(A) Diagram illustrating the conductance-based computational model that we used to model the effect of migraine mutations and of Hm1a by increasing the persistent Na⁺ current of the GABAergic neuron. gD,e and gD,i are glutamatergic conductances that model the baseline excitatory inputs (excitatory drive) of the pyramidal and of the GABAergic neuron, respectively. (B) Simulation, with GABAergic neuron’s physiologic persistent Na⁺ current (1%), of the effect of a constant external depolarizing input to the GABAergic neuron (gD,i = 0.1 mS/cm²) without input to the pyramidal neuron (gD,e = 0 mS/cm²): membrane potential of the GABAergic neurons (upper), membrane potential of the pyramidal neuron (middle), and extracellular K⁺ concentration (lower). (C) Same simulation with increased GABAergic neuron’s persistent Na⁺ current (6%, mimicking the effect of FHM3 mutations and of Hm1a); the right panels display with an enlarged time scale the first phase of the simulation shown in the panels on the left. The vertical dotted line indicates the beginning of the large [K⁺]_out increase that leads to depolarizing block. (D) Effect of an increase of the GABAergic neuron’s persistent Na⁺ current on the lowest external input to the GABAergic neuron (gD,i) sufficient to induce depolarizing block. (E) Effect of an increase of the GABAergic neuron’s persistent Na⁺ current on the latency of depolarizing block with gD,i = 0.349 mS/cm² (the lowest input to the GABAergic neuron able to generate CSD with 0% persistent Na⁺ conductance; see panel D). (F) Effect of the amount of persistent current on the firing frequency of the GABAergic neuron, in a simulation in which the pyramidal neuron was removed, reflecting the direct effect of the persistent current on the firing properties of the GABAergic neuron.