The set of equations describing the change of the state of activity of the neurons is
in which a is a time constant, is the strength of the synaptic connection from neuron j to neuron i, and is the additional feedforward input to the neuron besides those described by the feedback connection matrix . A second set of equations describes the way the synapses change with time due to neuronal activity. The learning rule proposed here is
in which B is a time constant and is the feedback learning signal as described in the following.
The feedback learning signal is generated by a Hopfield type associative memory network: , in which is the strength of the associative connection from neuron j to neuron i, which is the recent correlation between the neuronal activities and determined by Hebbian learning with a decay term [5,6,10]
in which is a time constant. The and are only involved in learning and do not directly affect the network outputs.