We have previously shown that the amplitudes of motor evoked potentials (MEPs) elicited by transcranial magnetic stimulation (TMS) were transiently decreased after exercise, indicating fatigue of motor pathways in the central nervous system. The responsible mechanism is apparently decreased efficiency in the generation of the descending volleys in the motor cortex. We also noted a progressive decrement in amplitude from the first to the fourth MEP. To further clarify the mechanism of this phenomenon, 5 subjects were studied with TMS delivered at the rates of 0.1, 0.15, 0.3, 1, 3, and 6 Hz. The effect was best demonstrated at 0.3 Hz, and occurred after both isometric and isotonic exercise. Three of the subjects also had 0.3-Hz percutaneous electrical stimulation of the brainstem, and a decrement in MEP amplitude did not occur. Further, the delivery of TMS during muscle contraction after muscle fatigue failed to produce a decrement. The results are similar to those found at the neuromuscular junction in myasthenia gravis and are consistent with a reduced safety factor of cortical synaptic transmission in central nervous system fatigue.