non-significant. However, the subjective effort of performing the auditive attention task increased significantly in subjects given placebo treatment, suggesting a compensatory arousal mechanism (Broadbent 1971). The effect of gender on temperature was non-significant. There was a significant interaction between gender and treatment in respect of subjective effort of performing the matrices task. In men caffeine decreased subjective effort and in women subjective effort was increased by caffeine. Experiment 3 was set up to investigate the hypothesis that negative effects of caffeine in women, observed in experiment 2, were due to over-optimal ('vigilance-related') arousal for the visual coding and matrices tasks. Ten female and eight male non-sleep deprived subjects were given 200 mg caffeine or placebos at 3 p.m. and tested at 4 p.m. Experiment 3 was not found to support the over-optimal 'vigilance-related-arousal' hypothesis. Effects of caffeine in performance and effort were non-significant in experiment 3. Combining data from experiments 2 and 3 gave a significant three-way interaction between caffeine, time for experiment and rule complexity in the visual coding task. When there was a complex rule, caffeine was found to have a positive effect in experiment 3 and a negative effect in experiment 2.