Fig. 2.

Relationship between the point process model and a soft-threshold integrate and fire model. A: illustration of the function g[I(t)] in Eq. 17. The parameter β controls the summation time constant of the model by decreasing the slope of g[I(t)] for negative currents. Refractory effects reduce the excitability of the neuron by decreasing the slope for all current values. B: illustration of the probability density function for the noisy variable η in the stochastic threshold analogy (see text). Smaller values of α generate broader distributions or equivalently a more variable spike initiation mechanism.