Clinical studies show that children, especially neonates are in a much higher risk than adults to develop seizures. Such seizures in the brain may be provoked by different factors: tumor, infection, anoxia, fever, trauma, cysts, vascular malformations... Seizures in neonates are also often resistant to treatments and available antiepileptic drugs (AEDs) are inefficient or even provoke and aggravate neonatal seizures. A fundamental concept in epilepsy is that the seizures generated in epileptogenic regions propagate to the other brain structures even to the contralateral side and may develop permanent epileptic focus in the naïve brain structures – secondary epileptic focus. Consequently, it is necessary to treat the neonatal seizures. Diazepam (DZP) and phenobarbital (PB) are extensively used as first and second line drugs to treat acute seizures in neonates and their actions are thought to be mediated by increasing the actions of GABAergic signals. Yet, their efficacy is different and variable with occasional failure or even aggravation of recurrent seizures questioning whether other mechanisms are not involved in their actions. We studied these issues in the intact interconnected hippocampal preparation from neonatal rats and mice. Using this preparation and three-compartment chamber we induced seizures in one hippocampus that propagated to the contralateral one. The propagation of recurrent seizures transformed the contralateral hippocampus into independent epileptogenic focus – mirror focus (MF) - that was capable of generating spontaneous seizures (« seizure beget seizure »). The formation of MF is associated with a permanent increase of the intracellular concentration of chloride and a shift of the actions of GABA from inhibitory to excitatory. Therefore determining how secondary epileptogenesis is induced will have major clinical impact as it will enable to develop tools that prevent selectively the pathogenic seizures.At first, we have determined the impact and the contribution of chloride co-transporter NKCC1 in seizure generation and secondary epileptigenesis. We have shown that the pharmacologically or genetically blockade of NKCC1 did not prevent neither the generation nor propagation of evoked seizures nor formation of MF. However, in the isolated MF, bumetanide effectively blocked spontaneous epileptiform activity. Bumetanide partially reduced DFGABA and therefore the excitatory action of GABA in epileptic neurons. Therefore, bumetanide is a potent anticonvulsive agent although it cannot prevent formation of the epileptogenic MF.Second using different electrophysiological and immunochemistry approaches we have demonstrated that the accumulation of chloride and the excitatory actions of GABA in mirror foci neurons are mediated by NKCC1 chloride importer and by a downregulation and internalisation of the chloride exporter KCC2.Finally using our MF model we have compared the actions of PB and DZP on neonatal seizures. We have revealed that PB but not DZP dramatically reduced initial propagating seizures and prevented formation of epileptogenic MF. We show that PB in contrast to DZP has a highly specific action on AMPA/kainate receptor mediated currents. This action underlies an important difference between the two AEDs as in contrast to PB, DZP aggravates early seizures reflecting the advantage of PB over DZP to prevent secondary epileptogenesis. Yet, after repeated seizures, once an epileptogenic MF has been formed, this difference is abolished because of the strong excitatory actions of GABA. Therefore, the history of seizures prior to GABA acting AED treatment determines its effects and rapid treatment of severe potentially epileptogenic neonatal seizures is recommended to prevent secondary epileptogenesis associated with KCC2 down regulation.