Neuromodulation of the agonistic behavior in two species of weakly electric fish that display different types of aggression

Agonistic behaviors that regulate conflict situations between conspecifics have shaped sociality across evolution (Lorenz, 1963; King, 1973). Conflicts arise because animals compete for different resources (territory, food, mates, breeding sites, etc.) and they are solved when one individual keeps the resource in struggle (dominant) and the other loses it (subordinate). Though the resources individuals compete for may be different and the behavioral patterns involved may be extremely diverse, agonistic encounters always follow the same phases (evaluation, contest and post-conflict resolution). Conflicts may be solved without escalated contests, but animals usually have to fight to elucidate which will become the winner or the loser. Therefore, aggression, defined as an overt behavior that leads to displacing, dominating or harming another individual (Nelson, 2006), is part of the contest phase of agonistic encounters. Given the diversity of aggressive behaviors and the variety of contexts in which aggression is expressed, it has not been possible to reach a general consensus in the definition of aggression and its classification, though several authors have provided useful theoretical frameworks in this concern (Moyer, 1968; Brain, 1979; Wingfield et al., 2006). Within the agonistic context, individuals can aggressively compete either for space or for a mate, or aggressively establish hierarchies in a flock. Thus, depending on the resources individuals compete for and the motivation towards combat, at least three distinctive types of aggression can be observed during agonistic encounters: territorial aggression, reproduction-related aggression and rank-related aggression (Wingfield et al., 2006). The type or types of aggression displayed by one species are correlated with its social structure. For example, it is not likely for gregarious species to display exclusive territorial aggression, nor for solitary species to display rank-related aggression. However, it is not always easy to discern between types of aggression, and to obtain model systems to test clear-cut examples of a distinctive type of aggression. Besides, both in natural performing animals and in laboratory settings, different types of aggression often intermingle, resulting, for example, in territorial aggression only observed between males during breeding while they are competing for mates, or rank-related aggression established by territory defense (Brown, 1964).

Several brain areas are known to control aggression, though it has not been possible to dissect a fixed neural circuit responsible for aggressive behavior. Rather, following Newman's hypothesis (Newman, 1999), we interpret aggressive behaviors as emergent properties of a complex social behavior network (SBN) that includes the medial preoptic area, lateral septum, anterior hypothalamus, ventromedial hypothalamus, periaqueductal gray, medial amygdala and bed nucleus of the stria terminalis (Newman, 1999; Nelson and Trainor, 2007). From this perspective, the distributed pattern of neural activity among the interconnected nodes of the SBN controls social behavior and the diversity in social behavior would be achieved by plasticity in the weighting of activity across the network (Newman, 1999; Goodson and Kabelik, 2009). These neural circuits, initially described in mammals, appear to be highly conserved, and putative homologs have been identified for all classes of vertebrates (Goodson, 2005; O'Connell and Hofmann, 2011). Also conserved among vertebrates are the modulators that shape the activity of the SBN. Several signaling molecules, including hormones (steroid hormones, hypothalamic neuropeptides) and neurotransmitters (serotonin, dopamine, GABA), have been linked to aggression (Goodson and Bass, 2001; Miczek et al., 2002; Summers and Winberg, 2006; Nelson and Trainor, 2007). The SBN hypothesis predicts that each behavior, context and/or phenotype will arise from distinct spatio-temporal patterns of activation of the SBN, product of a precise orchestration of neuroanatomical pathways and neuroendocrine messengers.

