Fig 1. Model framework and dynamics.

A. The model consists of a layer of input neurons, along which wavefronts of spiking activity propagate. Input neurons undergo a burst of spikes, generated stochastically, when the wavefront passes (red unit). Input spikes elicit excitatory postsynaptic potentials (EPSPs) in the output neuron (blue unit) that are scaled by the synaptic strength, w _ij. The output neuron generates spikes stochastically with a firing rate linearly proportional to its summed EPSP. B. Example spike raster from the simulation, showing one wave traveling past input neurons 1–10. C. Synaptic strengths are modified by either an asymmetric (top) or symmetric (bottom) STDP rule. D. Schematic for the influence of the wavefront on modifications at surrounding synapses. An input neuron (red) is recruited by a wave, which travels from left to right, and increases the firing rate of the output neuron (blue). When an asymmetric STDP rule is at play, output spikes at the current time point will cause synapses behind the wavefront to increase in strength (‘+’ symbol) because they were active at an earlier time. Similarly, synapses in front of the wave will decrease (‘−’ symbol), because they will become active at a later time. Thus, traveling waves map the STDP rule onto space. E. Schematic for the influence of surrounding inputs (colored light red), on synaptic modifications at the wavefront (dark red). F. Same as E, except that inputs to the right of the wavefront induce even greater synaptic strengthening as their respective synapses are stronger. G. Same as E, except for a symmetric STDP rule. Here, the greater strength of synapses either side of the wavefront induce more synaptic weakening.