Aim: The reduction in cerebral oxygenation (Cox) is associated with the cessation of exercise during constant work rate and incremental tests to exhaustion. Yet in exercises of this nature, ecological validity is limited due to work rate being either fully or partly dictated by the protocol, and it is unknown whether cerebral deoxygenation also occurs during self-paced exercise. Here, we investigated the cerebral haemodynamics during a 5-km running time trial in trained runners. Methods: Rating of perceived exertion (RPE) and surface electromyogram (EMG) of lower limb muscles were recorded every 0.5 km. Changes in Cox (prefrontal lobe) were monitored via near-infrared spectroscopy through concentration changes in oxy- and deoxyhaemoglobin ($δ$[O 2Hb], $δ$[HHb]). Changes in total Hb were calculated ($δ$[THb] = $δ$[O2Hb] + $δ$[HHb]) and used as an index of change in regional blood volume. Results: During the trial, RPE increased from 6.6 ± 0.6 to 19.1 ± 0.7 indicating maximal exertion. Cox rose from baseline to 2.5 km (↑$δ$[O2Hb], ↑$δ$[HHb], ↑$δ$[THb]), remained constant between 2.5 and 4.5 km, and fell from 4.5 to 5 km (↓$δ$[O2Hb], ↑$δ$[HHb], ↔$δ$[THb]). Interestingly, the drop in Cox at the end of the trial coincided with a final end spurt in treadmill speed and concomitant increase in skeletal muscle recruitment (as revealed by higher lower limb EMG). Conclusion: Results confirm the large tolerance for change in Cox during exercise at sea level, yet further indicate that, in conditions of self-selected work rate, cerebral deoxygenation remains within a range that does not hinder strenuous exercise performance. © 2009 Scandinavian Physiological Society.