Comparisons between the spike discharge of inspiratory neurons within the retrofacial area (RFN), and the membrane potential of expiratory neurones within the caudal medulla were made in pentobarbitone-anaesthetized, vagotomized, artificially ventilated cats. Spike-triggered averaging (STA) of synaptic potentials, triggered by the discharge of inspiratory RFN neurones, was utilized to test for synaptic connectivity. 2. Eighty-nine neurons with respiratory-phased discharge patterns were recorded in the vicinity of the RFN. Fifty-four neurones discharged at or slightly before the onset of the inspiratory burst activity of the phrenic nerve and continued firing throughout inspiration. Two continued to fire during post-inspiration. Forty-five of fifty-four inspiratory RFN neurones exhibited incrementing discharge patterns, six discharged with a plateau pattern, while only three neurones had a decrementing discharge pattern. 3. The membrane potential trajectories of caudal expiratory neurones revealed a typical wave of early inspiratory hyperpolarization. Occasionally, a second wave of hyperpolarization occurred during late inspiration, in conjunction with increased phrenic nerve activity. 4. Spike-triggered averaging revealed averaged inhibitory postsynaptic potentials (IPSPs), indicative of inhibitory synaptic connections, between eight and sixty-three pairs of RFN inspiratory and caudal expiratory neurones. 5. Inhibitory postsynaptic potentials detected by STA exhibited a relatively long latency and a slow time course. The IPSPs began, on average, 3.8 ms after an RFN action potential. The rise times, half-widths and durations of IPSPs were longer than expected for a monosynaptic somal input from myelinated axons of inspiratory RFN neurones. It is suggested that an inhibitory relay neurone in the immediate vicinity of the expiratory neurones is activated by a collateral of the RFN inspiratory neurone. 6. Retrofacial inspiratory neurones were antidromically activated only when high-intensity electrical stimulation was applied in the vicinity of caudal expiratory neurones. 7. The averaged IPSPs were preceded by diphasic and triphasic 'spike potentials'. The averaged spike potentials were highly entrained to the action potentials of RFN inspiratory neurones which triggered IPSPs. The spike potentials may be terminal potentials recorded from axons of RFN inspiratory neurones. 8. Evidence for convergence of synaptic inputs was obtained from STA tests in a caudal expiratory neurone receiving IPSPs from four RFN neurones. 9. The functional significance of this observation is discussed. We conclude that RFN inspiratory neurones exert a moderate inhibitory influence and act conjointly with other types of medullary inspiratory neurones.