Among neuromodulators of aggression, serotonin (5-hydroxytryptamine, or 5-HT) exerts the principal inhibitory control upon aggression (Nelson and Chiavegatto, 2001; Trainor et al., 2009). In several species in which the serotonergic system has been analyzed during agonistic interactions, serotonergic activity increases consistently in subordinates (Yodyingyuad et al., 1985; Blanchard et al., 1991; Summers and Greenberg, 1995; Overli et al., 1999; Summers et al., 2005b) and decreases or remains unchanged in dominants (van Erp and Miczek, 2000; Ferrari et al., 2003; Summers et al., 2003). The activation of 5-HT_1A receptors has been involved in the inhibitory effect of 5-HT on aggression in all vertebrate classes (Deckel and Fuqua, 1998; Sperry et al., 2003; Miczek and Fish, 2006; Clotfelter et al., 2007; Allee et al., 2008; Ten Eyck, 2008). In addition, the 5-HT deficiency hypothesis was formulated based on the correlation between low levels of the 5-HT metabolite 5-hydroxyindoleacetic acid and aggressive or violent behavior in humans and other primates (Krakowski, 2003). Although it is generally accepted that high levels of 5-HT inhibit aggression, there have been no attempts so far to determine whether the serotonergic system is differentially involved in the regulation of the different types of aggression displayed during agonistic interactions.

The hypothalamic nonapeptide arginine vasotocin (AVT) and its mammalian homolog, vasopressin, are key integrators underlying interspecific, sexual, individual and social context differences in behavior in vertebrates (Insel and Young, 2000; Goodson and Bass, 2001; Goodson, 2008; Goodson et al., 2012). Consistent patterns of activation of the vasopressin and AVT systems have emerged in mammals and birds, respectively. Differential central vasopressin signaling parallels differences in the social organization of closely related species of voles (Young et al., 1997). In estrildid finches, the AVT system encodes the valence of social stimuli and group-size preferences between species independently of other aspects of social organization (Goodson and Kingsbury, 2011). Vasotocin modulation of aggression has been demonstrated in several vertebrate classes (Stribley and Carter, 1999; Semsar et al., 2001; Lema and Nevitt, 2004b; Semsar and Godwin, 2004; Larson et al., 2006; Santangelo and Bass, 2006; Greenwood et al., 2008; Goodson and Kabelik, 2009; Filby et al., 2010). In birds, it has been reported that the AVT modulation of aggression depends on the type of aggression displayed (Goodson and Kabelik, 2009; Kabelik et al., 2009). In teleosts, though numerous studies have addressed the influence of the AVT system on agonistic behavior, no consistent patterns across species have yet emerged, probably due (at least in part) to the extremely diverse approaches and procedures used among species (Godwin and Thompson, 2012).

Teleost fishes provide valuable models to explore neuroendocrine bases of behavior as they are the most taxonomically diverse infraclass of vertebrates, they inhabit a wide range of environments, and they display an extraordinary diversity of social systems, features that offer almost endless possibilities of experimental models. However, these advantages of fish models may raise difficulties in the interpretation of data as it is difficult to identify generalities or conserved strategies across teleosts, which are an older and more diverse group than either birds or mammals.

We present a comparative analysis of the neuromodulation of aggression between two species of gymnotiform electric fish with different social structure that display different types of aggression. As a particular case of the theoretical prediction raised by Newman (Newman, 1999), we hypothesize that each type of aggression will have a distinctive spatio-temporal pattern of activation across the network. To approach this issue, we selected two clear-cut examples of different types of aggression: the non-breeding intrasexual and intersexual territorial aggression of Gymnotus omarorum (Batista et al., 2012) and the intermale reproduction-related aggression of Brachyhypopomus gauderio (Zubizarreta et al., 2012). We then explored the emergence of distinctive patterns by analyzing the modulation of the serotonin and AVT systems upon these two types of aggression.

