Figure 14.

Dynamic-clamp acquisition using simulated synaptic conductance trains. NeuroMatic Pulse tab R·P plasticity models were used to simulate 4 statistically different MF inputs with mean spike rate f. The 4 MF inputs were summed together and injected into a real GC via dynamic clamp. (A) Flowchart of the dynamic-clamp experiment. NeuroMatic sent precomputed G_AMPAR, G_NMDAR and G_GABAR waves to 3 DAC channels of a DAQ device, which were connected to the appropriate input channels of a dynamic-clamp SM1 amplifier. In real time, the SM1 amplifier converted the conductances to currents using V_mem of the GC, via the Multiclamp amplifier, and preset reversal potentials, i.e., I = G·(V_mem – V_rev). The sum of the 3 currents was sent to the command input of the Multiclamp amplifier (I_cmd). The Multiclamp amplifier recorded the GC V_mem response to I_cmd and sent V_mem to an ADC channel of the DAQ device, which was read by NeuroMatic. (B) Synaptic event times for 4 simulated MF inputs (top) computed by the Pulse tab. Event times had random Poisson intervals >1 ms with f = 20 Hz. The G_AMPAR train (middle) was computed from the 4 event trains, where each train had its own R·P plasticity model as described in Figure 13, and the resulting 4 conductance trains were summed together. The G_NMDAR train (bottom; before Mg²⁺ block) was computed in the same way, but included facilitation. Parameters for the G_AMPAR and G_NMDAR trains are as reported in Billings et al. (2014). (C) GC V_mem in response to injection of the G_AMPAR and G_NMDAR trains in (B) via dynamic clamp. The injection included a tonic G_GABAR of 0.27 nS. Reversal potentials of the AMPAR, NMDAR and GABAR channels were 0, 0 and −75 mV, respectively. The resting potential of this recording was −75 mV. GC experimental details: cerebellar slice from P24 rat, R_a = 26.7 MΩ, C_mem = 2.6 pF.