Increased wakefulness is known to suppress the initial ventilatory response to passive movement and the steady-state ventilatory response to exercise. However, the effect of increased wakefulness upon the integrated ventilatory response at the onset of exercise is not known. We hypothesized that increasing wakefulness via a cognitive task would attenuate the initial ventilatory response to exercise, and so we examined the response to active leg extensions under two conditions: with and without concurrently solving a puzzle. At rest before exercise, subjects demonstrated greater minute ventilation while solving a puzzle (mean ± S.E.M., 12.38 ± 0.55 versus 10.12 ± 0.511 min-1, P < 0.001), due to a higher mean breathing frequency (mean ± S.E.M., 17.1 ± 0.93 versus 13.6 ± 0.59 breaths min -1, P < 0.001). At the start of exercise, subjects did not increase their ventilation significantly while solving the puzzle (P = 0.170), but did by a mean ± S.E.M. of 6.16 ± 1.121 min-1 (P < 0.001) when not puzzle solving. The ventilation achieved at the start of exercise in absolute terms was also lower while solving the puzzle (14.6 ± 1.1 versus 16.3 ± 1.31 min-1, P = 0.047). Despite differences in the rapid ventilatory response to exercise between conditions, the steady-state responses were not different. We conclude that the performance of a cognitive task decreases the initial phase of exercise hyperpnoea, and suggest that this might occur because of either a competitive interaction between drives to breathe or a behavioural distraction from the 'task' of exercise.