Presynaptic activity in the inhibitory network impinging on the Mauthner (M-) cell was investigated in the goldfish medulla in vivo using extra- and intracellular recordings. The inhibitory presynaptic volley elicited by stimulation of the contralateral vestibular nerve consisted of multiple successive peaks at high frequency (up to 1,000 Hz). Less pronounced multicomponent responses were recorded after antidromic activation of the M-cell. Such high-frequency "oscillatory" field potentials also occurred spontaneously. 2. In intracellular recordings, a subset of inhibitory interneurons showed evoked and spontaneous burst discharge. Burst action potentials were correlated with the peaks in the extracellular volley, suggesting that repetitive firing of these cells is synchronized. Nonbursting cells, on the other hand, fired single action potentials in response to vestibular stimuli and were not activated via the M-cell collateral network. 3. Bursting cells were determined morphologically to be part of the feedback inhibitory circuit. Their responses to stimulation of the contralateral vestibular nerve thus suggest the existence of a crossed excitatory pathway to these interneurons. 4. Vestibular-evoked excitatory postsynaptic potentials (EPSPs) in bursting interneurons had a short latency of 0.781 +/- 0.08 ms (mean +/- SD, n = 18) but reached threshold at 2.25 +/- 1 ms (n = 21). These characteristics are suggestive of a chemically mediated EPSP. Indeed, the evoked synchronous repetitive activity of these cells was prevented by superfusion with excitatory amino-acid receptor antagonists. 5. Bursting neurons showed several characteristics that differentiate them from nonbursting cells, including brief action potentials, plateau responses, and intense spontaneous subthreshold activity. 6. With extracellular recordings, tetanization of contralateral vestibular primary afferents evoked a long-lasting potentiation of oscillatory population responses in 11 of 27 cases. Furthermore in three experiments, the frequency of occurrence of spontaneous bursts was enhanced and a similar facilitation was detected at the intracellular level. 7. We conclude that a subset of interneurons in this inhibitory network is capable of repetitive discharges and that evoked as well as spontaneous firing in this population is synchronized. Although electrical coupling between interneurons may mediate synchronization and intrinsic membrane properties may promote burst activity, our data suggest strongly that repetitive firing requires chemically mediated transmission. Furthermore they indicate that the mechanisms underlying evoked as well as spontaneous bursting in this population show activity-dependent plasticity.
