Long-term depression (LTD) of glutamatergic and electrotonic transmission can be induced at mixed synapses between eighth nerve fibers and the goldfish Mauth-ner (M) cell in vivo, by pairing weak presynaptic tetani with postsynaptic inhibition. This LTD can be reversed by stronger tetani that produce long-term potentiation (LTP). Moreover, the depression is more likely to occur and tends to last longer when the initial synaptic efficacy is high-that is, if the synaptic strength is first potentiated. In addition, when synaptic efficacy is initially elevated, a weak tetanization that usually results in a gradually developing potentiation instead produces no change in chemical transmission and even a depression of electrotonic coupling. Thus, the modifications in synaptic transmission caused by a certain tetanizing protocol depend upon the history of synaptic efficacy. This last concept provides an experimental basis for theoretical models concerned with pre-and postsynaptic contributions to the regulation of synaptic plasticity.Two opposing forms of activity-dependent modifications of synaptic transmission that are considered to be bases for neural plasticity are long-term potentiation (LTP) (1-4) and long-term depression (LTD) (5-8). One issue of concern has been the identification of conditions which, for the same synaptic connection, differentially favor the induction of one or the other of these alterations. Recently, it was shown that, in visual cortical slices, presynaptic tetani produce either homosynaptic LTD or LTP, depending upon the global level of disinhibition produced pharmacologically or upon the magnitude of postsynaptic depolarization by the tetanus (8). It was also suggested that there are depolarization thresholds for both phenomena, with that for homosynaptic LTD being closer to the resting potential than that for LTP. We provide evidence here for a concept not considered previouslynamely, that the thresholds for the two phenomena might vary with the initial level of synaptic efficacy of the involved connections. Specifically, while studying homosynaptic LTD of both electrotonic coupling and chemical transmission at mixed synapses onto an identified reticulospinal neuron, we found that the amplitude, duration, and sign of long-term changes in synaptic transmission produced by a certain tetanizing protocol depended upon the pretetanization level of efficacy. In addition, it is shown that postsynaptic inhibition is an important factor in induction of LTD in vivo.
MATERIALS AND METHODSLTP and LTD have been studied at the monosynaptic connection between eighth nerve fibers and the Mauthner (M) cell in goldfish (Carassius auratus). As shown in Fig. 1A, single afferents have both gap junctions and chemical synapses on the M cell's lateral dendrite. Standard surgical procedures were used, and intracellular recordings of the eighth nerve responses were obtained in vivo from the lateral dendrite about 200 to 250 ,um from the M cell soma, generally with a KCl (2.5 M)-containing microelectrode of ...