Skip to main content
. 2013 Nov 21;7:170. doi: 10.3389/fncom.2013.00170

Figure 8.

Figure 8

Susceptibility in the CD model under varying conditions. For the base parameters for SC23 we repeated the simulations to compute the rate of weight change under different conditions. Top row: Weight change with increasing baseline firing rate (A: 5 Hz, B: 20 Hz, and C: 40 Hz). At low firing rates, the synapse behaves similar to linear spSTDP; maximal LTP occurs at Δφ ≈ π/6 (compare Figure 4). With increasing firing rates, maximal potentiation also increases, and maximal LTP shifts toward slightly negative phase shifts (20 Hz: Δφ ≈ 0 40 Hz: Δφ ≈ −π/4), while being centered at fmod ≈ 5 Hz. Bottom row: Influence of model constituents on susceptibility. Neurons fire at an average rate of 5 Hz. (D) CD model without presynaptic adaptation. The synapse is strongly sensitive to Δφ, however, there is no lower bound on fmod. (E) Without postsynaptic adaptation. The susceptibility of the synapse is very similar to that of the undisturbed model (compare A). (F) Without activation q. The weight change is negative everywhere, and the synapse is not susceptible for some intermediate fmod. Plots (D,F) change only slightly with increasing baseline firing rates. We conclude that of the model constituents, postsynaptic adaptation is not necessary to explain theta susceptibility in the non-linear regime.