Gymnotus omarorum (Richer-de-Forges et al., 2009) and B. gauderio (Giora and Malabarba, 2009) cohabit Uruguayan freshwater habitats at the southern boundary of the gymnotiform continental distribution (30–35°S) (Silva et al., 2003). As shown in Fig. 1, the solitary G. omarorum emits sexually monomorphic triphasic electric pulses at around 25 Hz at 20°C (Caputi, 1999; Richer-de-Forges et al., 2009), whereas the gregarious B. gauderio emits sexually dimorphic biphasic electric pulses at around 15 Hz at 20°C (Hopkins et al., 1990; Caputi et al., 1998). In both species, a medullary pacemaker nucleus (PN) controls the timing of the electric organ discharge (EOD) (Bennett et al., 1967). The spatial organization and innervation pattern of the peripheral electric organ are responsible for the species-specific EOD waveform (Trujillo-Cenóz et al., 1984; Stoddard, 2002). Brachyhypopomus gauderio, but not G. omarorum, exhibits a sustained nocturnal increase of EOD rate (Fig. 1), along with a circadian rhythmicity of EOD waveform, both of which vary seasonally and are dependent on social context (Franchina and Stoddard, 1998; Franchina et al., 2001; Silva et al., 2007; Gavassa et al., 2013).

Though male and female G. omarorum display aggression all year round in intrasexual and intersexual interactions, we focused on the non-breeding dyadic agonistic encounters as a remarkable example of non-sex-biased territorial aggression (Batista et al., 2012). Several considerations must be taken into account to reliably diagnose this distinctive type of aggression and to exclude other possible interfering types. During the non-breeding season, resource value is not expected to depend on individuals' reproductive status and should thus be equal for males and females in this sexually monomorphic species in which interaction costs can also be considered to be symmetric between sexes. We designed controlled behavioral experiments to ensure that territory is the only resource animals fight for, by placing two fish (with the same previous experience and residence time) in equal-sized plain compartments separated by a removable glass gate. When the gate was removed, all fish (from either female–female, male–male or male–female dyads) defended this space vigorously and a clear dominance was established (Fig. 2). Theoretical asymmetric contest models (Maynard Smith and Parker, 1976; Parker and Rubenstein, 1981) predict that when sex asymmetries in resource holding potential (RHP) or in the resource value itself are not expected between contenders, the contest outcome should be sex-independent. However, until now, this prediction had not been demonstrated, mainly because in monomorphic species the value of the resource often differs between male and female contestants (Dunham, 2008). Gymnotus omarorum fits perfectly in this model, as dominance is not influenced by individuals' sex, and only weight asymmetry, the most direct indicator of RHP, influenced the contest outcome (Batista et al., 2012).

The reproduction-related intermale aggression displayed by B. gauderio (Zubizarreta et al., 2012) is probably part of the natural behavioral repertoire of this gregarious species, which has a polygynous breeding system in which males hold consistent home ranges typical of exploded lek polygyny (Miranda et al., 2008). Territory itself is not a sufficient incentive to elicit aggression in this species, but long-term residents aggressively defend their courting sites against intruders, and become dominants if size asymmetry favors them (Fig. 2).

As shown in Fig. 2, both species followed the three standard phases of agonistic encounters (evaluation, contest and post resolution) but with different timing: G. omarorum resolved contests in less than 3 min, whereas B. gauderio took around 15 min to establish dominance. Within the contest phase, both species interacted aggressively, though displays of overt aggressive behaviors were species-specific. For example, only G. omarorum displayed highly aggressive jaw locks and bites directed to the opponent's abdomen, whereas B. gauderio showed milder displays such as mutual attempts to bite the opponent's tail in an antiparallel position. Gymnotus omarorum produced submissive electric signals (offs and chirps) that showed a consistent temporal pattern (Fig. 2) (Batista et al., 2012). During the contest phase, only offs (electric silence) were produced as an early sign of submission. After contest resolution, the subordinate continued to produce more and longer-lasting offs, and added chirps, probably as a more unambiguous submissive signal (Fig. 2). The clear submissive information encoded by chirps in G. omarorum contrasts with previous findings in Apteronotus leptorhynchus (Zupanc et al., 2006; Triefenbach and Zakon, 2008; Hupé and Lewis, 2008), in which chirps were interpreted as threat signals associated to aggression. Brachyhypopomus gauderio produced chirps (either dominants or subordinates) and offs (only subordinates) but they did not seem to follow a predictable temporal pattern (Zubizarreta et al., 2012).

