A, systematically co‐varying and revealed that distinct regions in this 2‐D parameter space yield different spiking patterns. Boundaries shown here are based on testing with I
stim
= 60 μA cm−2. Traces in a–d show sample responses for parameter values labelled on the main plot. Large parameter variations that remain within a region yield the same spiking pattern; compare condition a ( = 3 mS cm−2, = 4 mS cm−2) with condition b ( increased by 4 mS cm−2). In contrast, small parameter variations that cross a boundary yield different spiking patterns; compare condition a with condition c ( reduced by 0.5 mS cm−2) or condition d ( increased by 0.5 mS cm−2). B, boundaries can shift because of stimulus intensity (I
stim), meaning a neuron with fixed values of and
K,A can exhibit different spiking patterns at different I
stim. To illustrate, each vertical arrow on the left panel represents a neuron: for neuron a, = 3 mS cm−2 and = 4 mS cm−2; for neuron b, = 3.5 mS cm−2 and = 2.5 mS cm−2. The spiking pattern at each I
stim (illustrated on the right) depends on which region the arrow passes through. C, boundaries can also shift because of pre‐stimulus membrane potential. For these simulations, a subthreshold ‘pre‐pulse’ (I
pre) was used to vary the membrane potential before the onset of suprathreshold stimulation. Each plane represents the response to I
stim
= 60 μA cm−2 after a different pre‐pulse (pre‐stimulus membrane potential is indicated beside each voltage trace). The vertical arrow represents a neuron with = 2 mS cm−2 and = 6 mS cm−2. Traces on the right show the reduced availability of g
K,A depending on I
pre. By partially inactivating g
K,A, subthreshold depolarization reduces the availability of those channels for activation during suprathreshold stimulation, effectively re‐scaling the y‐axis.