Inspiratory muscle output is downregulated when the mechanical load is reduced in awake humans. It is not known whether this is related to reduction in PCO2 or to removal of load-related neural responses. To address this issue, we did Read CO2 rebreathing tests in 13 normal subjects with and without unloading and compared respiratory output at identical end-tidal PCO2 (PET(CO2)) levels. Unloading was carried out with proportional assist ventilation (flow assist = 2 cm H2O/L/s plus volume assist = 4 cm H2O/L, representing approximately 50% reduction of the normal resistance and elastance). Ventilatory output (n = 13), total pressure of respiratory muscles (Pmus, n = 8), and transdiaphragmatic pressure (Pdi, n = 5) were computed at different PET(CO2) levels. Pmus was computed from esophageal pressure (Pes) using the Campbell diagram, and Pdi was measured from the difference between gastric pressure and Pes. Unloading caused an increase in ventilation (VI) and tidal volume (VT) at all PET(CO2) levels with no significant effect on slope (VI/PET(CO2) or VT/PET(CO2)) or respiratory rate. At low PET(CO2) (50 mm Hg), Pdi and Pmus waveforms did not differ with and without unloading. At high PET(C02) (59 mm Hg), peak Pdi and Pmus decreased by only 18.8 +/- 8.3% and 13.8 +/- 9.5%, respectively (NS, p > 0.05). Using a model that allows nonlinearity in the pressure-volume relation and for intrinsic muscle properties (force-length and force-velocity relations), we estimated the expected changes in mean VT and VI when the level of assist used in this study was applied in the absence of any change in neural output response to CO2. The predicted and observed changes in VT and VI were similar. We conclude that when chemical stimuli are rigorously controlled, unloading does not result in downregulation of respiratory muscle activation.