The medial nucleus of the trapezoid body (MNTB) plays an important role in the processing of interaural intensity differences, a feature that is critical for the localization of sound sources. It is generally believed that the MNTB functions primarily as a passive relay in converting excitatory input originating from the contralateral cochlear nucleus (CN) into an inhibitory input to the ipsilateral lateral superior olive. However, studies showing that the MNTB itself is also the target of inhibitory input suggest that the MNTB may serve more than a sign-converting function. To examine the fidelity of signal transmission at the CN-MNTB synapse, presynaptic calyceal potentials ("prepotentials"), reflecting the excitatory input to the MNTB neuron, and postsynaptic action potentials were simultaneously monitored with the same electrode during in vivo extracellular recordings from the gerbil's MNTB. Presynaptic activity differed from postsynaptic activity in several respects: (1) Spontaneous and sound-evoked discharge rates were greater presynaptically than postsynaptically. (2) Frequency tuning was sharper postsynaptically than presynaptically. (3) Calyceal terminals and MNTB neurons both showed phasic-tonic response patterns to tonal stimulation, but the duration of the onset response and the level of the tonic component were reduced postsynaptically. (4) Phase-locking to sound frequencies up to 1 kHz was greater postsynaptically than presynaptically. (5) The rate-intensity characteristics of pre- and postsynaptic activities differed significantly from each other in half of the MNTB neurons. To test the hypothesis that acoustically evoked inhibition of MNTB neurons contributed to the relatively lower levels of postsynaptic discharge, two-tone stimulation was applied, wherein the response to one tone-burst, set at the neuron's characteristic frequency, can be reduced by addition of a second "inhibitory" tone. The inhibitory tone caused a much larger reduction in post- than in presynaptic activity, indicating an acoustically evoked inhibitory influence directly on MNTB units. These findings show that transmission at the CN-MNTB synapse does not occur in a fixed one-to-one manner and that the response of MNTB neurons reflects the integration of their excitatory and inhibitory inputs.