An autocatalytic model of STDP timing from slow calcium-dependent signals

Data of spike timing-dependent plasticity (STDP) show a sharp temporal transition between potentiation and depression despite a relatively slow time course of calcium concentration. We show how autocatalytic amplification of initial concentration differences can enable a high degree of temporal selectivity and produce the sharp STDP weight change curve despite having a relatively slow time constant. This simple model is robust to parameter changes, noise and details of the model. The model correctly predicts the location of the maximum and minimum for STDP at ±10ms from coincidence.