As in other gymnotiforms (Hopkins, 1974; Hagedorn and Heiligenberg, 1985; Zakon et al., 1991; Fugère et al., 2011), both species establish ranks of EOD rate after contest resolution, in which dominants display higher discharge rates than subordinates (Perrone, 2012). Interestingly, we see interspecific differences in the way this post-resolution EOD rate rank is attained (Fig. 2). In the reproduction-related aggression of B. gauderio, the dominant fish increases its EOD rate during the contest to signal its dominance status. In contrast, in the non-breeding territorial aggression of G. omarorum, the subordinate fish decreases its EOD rate, adding a persistent signal of electric submission. We assume fish find advantage in advertising either dominance or subordination status via electric cues, thus avoiding extra time-consuming contests and injuries derived from fights. Each species establishes its post-resolution EOD rate rank in coherence with its social structure. In exploded lek systems (B. gauderio), it may be crucial for males to signal dominance to discourage other neighbor males to invade their home ranges, whereas in highly aggressive solitary species in which individuals rarely interact (G. omarorum), it may be more important to signal subordination unambiguously.

In summary, two clear-cut types of aggression displayed by two species with different social structure can be reliably tested in laboratory settings using distinct experimental protocols that simulate the natural conditions in which agonistic behaviors arise in each species. This approach, which allows a precise characterization of the agonistic encounter in each species (Fig. 2), also opens the possibility for a comparative evaluation of its modulation.

Though indisputable and ubiquitous across evolution, the 5-HT inhibitory control of aggression is now conceived to be much more complex than expected, as it appears to be dependent on brain region, time and context (Summers et al., 2005b). For example, the differential involvement of 5-HT in different behavioral phenotypes within a same species has been reported in lizards and mice (Summers et al., 2005b; Caramaschi et al., 2008; de Boer et al., 2009). However, no attempts to identify distinct patterns of activation of the serotonergic system among different types of aggression have been reported so far. Only one previous study in macaques compared the activity of the serotonergic system between species with different social structure (Westergaard et al., 1999).

Neurochemical data support the existence of a differential pattern of activation of the serotonergic system between the non-breeding territorial aggression of G. omarorum and the reproduction-related aggression of B. gauderio. In B. gauderio, as expected, the telencephalic serotonergic activity increases in subordinate males following a contest; however, in G. omarorum, 5-HT levels exhibited by both subordinates and dominants after territorial aggression were significantly lower than those exhibited by controls (Zubizarreta et al., 2012). Taken together, the conventional 5-HT profile found in B. gauderio and the unexpected 5-HT profile found in G. omarorum constitute the first evidence of the differential activation of the serotonergic system between types of aggression.

As the involvement of 5-HT_1A receptors in the serotonergic inhibition of aggression is well known across vertebrates (Nelson and Trainor, 2007), we tested the effects of the administration of the 5-HT_1A agonist 8-hydroxy-2-(di-n-propylamino)tetralin (8-OH-DPAT) to potential dominants prior to the contest. As expected, 8-OH-DPAT decreased aggression in G. omarorum (Fig. 3). This treatment reversed the expected contest outcome in seven of 12 dyads, a phenomenon that in other vertebrates has been associated with 5-HT changes (Summers et al., 2005a). These results confirm that the inhibition of aggression by pharmacological manipulation of 5-HT levels shown in a previous report in G. omarorum (Capurro et al., 1997) is mediated at least partially by the activation of 5-HT_1A receptors. However, the resident male B. gauderio failed to show consistent effects on any aspect of the agonistic encounter after the administration of 8-OH-DPAT. Pharmacological stimulation of the 5HT_1A receptors in B. gauderio did not affect contest outcome (seven out of eight injected residents won the contest), aggression levels or motivation towards combat (Fig. 3). Previous reports in B. gauderio showed that 8-OH-DPAT diminishes amplitude and duration of the EOD waveform (Allee et al., 2008), which is an important proxy of male body size and endocrine status (Gavassa et al., 2011). The lack of effect of the 5-HT_1A agonist in B. gauderio is relevant because this is the first species in which the administration of 5-HT_1A receptor agonists does not clearly diminish aggression in realistic dyadic encounters. These results demonstrate interspecific differences in the organization of the 5-HT system in two species that display different types of aggression.

