Figure 5. The relative contribution of input pathways with different temporal coherence and initial synaptic strength to postsynaptic spike events is dependent on inhibition.

In these simulations, P1 (red, σ = 3) is more coherent than P2 (blue, σ = 6), as in Input Regime III (Figure 3G), but is weaker (ISS_P2/P1 = 1.5). The relative number of presynaptic spikes in P1 or P2 that preceded postsynaptic spike events was counted in 1 millisecond time bins and plotted as a function of time preceding the postsynaptic spike for two levels of inhibition, low (A,C,E, gI/gmax_ex = 0.264) and high (B,D,F, gI/gmax_ex = 0.792). The analysis was first done in the absence (A, B) and then in the presence (C–F) of the STDP learning rule, (C–F) are examples of individual STDP simulations trials. At a low level of inhibition, P2 had a slight advantage in all time bins in the absence of STDP (A), and went on to dominate in 50% of the STDP trials; an example of P2 domination shown in (C, E). At a high level of inhibition, P1 had a slight advantage over P2 in contributing to postsynaptic spike events 2–8 ms preceding the postsynaptic spike (B); and this advantage was dramatically magnified by STDP (D,F). (G–H) The total number of presynaptic spikes in P2 that followed postsynaptic spike events is greater than that in P1 in the absence of STDP learning rules. The relative number of presynaptic spikes was calculated as in (A–B), except that 1 ms bins counts were made for a time window of 20 ms following postsynaptic spikes.