The aims of this study were to (1) quantify any central fatigue that occurs following prolonged dynamic exercise, i.e. reduced muscle force caused by impaired motor drive from the central nervous system and (2) determine whether decreased cortical arousal, assessed using critical flicker fusion threshold (CFF), occurs and is related to impaired exercise performance. Fifteen healthy men cycled at 70% VO2peak until exhaustion. The peak force of maximum voluntary isometric contractions (MVC) of the quadriceps muscle group was reduced by 30% at exhaustion. The voluntary activation ratio determined using superimposed tetanic stimulation fell from 0.99 to 0.86 at exhaustion. The central fatigue (%) at exhaustion was 33+/-12% (+/- SD) (assessed via the tetanus interpolation technique) and 54+/-32% (assessed via the relative decline of MVC and peak tetanic force) of the total fatigue. The MVC only partially recovered and central fatigue persisted at 30 min post-exercise. CFF increased from 39.2+/-2.3 to 41.8+/-3.5 Hz at exhaustion, but did not correlate with central fatigue. Every subject reached the highest rating of perceived exertion (RPE) at exhaustion of 20 on the Borg scale. The time to exhaustion was related to how quickly the RPE increased and to the ability to sustain exercise at very high RPE. These data suggest that with prolonged cycling: (1) there is considerable and a persistent form of central fatigue, (2) there is an increased level of cortical arousal, and (3) exhaustion is linked to very high subjective RPE.