A circle criterion observer for estimating the unmeasured membrane potential of neuronal populations

Michelle Chong (1) , Romain Postoyan (2) , Dragan Nesic (1) , Levin Kuhlmann (1, 3) , Andrea Varsavsky (1, 4)

1 Department of Electrical and Electronic Engineering [Melbourne]
2 CRAN - Centre de Recherche en Automatique de Nancy
3 Department of Optometry and Vision Science
4 Department of Otolaryngology

A circle criterion observer is designed for estimating the unmeasured membrane potential of neuronal populations using the electroencephalogram (EEG) from a class of parametric models that replicates patterns seen on the EEG. Compared to existing similar designs, we provide a less conservative linear matrix inequality (LMI) condition that is shown to be fulfilled for the neural models we consider. The designed observer is robust towards disturbance in the input and measurement, as well as model uncertainty. We show that the observer can be designed for a model that reproduces alpha rhythms in the EEG as an illustrative example.

Mots clés: Nonlinear observer; neural mass models

Domaines: Automatique / Robotique

Michelle Chong, Romain Postoyan, Dragan Nesic, Levin Kuhlmann, Andrea Varsavsky. A circle criterion observer for estimating the unmeasured membrane potential of neuronal populations. Australian Control Conference, AUCC 2011, Nov 2011, Melbourne, Australia. pp.CDROM. ⟨hal-00637982⟩