Nelson Niemeyer: Reconstructing Phase Response Curves in a Small Motor Network

BCCN Berlin and TU Berlin

 

Abstract

Many behaviours, such as walking, swimming or breathing require stable temporal patterns. One powerful way to describe neural networks that generate such patterns uses systems of coupled oscillators. In these models phase response curves (PRCs) describe how one oscillator responds to inputs from another oscillator. PRCs contain the essential properties that determine synchronisation states and explain how output patterns are generated. As of now, determining PRCs experimentally requires full control over the input and involves patch-clamping of isolated units. This process is laborious, time-consuming and takes the neuron outside of its network context. In this project we develop two approaches to extract PRCs from spike times of the intact network alone. Subsequently, these methods are tested on simulated data. Method I is based on least squares optimisation and provides fast solutions that are accurate under low-noise conditions. Method II is based on maximum likelihood estimation and covers slightly larger noise ranges. Finally, both methods are applied to experimental recording data from neurons innervating a Drosophila flight muscle. Neither method provides reconstructions that capture the salient features of this network. We discuss potential reasons for this lack of success and suggest future improvements to both methods.

Organized by

Susanne Schreiber / Robert Martin

Location

BCCN Berlin, lecture hall, Philippstr. 13 Haus 6, 10115 Berlin