Spiking statistics of the stimulus and neural populations.

(a) The demonstration network consists of populations of excitatory (red dots), inhibitory (blue dots) and stimulus neurons (black dots). Each neuron connects to neighbours within a fixed range indicated by black circles. Example input connections to single neurons are shown from excitatory (red lines), inhibitory (blue lines) and stimulus (black lines) populations. All synapses are plastic. (b), (c), (d) Spike statistics recorded during a single pass through the dataset for the stimulus (black), excitatory (blue) and inhibitory populations (red), at the beginning of the simulation (dashed lines) and after a 3100 seconds of learning (solid lines). At the beginning of the simulation the network is initialised with very strong synapses from the stimulus which leads to very high firing rates and low variability. After learning variability increases and firing rates decrease. All distributions are across neurons in each population. (b) Distributions of population coefficients of variation of interspike intervals. (c) Distributions of population firing rates. (d) Population Fano-factors after 3100 seconds of learning.