Gamma Oscillations Produce Binomial Spike Statistics for LIF Neurons

(A) Mean spike count plotted against static input signals s without (A = 0, thin line) and with (A = 0.3, thick line) gamma modulation. Note the knee in the curve for mean spike counts of approximately μ = 100 (= N, dashed line) when gamma modulation is present.

(B) Relation between the mean spike count and the spike-count variance for the simulated network without (squares) and with (circles) gamma modulation. Open and closed symbols correspond to c = 0 and c = 0.12, respectively. The dashed line corresponds to the mean-variance relation for a Poisson distributed spike count, while the solid line for a binomial distributed spike count. For a fixed s, we computed the spike-count statistics from 3,000 gamma cycles.