Status epilepticus (SE), an acute condition characterized by repetitive or prolonged seizures, affects between 120,000 to 200,000 people per year in the United States (Neill et al., 2005; Lowenstein, 2006). It occurs more often in children than in adults, and 40–50% of children under the age of 2 years (Shinnar et al., 1997). The incidence of neonatal seizures varies according to the age, severity of the etiologic factor, and of the population studied. The clear incidence has not yet been established, although it has been estimate that varies between 1.1–8.6/1000 live births (Saliba et al., 1996, 1999; Sheth et al., 1999; Mosley, 2010). Furthermore, the prognosis of neonatal seizures is highly variable (Costa et al., 2001), and persists the controversy whether the severity, duration or frequency of seizures may contribute to brain damage independently of its etiology (Udani, 2008). In general terms, half of the cases exhibit minimum sequels and the other half evolves into death or serious squeals (Costa et al., 2001). In addition, newborns with transient focal ischemia, metabolic abnormalities without a clear etiology seem to evolve in a more satisfactory manner, unlike those with hypoxic-ischemic encephalopathy, infection of the nervous system and prenatal brain dysgenesis that generally have a worse prognosis (Udani, 2008). From a clinical standpoint, most studies fail to show that seizures per se regardless the severity of the etiologic factor is able to disrupt brain development. In this context, animal models allow the observation and analysis of variables under controlled conditions and have been used to explore these issues. Thus, the animal models with neonatal injury that results in permanent impairment, especially of social interaction are considered useful for the study of neurodevelopmental disorders, such as schizophrenia and autism (Schneider and Koch, 2005; Tordjman et al., 2007).

Following these lines of reasoning, animal data indicates that seizures during development are associated with a high probability of long-term adverse effects such as, learning and memory impairment, behavioral changes and epilepsy that are strongly age dependent, as well as the severity, duration, and frequency of seizures (Holmes, 2004, 2009). The mechanisms that underlie these changes are not yet completely understood, but neonatal seizures do not lead to cell loss, rather than considerable synaptic reorganization. Studies focused primarily on behavioral changes, rather than epileptogenic activity, are less frequent. Two studies have shown that neonatal seizures produce anxiety-related behavior (Sayin et al., 2004; Shi et al., 2007), and another study demonstrated deficit of cognitive flexibility (Kleen et al., 2011). Recently, we demonstrated that the neonatal SE (Castelhano et al., 2010) leads to impairment of social interaction in adolescent male rats. Our study brings an important contribution to demonstrate that sequels of neonatal seizures extending beyond the learning and memory deficits, increased anxiety and loss in cognitive flexibility. Animals studies reveal that early life seizures can disrupt some brain structure such as the hippocampus (Holmes et al., 1998; Lynch et al., 2000; Santos et al., 2000; Villeneuve et al., 2000; Sogawa et al., 2001; Cornejo et al., 2007; Nishimura et al., 2011) neocortex (da Silva et al., 2005; Isaeva et al., 2010), prefrontal cortex (PFC) (Kleen et al., 2011) and thalamus (Santos et al., 2000; Kubová et al., 2001) allowing the argument that the behavioral implications are broad, impacting the cognitive, and adaptive performance. In human condition, different neonatal conditions are likely to cause neurological damage e alter brain development. Atladóttir et al. (2012) using cohort study design investigated neonatal conditions and the risk for Autism Spectrum Disorders (ASDs). They found an increased risk of ASD after exposure to perinatal hypoxia, neonatal seizures, intracranial hemorrhage, neonatal hypoglycemia, neonatal septicemia, and meningitis.

Based on accumulated evidence, animal models submitted to neonatal seizures may be relevant for studies of neurodevelopmental disorders. In this context, it is necessary a more comprehensive study on the behavioral repertoire, particularly in the area of sociability, repetitive behaviors, and anxiety. On the balance, the aim of the present study was to investigate the existence of a persistent deficit and if the anxiety-related behavior could be associated with that impairment.

In the present study we used a three—chambers choice test to assess social behavior, elevated plus maze and open field to evaluated general anxiety. Adult animals exposed to neonatal SE showed less preference for social novelty, increased emotionality in an unfamiliar and threatening environment, increased grooming episodes and immobility, reduced locomotor activity in situations of higher risk. Taken as a whole, the data suggest that early life seizures in rats leads to increased emotionality and abnormal social behavior.

