001 (measured at a resting potential of −54.2 ± 0.4 mV; input resistance 32.7 ± 0.6 MΩ, n = 111 pairs). We then activated excitatory synaptic input to the neurons and increased stimulus strength until the evoked excitatory postsynaptic potentials (EPSPs) reliably induced spiking (spiking probability averaged over both cells ≥0.5). For each pair, we recorded the electrical coupling by alternating delivery of negative current pulses to each cell for a baseline period.

This was followed by an induction protocol consisting of 50 synaptic stimuli at 1 Hz, with steady-state depolarizing current such that the neurons fired the characteristic burst of spikes observed in olivary neurons in response to sensory stimulation in vivo (Chorev et al., LY294002 2007 and Khosrovani Dabrafenib et al., 2007). Following the induction protocol, the average coupling coefficient was significantly lower in nine out of ten pairs of connected cells (Figure 1; average reduction of 47% ± 9.9% from an initial

coupling coefficient of 0.012 ± 0.003; p < 1 × 10−5, n = 10 pairs). This depression of coupling was sustained for more than 15 min, with the longest recordings showing plasticity 25 min after induction. Consistent with the reduction in coupling, input resistance was also increased following induction (mean input resistance 38.4 ± 18 MΩ after induction; 21.8% ± 7.5% increase; p < 0.01). Only small changes were seen in resting membrane potential (6% ± 3% hyperpolarization, n = 20 cells, p = 0.067) and sag ratio (decrease by 13.9% ± 4.5%, n = 20 cells, p = 0.054)

following induction. Changes in coupling can result from either changes in input resistance or junctional conductance, and this conductance can be estimated indirectly by combining transfer resistances and input resistance. Using this estimate of gap junctional conductance confirms that coupling remains significantly reduced after induction (48% ± 12% reduction; p < 1 × 10−4; Figure S1A available online). Experiments were also performed in voltage clamp to provide a more direct readout of coupling (Figures S1B–S1D). Cells were held at −55 mV during the baseline and postinduction period. For plasticity induction, synaptic stimuli were paired with short (10 ms), 10 mV depolarizations to allow the cells to fire bursts of unclamped spikes. After induction, coupling was significantly first reduced (by 15% ± 1% of control; p < 0.01, n = 7 cells), consistent with our current-clamp experiments. Finally, we tested the effect on coupling coefficient of higher-frequency olivary spiking in the presence of more intense synaptic input. Spikes at 4 Hz, evoked by depolarizing current pulses, were paired with 25 Hz bursts of synaptic input timed to provide synaptic glutamate release throughout the postsynaptic spike (Figure S2A). This “theta”-like activity was designed to mimic pairing protocols that are typically used to induce synaptic plasticity in other brain areas.