The purpose of the study was to examine the effects of chemical feedback on respiratory motor and ventilatory output in conscious subjects ventilated on various modes of assisted mechanical ventilation. Seven subjects were connected to a ventilator and randomly ventilated on assist-volume control (AVC), pressure support (PS) or proportional assist ventilation (PAV). On each mode, the assist level was set to the highest comfortable level. Airway and oesophageal (Poes) pressures, tidal volume, respiratory frequency (fR) and end-tidal carbon dioxide tension (PET,CO2) were measured breath-by-breath. When the subjects were stable on each mode, the fraction of inspired carbon dioxide (FI,CO2) was increased stepwise, and changes in minute ventilation (V'E) and respiratory motor output, estimated by the pressure-time product of all the respiratory muscles per breath (PTPrm) and per minute (PTPminute), were observed. At zero FI,CO2, PTPminute/PET,CO2 did not differ between modes, while V'E/ PTPminute was significantly lower with PAV than that with PS and AVC. As a result V'E/PET,CO2 was significantly lower with PAV, preventing, unlike AVC and PS, a significant drop in PET,CO2. With PAV, independent of CO2, V'E/PTPminute remained constant, while it decreased significantly with increasing CO2 stimulus with PS and AVC. At high PET,CO2 respiratory effort was significantly lower with PAV than that with PS and AVC. In conclusion, the mode of mechanical ventilation modifies the effects of chemical feedback on respiratory motor and ventilatory output. At all carbon dioxide stimulus levels neuroventilatory coupling was better preserved with proportional assist ventilation than with pressure support and assist-volume control ventilation.