A NEW METHOD FOR DETECTING NEURAL INTERCONNECTIVITY

We propose a class of counting process models for analyzing firing tim es of an ensemble of neurons. We allow the counting process intensitie s to be unspecified: unknown functions of the times passed since the m ost recent firings. Under this assumption we derive a class of statist ics with their respective thresholds as well as graphical meth ods for detecting neural connectivity. We introduce a model under which detec tion is shown to be certain for long series of observations. We sugges t ways to classify interactions as inhibition or excitation and to est imate their strengths. The power of the proposed methods is compared b y simulating observations from artificial networks. By analyzing empir ically obtained series we obtain results which are consistent with tho se obtained from cross-correlation-based methods but in addition obtai n new insights on further aspects of the interactions.