CommentaryHypoxia-induced seizures are among the commonest causes of neonatal seizures and raise significantly the risk for subsequent epilepsy and poor neurodevelopmental outcomes (1). Experimental models of early life encephalopathies with seizures are indispensable in the effort to develop effective disease-modifying treatments. The research group authoring this report has pioneered the characterization of a rat model of neonatal hypoxia-induced seizures. Postnatal day 10 (PN10) rats exposed to hypoxia develop acute seizures, subsequent cognitive deficits, increased seizure susceptibility, and spontaneous seizures later in life (2).Prior studies have highlighted the role of AMPA receptors in the pathogenesis of cognitive dysfunction and proepileptogenic diathesis of rats subjected to neonatal hypoxic seizures (3,4). Starting within the first hour following PN10 hypoxic seizures, AMPA receptor spontaneous excitatory postsynaptic currents (sEPSCs) increase in the hippocampus, partially as a result of the enhanced phosphorylation of AMPA receptor subunits, GluA1 and GluR2. This observation correlates nicely with the parallel, but transient, early enhancement of long-term potentiation (LTP) in this model. Furthermore, early systemic administration of AMPA receptor inhibitors, such as NBQX or topiramate, during the two first posthypoxia days prevents the appearance of cognitive impairment and seizure susceptibility following hypoxia. Even though these results provided a strong case for the pathogenic role of AMPA receptors in the early stages of this model, little was known about the molecular and pathophysiological changes during the period of cognitive impairment.Zhou et al. report that 2 to 3 days following hypoxic seizures in PN10 pups, the frequency of sEPSCs is still increased in the CA1 stratum radiatum, through further enrichment of the NMDA receptor (NR)-positive synapses with AMPA receptors, such as GluA1. Through an elegant series of experiments, they demonstrate an increased ratio of AMPA receptor/NR-type sEPSCs, and increased synaptic colocalization of GluA1 and NR1, the obligate NR subunit. It is known that the presence of functional AMPA receptors in NR-containing synapses "unsilences" the synapses in the face of glutamate release, through activation of AMPA receptors, depolarization, and release of the Mg 2+ blockade of NRs (5). Indeed, the authors proceed to show that 2 to 3 days following PN10 hypoxic seizures, the failure rates of eliciting glutamatergic sEPSCs after synaptic stimulation are selectively lower for AMPA receptor-mediated sEPSCs but not for NR-mediated sEPSCs. As the authors comment, such results indicate a reduced number of "silent" glutamatergic synapses following post-hypoxic seizures. Given the importance of silent synapses for the generation of LTP, the authors demonstrate that LTP is significantly impaired in the hippocampus at the time when silent glutamatergic synapses are reduced. Exposure to hypoxia, in the absence of seizures, fails to elicit similar effects, indicating ...