The purpose of this study was to reproduce and extend two earlier studies of the effects of human exposure to 50 Hz magnetic fields (MF). In a recent paper, we described results of two double-blind investigations performed to examine effects of 100 microT(rms) 50 Hz MF exposure on psychological parameters in the same group of healthy human volunteers. In each exposure session, at 1 week intervals, with sham, continuous, and intermittent (15 s ON/OFF cycles) MF conditions, mood ratings, performance measures, and electrophysiological measures were taken. In the first study, significant amplitude changes were observed in the event-related brain potentials (ERP) recorded during a dichotic listening task. In the second study, latency and reaction time (RT) slowing were seen on a visual discrimination task (P(300) paradigm). Although these results were little related to the number of parameters analysed, they indicate that low level 50 Hz MF might have a slight influence on ERP and RT under specific circumstances of sustained attention. Before concluding that moderately strong MF exposure can influence cognitive function, previous results should be replicated, using the same paradigms with another group of healthy volunteers. In the present study, 18 healthy subjects were exposed to three experimental sessions of 30 min each, given at 1 week intervals. The sessions consisted of continuous 100 microT(rms) 50 Hz MF exposure, sham condition, and bright light (5000 lux) exposure. The study was performed double-blind, with the exposure order counter-balanced. The data on mood, ERP, RT, and other performance measures did not show any differences among the sham exposure, light exposure, and MF exposure conditions. The results of this study do not support the hypothesis that extremely low frequency (ELF) MF exposure affects the brain's electrical activity or cognitive function at field strength (100 microT(rms)) similar to that found in very close proximity of some household and industrial electrical appliances and well in excess of the average MF strength (c. 0.1 microT) found in homes. The sensitivity of the experiment was possibly not sufficient to detect an effect at this relatively low MF, and larger sample sizes would be required in further studies.