AIM: Changes in sensory information from active muscles accompany fatiguing exercise and the force-generating capacity deteriorates. The central motor commands therefore must adjust depending on the task performed. Muscle potentials evoked by transcranial magnetic stimulation (TMS) change during the course of fatiguing muscle activity, which demonstrates activity changes in cortical or spinal networks during fatiguing exercise. Here, we investigate cortical mechanisms that are actively involved in driving the contracting muscles. METHODS: During a sustained submaximal contraction (30% of maximal voluntary contraction) of the elbow flexor muscles we applied TMS over the motor cortex. At an intensity below motor threshold, TMS reduced the ongoing muscle activity in biceps brachii. This reduction appears as a suppression at short latency of the stimulus-triggered average of rectified electromyographic (EMG) activity. The magnitude of the suppression was evaluated relative to the mean EMG activity during the 50 ms prior to the cortical stimulus. RESULTS: During the first 2 min of the fatiguing muscle contraction the suppression was 10 +/- 0.9% of the ongoing EMG activity. At 2 min prior to task failure the suppression had reached 16 +/- 2.1%. In control experiments without fatigue we did not find a similar increase in suppression with increasing levels of ongoing EMG activity. CONCLUSION: Using a form of TMS which reduces cortical output to motor neurones (and disfacilitates them), this study suggests that neuromuscular fatigue increases this disfacilitatory effect. This finding is consistent with an increase in the excitability of inhibitory circuits controlling corticospinal output.