Remembering and Planning: A Neuronal Network Model for the Selection of Behaviour and Its Development for Use in Human Language

A description is given of the frame/content relationship between the various temporal orders of magnitude in behaviour, from long term contexts down to the duration of single synaptic potentials. The time frame model of motor organization, consisting of nested levels operating at different time scales, is based on this description and deals with how long term contexts influence instantaneous motor output. This model can be simulated by a neuronal network based on the Hopfield algorithm. The time frame approach and the same algorithm can also be used for memory, resulting in a general model for remembering and planning actions in animals. Simple elaborations of this model, consisting of replication and new interconnections of existing components, lead to symbolization and increased categorising power and so allow application of the model to human memory and motor organization. These elaborations are both evolutionarily plausible and supported by neurological findings. Further, it is shown that language can also be described in terms of time frames and that words may be organized in the same way as actions. Thus, the model can also be extended to remembering and planning choice of words in language.