Diagnostics of epileptic activity is usually based on visual and/or automated analysis of brain electromagnetic signals, including EEGs, MEGs, local field potentials and single/multi neuron unit recordings. In primary generalized epilepsies the pathological activity is detected in both hemispheres. However, in some animal models this was not tested properly. Recently, we have found that in pentylenetetrazol rat model of spike-wave discharges (SWDs) in half of subjects a significant part of SWDs was expressed in the cortex of only one hemisphere. The purpose of this paper is to test whether such discharges differ from background activity. In this paper we use both indirected (phase coherence index, mutual information function) and directed (nonlinear adapted Granger causality) nonlinear measures which became popular in neuroscience last decades. Studying connectivity changes accompanying discharge initiation we found increase in interhemispheric coupling similar to that found previously between different cortical regions of the same hemisphere, with no specific preictal connectivity dynamics distinguishable. We also showed that there was no significant difference in interhemispheric connectivity between unilaterally and bilaterally expressed discharges. In pentylenetetrazol rat model of SWDs there is a significant increase in connectivity between hemispheres of somatosensoty cortex during ictal stage. This increase is expressed at the same level for both unilaterally and bilaterally visible discharges.