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.