Determine if a series of trials with fraction of inspired oxygen (FiO2) content deception could improve 4000-m cycling time-trial (TT) performance. Fifteen trained male cyclists (mean ± SD: body mass 74.2 ± 8.0 kg; peak oxygen uptake 62 ± 6 mL.kg-1.min-1) completed six, 4000-m cycling TTs in a semi-randomised order. After a familiarisation TT, cyclists were informed in two initial trials they were inspiring normoxic air (NORM, FiO2: 0.21), however in one trial (deception condition) they inspired hyperoxic air (NORM-DEC, FiO2: 0.36). During two subsequent TTs, cyclists were informed they were inspiring hyperoxic air (HYPER, FiO2: 0.36), but in one trial normoxic air was inspired (HYPER-DEC). In the final TT (NORM-INFORM) the deception was revealed, and cyclists were asked to reproduce their best TT performance while inspiring normoxic air. Greater power output and faster performances occurred when cyclists inspired hyperoxic air in both truthful (HYPER) and deceptive (NORM-DEC) trials compared to NORM (P < 0.001). However, performance only improved in NORM-INFORM (377 W [95% CI 325, 429]) vs NORM (352 W [299, 404]), P < 0.001) when participants (n = 4) completed the trials in the following order: NORM-DEC, NORM, HYPER-DEC, HYPER. Cycling performance improved with acute exposure to hyperoxia. Mechanisms for the improvement were likely physiological, however improvement in a deception trial suggests an additional placebo effect may be present. Finally, a particular sequence of oxygen deception trials may have built psycho-physiological belief in cyclists such that performance improved in a subsequent normoxic trial.