Statistical Estimation of the Interaction Graph
摘要
The classical question of inferring the network structure from the observation of spike sequences of (some of) its neurons is briefly reviewed. It is then studied in the discrete time framework of Chap. 2 , leading to rigorous results in a remarkably economical way. A less economical, but much more general approach—thanks to the concepts of stochastic chains with memory of variable length and the context algorithm—is presented in detail in Appendix F.