In summary, G. omarorum is characterized by a low basal serotonergic activity and, after agonistic encounters, shows a slow recovery of 5-HT levels. This is consistent with the clear anti-aggressive effect of the 5-HT_1A agonist obtained with the pharmacological manipulation in this species. The reproduction-related aggression of B. gauderio shows a dynamic response of the serotonergic system, in which subordinates exhibit the expected 5-HT levels higher than controls. Brachyhypopomus gauderio shows a high basal serotonergic activity, which may explain the non-responsiveness of reproduction-related aggression after 5-HT_1A agonist administration.

Three populations of AVT neurons in fish brains are distributed along the third ventricle in the ventral portion of the preoptic area in the anterior hypothalamus (Lema and Nevitt, 2004a; Dewan et al., 2008; Thompson and Walton, 2009). The parvocellular AVT neuron population (homologous to the supraoptic nucleus in tetrapods) is located at the rostral and ventral portion of the preoptic area; magnocellular and gigantocellular AVT neuron groups (homologous to the paraventricular nucleus in tetrapods) are found more caudally and dorsally (Kapsimali et al., 2001; Urano and Ando, 2011; Godwin and Thompson, 2012). The projections of these neurons reach the neurohypophysis, where AVT is released into the bloodstream to exert peripheral effects, and several widespread brain areas (from telencephalon to hindbrain) where AVT is postulated to modulate social behavior (Goodson and Bass, 2001; Saito et al., 2004). As recently reviewed, several studies have demonstrated the involvement of the AVT system in the modulation of the aggressive behavior in different species of teleosts (Godwin and Thompson, 2012). Given the heterogeneity in both the social structure of fish species studied and the experimental approaches used, it has been difficult to identify general strategies in the AVT control of aggression in teleosts so far. However, based on pharmacological and neuroanatomical data, it has been postulated that two subsystems of AVT neurons alternatively activate and control different aspects of social behavior (Greenwood et al., 2008; Godwin and Thompson, 2012). The magno-gigantocellular AVT populations produce and release AVT in circuits that stimulate courtship and/or aggressive behaviors, whereas parvocellular neurons most likely release AVT in circuits that induce behaviors related to social subordination, particularly social withdrawal. According to this model, different types of aggression, and the responses dominants and subordinates display, would depend on both the relative development of these different two subsystems across species and phenotypes, and their relative activation in particular social contexts. For example, greater levels of AVT expression were found in territorial males of Astatotilapia burtoni than in non-territorial males (Greenwood et al., 2008). Territorial butterflyfish (Chaetodon multicinctus) show larger magnocellular and gigantocellular AVT neurons and greater density of AVT projections than the non-territorial species Chaetodon miliaris (Dewan et al., 2008). In Cichlasoma dimerus, non-territorial males show larger parvocellular AVT neurons than the territorial males (Ramallo et al., 2012). In the peacock blenny (Salaria pavo), territorial males present smaller parvocellular neurons than non-territorial ones (Grober et al., 2002). In zebrafish (Danio rerio), dominants express AVT in magnocellular neurons whereas subordinates express AVT in parvocellular preoptic neurons (Larson et al., 2006).

