Synchrony Measurement and Connectivity Estimation of Parallel Spike Trains from in vitro Neuronal Networks

Abstract

The goal of this doctoral thesis is to identify appropriate methods for the estimation of connectivity and for measuring synchrony between spike trains from in vitro neuronal networks. Special focus is set on the parameter optimization, the suitability for massively parallel spike trains, and the consideration of the characteristics of real recordings. Two new methods were developed in the course of the optimization which outperformed other methods from the literature. The first method “Total spiking probability edges” (TSPE) estimates the effective connectivity of two spike trains, based on the cross-correlation and a subsequent analysis of the cross-correlogram. In addition to the estimation of the synaptic weight, a distinction between excitatory and inhibitory connections is possible. Compared to other methods, simulated neuronal networks could be estimated with higher accuracy, while being suitable for the analysis of massively parallel spike trains. The second method “Spike-contrast” measures the synchrony of parallel spike trains with the advantage of automatically optimizing its time scale to the data. In contrast to other methods, which also adapt to the characteristics of the data, Spike-contrast is more robust to erroneous spike trains and significantly faster for large amounts of parallel spike trains. Moreover, a synchrony curve as a function of the time scale is generated by Spike-contrast. This optimization curve is a novel feature for the analysis of parallel spike trains.