The social interaction test developed by File and Hyde (1978) was the first to measure experimental anxiety based on ethologic concept. Posteriorly, Haller and Bakos (2002) proposed a novel animal model of anxiety in a social context (Social Approach-Avoidance test). In this paradigm, prestressed Wistar rats spent significantly less time as compared to control in a social compartment containing a unfamiliar rat confined behind a transparent wall, demonstrating the anxiogenic-like effect of the stress on approach-avoidance behavior (Nicolas and Prinssen, 2006). In the sociability paradigm, the animals exposed to neonatal SE displayed less preference for social novelty than controls. We previously demonstrated that pilocarpine-induced neonatal SE impaired social play behavior (Castelhano et al., 2010). It is well-established that adolescence is an important period for the development of adult social competence and that an inadequate social development, in addition to seizures itself or maladaptive response following seizures, may contribute to social anxiety in adulthood. In rodents, higher levels of social anxiety can disrupt the preference between two stimuli in a preference test (Mines et al., 2010). This effect could be related to decreased motivation, increased emotionality or impaired social memory. Lin et al. (2009) reported in lithium-pilocarpine-induced neonatal seizures reduced levels of dopamine in the PFC. It is well-established that the dopamine found in the PFC originates from the ventral tegmental area, a structure involved in neural network of social behaviors (Goodson and Kabelik, 2009) is also involved in and emotional responses (Baskerville and Douglas, 2010). In this sense, dopamine reduction in PFC could lead to less engagement in social interactions. However, taking in account that in social approach phase experimental group demonstrated being as interested as the control for a social encounter, the probable mechanisms to explain reduced preference for social novelty are the increased state of anxiety associate with the deficit in social discrimination. There is emerging in the literature a discussion about the bi-directional relationship and possibly cyclic, between social interaction impairment and the degree of anxiety or emotionality. Anxiety intensifies social impairment, and poor social function contributes to anxiety (Nicolas and Prinssen, 2006; White et al., 2010). In this sense, animals were evaluated in the open field and the elevated plus-maze. Experimental animals avoided the central area of the open field in both exposures, suggesting no habituation to anxiogenic environment, but they performed similarly to control in total locomotion suggesting no locomotor activity changes. In addition, they displayed higher immobility and stereotyped behavior. In EPM, they spent less time and reduced number of entries on the open arms, suggesting the presence of general anxiety. Reduced locomotor activity on sociability test, even after the habituation period and before social presence, suggest that the new environment increased the state of anxiety and that no habituation to anxiogenic space was reached, as observed in the central zone of the open field. Rodents also have emotional behaviors similar to humans when exposed to specific experimental situations (Ramos and Mormède, 1998). In this regard, Nicolas and Prinssen (2006) evaluated the social approach-avoidance behavior in a high-anxiety strain, F-344 rats. The animals exhibited spontaneous avoidance behavior that was sensitive to benzodiazepine agonist and inverse agonist in a bidirectional manner. To the best of our knowledge, we are the first group to demonstrate abnormal social behavior following early life seizures. However, general anxiety was previously reported. Corroborating with our results, Sayin et al. (2004) demonstrated anxiety-related behavior without locomotor activity changes in adult animals submitted to SE by kainic acid injection during early stage of development. Oliveira et al. (2007) reported anxiety-like behavior in adult rats submitted to lithium-pilocarpine model at 15 days postnatal. The enhanced emotionality could be associated to changes in hippocampus function. Lugo et al. (2012) and Benedikt et al. (2012) demonstrated that hippocampal damage leads to anxiety-like behavior. Beyond of the hippocampus, others brain regions exhibit maladaptive response following neonatal seizures. For example, Kleen et al. (2011) observed that adults male Wistar rats aged between 1 and 10 days postnatal that undergo a series to of neonatal seizures by fluorothyl administration exhibited structural changes of the PFC that were positively correlated with the impairment of cognitive flexibility and insistence to maintain a routine previously learned. Individuals with autism often have repetitive behaviors and strong resistance to the change of routine. Failure to change routine in rodents may be analogous to the cognitive inflexibility. In stressful situations is common the emergence or intensification of stereotypes. In our study, we observed increase in self-grooming when animals were exposed to an anxiogenic environment. The association of grooming behavior with anxiety or stress is well-established in animal models. Stressful events or environments (Langen et al., 2011) as well as appetitive situations can cause an animal to develop abnormal repetitive behavior (Spruijt et al., 1992; Van Erp et al., 1994). Additionally, the enhanced stereotyped behavior could be also due of imbalance both within and between the motor, cognitive and limbic corticostriatal circuits. The corticostriatal circuits include striatum, globus pallidus, substancia nigra, thalamus, and cortex. It is well-established that corticostriatal circuits are functionally divided into three circuits related to the predominant cerebral cortical input to striatum: sensoriomotor circuit functionally related to movements (comprising the motor and oculomotor loops), associative circuit functionally related to cognitive functions (dorsolateral prefrontal loop) and the limbic circuit functionally related emotional-motivational behavior (lateral orbitofrontal and anterior cingulate loops) (for an extensive review see; Langen et al., 2011). Abnormal repetition of behavior can result from damage to any of the corticostriatal circuits. Experimental data demonstrated that SE can damage brain regions involved with corticostriatal circuit. Fernandes et al. (1999) observed hypermetabolism in cortical and forebrain regions plus the substantia nigra following lithium–pilocarpine-induced SE in immature rats. Kubová et al. (2001) demonstrated the presence of silver-positive cells predominantly in the mediodorsal, but also too in the ventrolateral and ventromedial thalamic nucleus in rats that experienced pilocarpine-induced SE at the age of 12 days postnatal. Interestingly, the mediodorsal and ventrolateral nucleus of thalamus are part of limbic and sensoriomotor circuit, respectively. Taking in account that stereotypies can function as coping mechanism to reduce animal's arousal level when it is exposed to stressful events or environments (Langen et al., 2011), it is licit to suppose that animals with damaged thalamus nuclei could be more prone to experience stereotypies under these situations.

As previously mentioned, the decreased preference for social novelty could be also associated to inability of animals exposed to neonatal SE to distinguish a familiar from an unfamiliar conspecifics. In support to this hypothesis, Sankar et al. (1998) reported in rats submitted to pilocarpine-induced seizures during second week of life, damage in the amygdala and hippocampus, brain regions implicated in social recognition memory an emotionality (Ferguson et al., 2002; Yang et al., 2013). Thus, the abnormal social behavior noted in animals exposed to early life seizures could be related to high state of anxiety and impaired social memory.

Taken as a whole, adult animals exposed to neonatal SE displayed less preference for social novelty than control, anxiety-related behavior, increased stereotyped behavior in anxiogenic environment without locomotor activity changes. The data suggest that neonatal SE in rodents leads to altered anxiety-related and abnormal social behaviors.