The algorithm of backpropagation error allows training a hidden layer neural network (perceptron type), but still far from the complexity of biological networks in which the property reentrancy is paramount and for which it would be unrealistic want to follow the flow of nerve impulses. This is why global methods have been proposed on networks almost completely connected (ie where the axon of each neuron is connected to almost all the dendrites of other neurons) as the Kohonen networks.


Assuming that an organization is satisfied if it is consistent with its environment (whether to eat his perception of the outcome of its actions on the environment is consistent with them) can model this situation by assuming the flow of actions is consistent with the flow of perceptions. In terms of signal processing we say that these two flows are in phase.
almost periodic signals we seek to align their periods and for any signals one try to parallelize their variations (they wax and wane at the sametime).


I employed this method in the connexionist mime making coherent the video image analyse of the real mime and the analyse of the virtual mime actions.