In healthy humans, cerebral oxygen desaturation during exercise affects motor unit recruitment, while oxygen supplementation enhances cerebral oxygenation and work capacity. It remains unknown whether in patients with chronic obstructive pulmonary disease (COPD), the well-documented improvement in exercise tolerance with oxygen supplementation may also be partly due to the increase in cerebral oxygenation. Using near infrared spectroscopy, we measured both frontal cerebral cortex blood flow (CBF) using indocyanine green dye and cerebrovascular oxygen saturation (S(t,O(2))) in 12 COPD patients during constant-load exercise to exhaustion at 75% of peak capacity. Subjects exercised while breathing air, 100% oxygen or normoxic heliox, the latter two in balanced order. Time to exhaustion while breathing air was less than for either oxygen or heliox (mean±sem 394±35 versus 670±43 and 637±46 s, respectively). Under each condition, CBF increased from rest to exhaustion. At exhaustion, CBF was higher while breathing air and heliox than oxygen (30.9±2.3 and 31.3±3.5 versus 26.6±3.2 mL·min(-1) per 100 g, respectively), compensating for the lower arterial oxygen content (C(a,O(2))) in air and heliox, and leading to similar cerebral cortex oxygen delivery (CQ(O(2)) for air was 5.3±0.4, for oxygen was 5.5±0.6 and for heliox was 5.6±1.0 mL O(2) per min per 100 g). In contrast, end-exercise S(t,O(2)) was greater while breathing oxygen compared with air or heliox (67±4 versus 57±3 and 53±3%, respectively), reflecting C(a,O(2)) rather than CQ(O(2)). Prolonged time to exhaustion by breathing oxygen and heliox, despite these having a similar CQ(O(2)) to air, a lower S(t,O(2)) with heliox than oxygen, and yet similar endurance time and similar S(t,O(2)) in air and heliox despite greater endurance with heliox, do not support the hypothesis that an improvement in cerebral cortex oxygen availability plays a contributing role in increasing exercise capacity with oxygen or heliox in patients with COPD.