Sustained or repetitive use of skeletal muscles leads to muscle fatigue. For human exercise, muscle fatigue can be defined as a decrease in maximal voluntary force or power of a muscle or muscle group. This definition implies that not just the muscle itself but also the nervous system has a role in exercise-related fatigue. During fatiguing exercise, the central nervous system can compensate for some weakness and slowness of the muscle to allow continued performance of a task, but may also become “fatigued” and contribute to a fall in maximal voluntary force. Twitch interpolation has been used to demonstrate central fatigue by showing that fatigue increases the increment in force evoked by electrical stimulation of a muscle during the performance of a maximal voluntary contraction. This interpretation has been challenged by studies in mouse muscle fibres (Place et al. 2008). These studies suggested that changes in the intracellular force-Ca2+ relationship could provide a peripheral mechanism by which interpolated twitches could grow in size with fatigue. However, interpolated twitches do not grow in human adductor pollicis when the muscle is fatigued by a sustained contraction evoked by constant frequency stimulation of the ulnar nerve (Gandevia et al. 2013). Rather, the twitches diminish, in part because of failure of the muscle fibre action potential. Therefore, at least for sustained maximal efforts, growth of the interpolated twitch can only occur with a reduction in neural drive to the muscle and so reflects a central component of fatigue. The motoneurones are one site at which central fatigue may impair motor output. Our recent studies have shown that the responses of motoneurones to single corticospinal volleys are reduced during fatiguing maximal and submaximal isometric contractions (McNeil et al. 2009, 2011a). The reduction was most pronounced when descending voluntary drive was transiently interrupted but some reduction was also seen when voluntary drive was ongoing. These changes are unlikely to be the result of altered muscle spindle or small-diameter muscle afferent feedback (McNeil et al. 2011b). Furthermore, during submaximal contractions, smaller responses were more affected than larger responses. This differential effect suggests that not all of the motoneurones in the pool become less responsive but that it is mainly the motoneurones that are active during the contraction. The altered input-output relationship of the motoneurones may contribute to the mismatch between perceived effort and the level of voluntary EMG that occurs during sustained submaximal efforts. The firing of small-diameter muscle afferents is perceived as muscle pain and/or fatigue and causes reflex cardiorespiratory responses to exercise. For the elbow flexor muscles, the ability of subjects to drive the muscle maximally in isometric maximal voluntary efforts remains low after a fatiguing contraction if the firing of fatigue-sensitive muscle afferents is maintained by block of blood flow to the fatigued muscle. Thus, these afferents contribute to central fatigue. However, this influence appears to be at a supraspinal level (Gandevia et al. 1996). Indeed, the motoneurones of the elbow flexors are facilitated by the afferent firing (Martin et al. 2006). Recent experiments have shown that the supraspinal influence of small-diameter afferent firing is not confined to the fatigued muscle. Block of blood flow to the elbow extensor muscles when they were fatigued resulted in impaired voluntary drive to the non-fatigued elbow flexors, as did maintained firing of small-diameter muscle afferents from the hand. Therefore, afferent feedback from fatigue of one muscle in a limb can alter drive to other muscles in the same limb. This is likely to be important in the development of central fatigue in most coordinated actions. Multiple mechanisms at multiple levels of the nervous system contribute to central fatigue. They include repetitive activity, which creates a specific deficit at the motoneurone level, and sensory feedback from fatigued muscles, which has wider effects on the motoneurones and supraspinally, as well as on autonomic responses to exercise.