Our objective is to analyze EEG signals recorded with depth electrodes during seizures in patients with drug-resistant epilepsy. Usually, different phases are observed during the seizure process, including a fast onset activity (FOA). We aim to determine how cerebral structures get involved during this FOA, in particular whether some structures can "drive" some other structures. We compare a transfer entropy based measure with a measure related to linear Granger causality index to detect causal interdependences in multivariate signals generated either by a linear autoregressive model or by a physiology-based model of coupled neuronal populations. Experimental simulation results support the relevance of the new measure for characterizing the information flow for direct and indirect relations.