Fig 5. The neural resource model.

(a & b) Stimulus features were encoded in the activity of idealized contrast sensitive and orientation selective neurons; preferred orientations were evenly distributed in orientation space. (c) Normalisation operated across the whole population, scaling summed activity to a fixed level determined by a gain constant. (d) Neurons generated spikes stochastically according to a Poisson process, with mean firing rate determined by the normalized input of each neuron. Orientation recall was modeled as maximum likelihood decoding of the spiking activity over a fixed time window. (e) Examples of error distributions predicted by the neural resource model.