Figure 3

Figure 3 Exploration and optimization of model parameters.

(A) Burst rate (left) and number of participating neurons (right) as a function of inhibitory space constant (σ ), inhibitory time constant (τ^(I)) and inhibitory gain (g^(I)). Red star indicates the parameters of the baseline model (as per Figure 2), and black regions indicate parameter combinations that failed to produce bursting. Dashed squares indicate regions approaching uniform inhibition (red), non-decaying inhibition (orange) and lack of inhibition (green). (B–D) Model simulation evaluated against optimization objectives: post-burst activity suppression (B), burst sizes distribution (C) and burst durations distribution (D). (E–F) Model simulation vs experimental data for two additional features: Burst-triggered burst participation (E, see Figure 1L) and burst temporal profile (F, see Figure 1H). (G–H) Temporal (G) and spatial (H) profiles of the intercellular interactions in the optimized model. (I) Timeseries illustrating the influence of a single spike (in cell marked with red spot) on the linear drive of surrounding cells in the optimized model network. Plots show mean and shaded areas indicate SD for n=10 simulation runs.