The effect of central fatigue on the non-exercising arm following fatigue of the contralateral arm has previously been investigated using transcranial magnetic stimulation (TMS) (Humphry et al. 2004). The aim of this study was to examine whether central fatigue and recovery from it is different in aerobic and anaerobic trained subjects compared with sedentary controls. Twenty-four healthy human volunteers in 3 equal groups of 8 age-matched subjects (control, aerobic and anaerobic athletes) took part in two separate sessions. Both sets of trained athletes routinely performed over 8 h exercise per week on average, specific to either aerobic or anaerobic sports. Both groups also averaged over 8 years of training. Control subjects averaged less than 1 h of physical activity per week. In session 1, a 4.5 kg weight was attached to the non-dominant forearm and subjects performed bicep curls until exhausted. In session 2, the dominant arm was exhausted first, followed by the non-dominant arm. The central limit of endurance (CLOE) was calculated as the percentage difference between the time taken by the non-dominant arm to fatigue in session 1 and 2. Digital dexterity and maximum voluntary contraction (MVC) were also measured before and after exhaustion. TMS was applied using a MagStim 200 connected to a 9 cm circular coil centred over the vertex and the change of motor-evoked potentials (MEPs) in biceps brachii and thenar muscles was monitored bilaterally during both sessions and for a recovery period of 30 min. All three groups showed significant reductions in MEPs for the non-exercising biceps after exhaustive exercise of the opposite biceps (P<0.05, repeated measure ANOVA). MEP size recovered 20 min post-exhaustion in aerobic athletes, 30 min in controls but was not recovered in anaerobically trained athletes at 30 min (P<0.05, post hoc Holm-Sidak). The CLOE (±SEM) was significantly different between groups (aerobic athletes 106.8±7.1%; anaerobic athletes 69.4±6.4%; controls 83.9±5.7%, repeated measures ANOVA on ranks, P=0.002). There was no central fatigue evident for thenar MEPs and no effect on digital dexterity or MVC after exhaustive exercise for any group. Thus, in comparison to control and anaerobic subjects, aerobically trained athletes demonstrate reduced central fatigue and faster recovery from it, whereas anaerobic athletes more closely resemble control individuals. We propose that long-term aerobic training results in attenuation of the normal central fatigue response to exercise.