The construction of a simultaneous functional order in nervous systems. I. Relevance of signal covariances and signal coincidences in the construction of a functional order (Q1094346)

A broad range of algebraic and topological concepts and constructions is used to transform successively the information given by covariances and coincidences of signals between neurons into structures which may represent functional and even anatomical (geometrical) order in the network. Thus it is possible for the brain to develop functionally relevant and structurally efficient connections by self organizing mechanisms, depending on activities directly perceived by the neural system. The presented results thus support the hypothesis that algorithms rather than specific neuronal connections are genetically preprogrammed. In Part I, an unordered collection of neurons is transformed into a poset of blocks, two neurons of the same block having the same covariance/coincidence relation to all other elements of the block. By neuron merging and recruitment, the poset is further converted into a lattice. In Part II, see the following entry, Zbl 0629.92006, functional analogies for some geometrical concepts like `between', `dimension' etc. are defined and computed from the lattice order; certain subsets of the lattice are identified with an abstract geometrical complex, reflecting the underlying structure of detectors.