Extracellular neuronal cell cultures provide insights in the neurophysiological information processing. For this reason multi-electrode arrays (MEAs) are used to acquire neuronal activities, the growth process of cells and for measuring the reactions of neurons to stimulation. Thus, MEAs are required for the recording of neuronal network activity. The objective of this thesis is to analyze the neuronal activities regarding their causal correlations. For this purpose the used procedure is the parallel factor analysis 2 (PARAFAC2) which is a decomposition model for multidimensional data. To evaluate this method artificial datasets composed of neuronal activity are generated. PARAFAC2 will be analyzed in regard to the tracking of spatial moving sources and its noise suppressing characteristics. In addition a procedure is presented that allows identifying connections between the recorded channels by the time signature of the PARAFAC2 components. These results are compared with the conditional Granger causality. In this context the conditional granger causality is also evaluated on simulated neuronal activities.