Combinatorial analysis of temporal sequences for pulse-transmission neural networks

Summary form only given, as follows. An approach is described to analyzing temporal patterns in pulse-transmission neural networks. Such networks appear in biology and in neural models in which processing elements transmit pulses to one another. Combinatorial mathematics is used to characterize the amount of freedom there is in specifying temporal subsequences in impulse trains. A set of dependency equations is described and related to a test for favored and suppressed temporal patterns. The concepts of independently prescribable sets and good graphs are given and related to temporal sequences.<>