To determine whether hypoxia has a direct influence on the central command independently of the working muscles, 16 subjects performed intermittent isometric unilateral knee extensions until exhaustion either in normobaric hypoxia (inspired O(2) fraction=0.11, arterial oxygen saturation approximately 84%) or in normoxia while the knee extensor muscles were exposed to circulatory occlusion with a 250 mmHg cuff. Among the subjects, 11 also performed the tests in hypoxia and normoxia without occlusion. Single electrical stimulations were regularly delivered to the femoral nerve to measure the changes in the knee extensor peak twitch force. With the cuff, the average slope of decrease in peak twitch did not depend on the inspired oxygen fraction. Performance was slightly but significantly lower during hypoxia than in normoxia (8.2+/-2.6 vs 9.4+/-3.1 repetitions, P<0.05) with the cuff on. The number of repetitions was much higher during hypoxia with maintaining leg blood flow (15.6+/-4.5 repetitions) than with circulatory occlusion in normoxia. In conclusion, this study showed that a direct effect of hypoxia in reducing the motor drive to the working muscles exists but this effect is moderate.