Maturation of neural circuits requires activity-dependent processes that underpin the emergence of appropriate behavior in the adult. It has been proposed that disruption of these events, during specific critical periods when they exert maximal influence, may lead to neurodevelopmental diseases, including epilepsy [1-3]. However, complexity of neurocircuitry, coupled with the lack of information on network formation in mammals, makes it difficult to directly investigate this hypothesis. Alternative models, including the fruit fly Drosophila melanogaster, show remarkable similarities between experimental seizure-like activity and clinical phenotypes [4-6]. In particular, a group of flies, termed bang-sensitive (bs) mutants have been extensively used to investigate the pathophysiological mechanisms underlying seizure [7-12]. Seizure phenotype can be measured in larval stages using an electroshock assay, and this behavior in bs mutants is dramatically reduced following ingestion of typical anti-epileptic drugs (AEDs; ). In this study we describe a critical period of embryonic development in Drosophila during which manipulation of neural activity is sufficient to significantly influence seizure behavior at postembryonic stages. We show that inhibition of elevated activity, characteristic of bs seizure models, during the critical period is sufficient to suppress seizure. By contrast, increasing neuronal excitation during the same period in wild-type (WT) is sufficient to permanently induce a seizure behavior. Further, we show that induction of seizure in WT correlates with functional alteration of motoneuron inputs that is a characteristic of bs mutants. Induction of seizure is rescued by prior administration of AEDs, opening a new perspective for early drug intervention in the treatment of genetic epilepsy.