One way in which animals localize sounds along the horizon is by detecting the level differences at the 2 ears. Neurons in the lateral superior olive (LSO) encode this cue by integrating the synaptic drive from ipsilateral excitatory and contralateral inhibitory connections. This synaptic integration was analyzed in 400–500-microns brain slices through the gerbil superior olive. Intracellular recordings from LSO neurons were obtained during the application of independent or conjoint electrical stimuli to the excitatory afferent and inhibitory afferent pathways. Stimulation of ascending fibers from the ipsilateral cochlear nucleus reliably evoked EPSPs and action potentials. Stimulation of the medial nucleus of the trapezoid body (MNTB) consistently evoked IPSPs. The evoked postsynaptic potentials differed in that IPSPs were 2 times the duration of EPSPs. An electrophysiological estimate of convergence indicated approximately 10 excitatory and 8 inhibitory afferents per LSO neuron. MNTB stimulation suppressed synaptically evoked action potentials. When stimulus amplitude was increased to the excitatory pathway, it was generally found that a greater MNTB stimulus was necessary to suppress the action potential. A similar commensurate rise in ipsilateral and contralateral acoustic stimulation was also found to be necessary to give the same criterion response. These results confirm that the LSO can integrate evoked action potentials and IPSPs to encode interaural level. Increasing stimulus voltage was found to decrease both action potential and IPSP latency, suggesting that intensity information may be encoded with temporal cues in the nervous system. It was also found that an evoked burst of action potentials could be inhibited in such a way as to yield intermediate discharge rates, dependent on contralateral stimulus level. Taken together, these results suggest that certain properties related to level-difference coding may be available for intracellular analysis using the brain- slice preparation. Several temporal characteristics of the synaptic potentials, including latency and duration, may play a critical role in this simple computation.