CPM Seminar

Correlations in point processes, implications for signal processing.

Maurice Chacron

Department of Physiology
McGill University

There has been much interest to study problems related to the escape from metastable states in recent years. In many cases, the output consists of series of stereotyped responses which can be described by a point process. However, many such processes are correlated over various time scales and must furthermore operate under the presence of significant noise. Here we present theoretical results supported by numerical simulations and experimental data on the effects of these experimentally observed correlations on signal transmission. We show that correlations will shape the noise landscape over different frequency ranges, thus allowing greater or worse signal transmission. Our results have a wide range of applications in different systems such as Sigma-Delta Modulators, Josephson junctions, and neural systems.