Fig. 8
SOM-interneuron-evoked nonlinear GABAergic inhibition powerfully regulates NMDAR-mediated dendritic spikes. a Computational pyramidal-cell model with stimulated glutamatergic (red) and GABAergic synapses (blue). b The local membrane potential as a result of glutamatergic and GABAergic (open circles) inputs at 20 ms after a short burst stimulation (3@200 Hz) is shown as a heat map across the tuft dendrites. c Membrane potential responses from the same simulation are shown for the left apical dendritic trunk (blue traces) and two tuft dendrites, one without inhibition (left, magenta traces) and one with strong inhibition (right, green traces). Recording sites of the local membrane potential are indicated by colored pipettes in a. The family of different voltage traces show brief-burst activation of an increasing number of glutamatergic inputs. The number of active synapses during the first stimulus (15, 45, 75, 105, 135, and 165) represent about 1–10% of all available tuft synapses. d When the GABAR conductance is constant, membrane potential responses exhibit larger NMDAR-mediated plateau depolarizations in both the disinhibited and the inhibited tuft dendrite. e Inhibition with PV-like dynamics using 5-fold faster time course of the slow component (rise τ = 0.2 ms; decay τ = 6 ms) and 5-times larger peak conductance, fails to prevent an NMDAR-mediated dendritic spike in the inhibited tuft dendrite if the number of glutamatergic inputs exceeds 5.1% active synapses (>75 of 1464 distributed over the whole tuft). f–h The dependence of the burst PSP integral measured in the different compartments on glutamatergic input number for all three conditions. Each point is the mean of five simulations with randomized glutamatergic synapse locations, while the location of GABA synapses was kept constant. The SEM is indicated. magenta: dend1; green: dend2; blue: apical trunk; black: soma
