An optimization method to simultaneously estimate electrophysiology and connectivity in a model central pattern generator

Abstract

Central pattern generators (CPGs) appear to have evolved multiple times throughout the animal kingdom, indicating that their design imparts a significant evolutionary advantage. Insight into how this design is achieved is hindered by the difficulty inherent in examining relationships among electrophysiological properties of the constituent cells of a CPG and their functional connectivity. That is: experimentally it is challenging to estimate the values of more than two or three of these properties simultaneously. We employ a method of statistical data assimilation (D.A.) to estimate the synaptic weights, synaptic reversal potentials, and maximum conductances of ion channels of the constituent neurons in a multi-modal network model. We then use these estimates to predict the functional mode of activity that the network is expressing. The measurements used are the membrane voltage time series of all neurons in the circuit. We find that these measurements provide sufficient information to yield accurate predictions of the network's associated electrical activity. This experiment can apply directly in a real laboratory using intracellular recordings from a known biological CPG whose structural mapping is known, and which can be completely isolated from the animal. The simulated results in this paper suggest that D.A. might provide a tool for simultaneously estimating tens to hundreds of CPG properties, thereby offering the opportunity to seek possible systematic relationships among these properties and the emergent electrical activity.