Vasotocin is probably acting directly by influencing the activity of SBN nodes and descending motor output pathways, and indirectly by changing peripheral body states that feed back to the brain, but it has been challenging to determine the precise target areas of AVT effects in teleosts. As the pattern of emission of the electromotor output in electric fish is itself a social display [as it signals species, sex, season, status, male quality and social context (Silva et al., 2007; Perrone et al., 2009; Gavassa et al., 2012)], which relies on the activity of a well-known and accessible medullary medial nucleus (the PN) (Pouso et al., 2010; Quintana et al., 2011a; Quintana et al., 2011b), electric fish appear as a good model system to target specific brain areas of AVT actions. In a previous study, Perrone et al. (Perrone et al., 2010) demonstrated that AVT induced a long-lasting and progressive increase of EOD rate in isolated B. gauderio, but had no effects in G. omarorum. The administration of AVT in an in vitro preparation of the isolated PN of B. gauderio also induced a long-lasting increase in its firing rate, strongly suggesting that the PN is a direct target of AVT projections (Perrone et al., 2010).

Gymnotus omarorum and B. gauderio differed in the effects of the administration of AVT to potential dominants prior to the contest. In the non-breeding territorial aggression of G. omarorum, AVT injection increased the motivation towards aggression revealed by a decrease in the first attack latency and a marginal decrease in contest duration, but no changes in aggression levels (Fig. 4) (Perrone, 2012). Vasotocin injection in the resident male had no consistent effects on aggression levels or motivation towards combat of B. gauderio (Fig. 4) (Perrone, 2012). More interestingly, the electric submission and dominance observed in G. omarorum and B. gauderio, respectively, were both AVT-dependent in a distinctive way (Fig. 5) (Perrone, 2012). In the territorial aggression displayed by G. omarorum, AVT injected into the predicted subordinate prevented the electric submission observed after conflict resolution, suggesting that the sustained decrease in EOD rate observed in intact subordinates is achieved by the inhibition of an AVT input to the PN. In the reproduction-related aggression, the AVT receptor V1a antagonist Manning compound administered to the predicted dominant male B. gauderio prevented the electric dominance display, allowing us to hypothesize that the sustained increase in EOD rate observed in intact dominants is caused by direct stimulation of the PN via AVT fibers.

As observed in teleost-related species with different social structure (Dewan et al., 2008) and among dominants and subordinates within the same species (Greenwood et al., 2008; Filby et al., 2010; Ramallo et al., 2012), there is a differential pattern of activation of the AVT system between the non-breeding territorial aggression of G. omarorum and the reproduction-related aggression of B. gauderio. Our results match coherently with the above-described hypothesis that AVT's control of social behavior in teleosts relies on the balanced activity of the magno-gigantocellular and parvocellular subsystems of AVT neurons (Greenwood et al., 2008; Godwin and Thompson, 2012). Assuming that AVT fibers reach the PN in both species (an assumption that will be investigated shortly by immunohistochemical techniques), an excitatory AVT-ergic input, probably from the magno-gigantocellular subsystem, can exert a tonic control of EOD rate that is expected to be stronger in G. omarorum than in B. gauderio (Fig. 6). As the administration of AVT to resting isolated G. omarorum (Perrone et al., 2010) or putative dominants (Perrone, 2012) does not affect EOD rate, we assume there is a permanent input of AVT to the PN in this species that occupies all the binding sites of AVT receptors (saturated). In contrast, in B. gauderio the non-saturated AVT input to the PN may be further activated in dominants after conflicts and thus explains their electric dominance (Fig. 5). In subordinates, the prevalence of the activation of the parvocellular subsystem of AVT neurons may result in the inhibition of the AVT input to the PN, a hypothesis confirmed in G. omarorum in which the injection of AVT prevented the establishment of their electric submission behavior (Fig. 5). Another interesting difference in the activation of the AVT system among dominants of both species is that only in G. omarorum (but not in B. gauderio) does activation of the magno-gigantocellular subsystem of AVT neurons appear to enhance overt aggression (Fig. 6). This global picture of the distinct activation patterns of the AVT system between species may be part of a more general strategy; i.e. our results indicate the existence of specific spatio-temporal patterns of activation of the SBN that may characterize different types of aggression in concert with the social structure of each species.

Multiple neuromodulators, acting via both fast wired circuits and slow global diffusive ways, shape the spatio-temporal pattern of activity of the network controlling social behavior, thus providing the emergence of environmental-, ontogenic-, social-context- and phenotype-dependent behaviors (Newman, 1999; O'Connell and Hofmann, 2011; Goodson et al., 2012). Our study provides an interesting contribution to assess the complexity of the control of social behavior showing distinctive patterns of activation of both the 5-HT and the AVT systems in the modulation of different types of aggression in species with different social structure. We have demonstrated that these different patterns of activation arise not only between dominants and subordinates of each species but also among dominants, depending on the type of aggression they have displayed (Fig. 6). For example, the activation of the serotonergic system via 5-HT_1A receptors inhibits overt aggression of dominants in the territorial aggression but not in the reproduction-related aggression. Conversely, AVT induces electric dominance in the reproduction-related aggression but not in the territorial aggression. In addition, we have identified different target areas in which these modulatory actions are taking place (Fig. 6). Besides the effects of 5-HT and AVT on SBN nodes that control the locomotor pathways involved in overt aggression, AVT is also acting on the PN (shaping either electric dominance or submission).

In mammals, 5-HT and vasopressin are known to act as functionally opposed neurotransmitter systems in the control of aggressive behavior, and we have considerable evidence of direct and indirect connections between neurons of both systems (reviewed in Ferris, 2006; Trainor et al., 2009). Though the interaction between 5-HT and AVT has not been extensively analyzed in teleosts, raphe neurons expressing AVT receptors have been recently described (Kline et al., 2011). Following the same trend reported in mammals, we observed opposing actions of 5-HT and AVT upon overt aggression in the territorial aggression of G. omarorum, probably acting by enhancing and decreasing the activity of the same target areas across the SBN (Fig. 6). However, we failed to find opposing effects of both systems on the overt aggression of B. gauderio, as neither 5-HT nor AVT affected dominants' aggression. Interestingly, an asymmetric control of 5-HT and AVT systems was evidenced at the level of the PN (Fig. 6). Our data suggest an important excitatory function for AVT in both species (even stronger in G. omarorum, as discussed above), but provide no evidence that 5-HT modulates EOD rate. Previous studies in both species evaluating the effects of the activation of the 5-HT system on locomotor and electric behaviors did not report 5-HT actions on EOD rate in Brachyhypopomus (Allee et al., 2008), whereas pharmacologically increased levels of 5-HT evoked a decrease in EOD rate in Gymnotus (Capurro et al., 1997).

The advantages of our model species allowed us to combine behavioral, neurochemical and pharmacological techniques to identify precise target areas and mechanisms of the neuromodulation of different types of aggression that may represent more general strategies of the control of social organization across vertebrates.

We thank Laura Quintana for her generous revision and suggestions to our manuscript. We thank Pilar Fernández and Laura Nozar from Loopita Design for their help in the artwork. All research procedures complied with ASAP/ABS Guidelines for the Use of Animals in Research and were approved by the Universidad de la República Institutional Ethical Committee (Comisión Bioética, Instituto Clemente Estable, MEC, 07-28-2008 and 007/02/2010) and by the Institutional Animal Care and Use Committee of Florida International University, Miami, FL (protocol no. 09-023).

List of abbreviations

 
• 5-HT

  5-hydroxytryptamine (serotonin)

 
• 8-OH-DPAT

  8-hydroxy-2-(di-n-propylamino)tetralin

 
• AVT

  arginine vasotocin

 
• EOD

  electric organ discharge

 
• PN

  pacemaker nucleus

 
• RHP

  resource holding potential

 
• SBN

  social behavior network