This study showed that monogamous male and female pearl cichlids have similar brain volumes as well as aggressive and cognitive performances, probably due to their similar roles in parental care. Hormone levels were expected to be different, at least to what androgens were concerned. This difference was observed for 11KT and for the ratio of conversion 11KT/T, which were lower in females compared to males, yet not for cortisol and T levels. Even though males and females were similar in most of the variables tested herein, we found distinct correlations between sexes that deserve attention. While cognitive performance and brain size are highly correlated to cortisol levels in females, males’ social and cognitive behaviors, as well as brain sizes, are strongly related to T. These results suggest that, besides similar pressures provided by monogamy and biparental care, male and female pearl cichlids still have behavioral and physiological particularities coordinated by different hormonal patterns and brain areas.
As a monogamous species, female and male pearl cichlids have similar social and cognitive abilities
Males and females of G. brasiliensis showed similar aggressiveness and inhibitory control. These results corroborate our predictions, since both parents must coordinate their behavior to defend and assist the offspring while chasing away predators 26. This coordination is a highly demanding cognitive exercise for social animals, which is hypothesized to have shaped the brain in monogamous species 16. Moreover, both parents seem to receive similar cognitive demands during parental care synchronization, as expected, since males and females performed very much alike in a test considered to be highly cognitively demanding, such as the inhibitory control task, which is the opposite scenario observed for other species that do not perform biparental brood care (e.g., Nile tilapia with female-only care: 27) or perform no brood care at all (e.g., guppies: 28,29). There were no differences regarding specific brain volumes. Therefore, both sexes from this species seem to be under a similar evolutionary pressure, which concerns sociability, cognition, and brain size.
Cortisol and testosterone are similar in both males and females, although males present higher plasma levels of 11KT
Males showed similar T levels compared to females, and higher 11KT levels and 11KT/T rates, thus in accordance with a teleost fish pattern, in which 11KT is considered the main androgen associated to reproductive behavior in males 5,25. This second androgen is usually found in very low levels or is sometimes absent in females 5. The effect of sexual steroids on the social and reproductive behavior, however, is highly variable among species, since castration can either impair, increase, or not affect at all aggressiveness, depending on the species 12,30. Sapolsky (2017) states that T has a very flexible utility in vertebrates, and could be more related to the maintenance of an individual’s status in a challenging scenario than to the aggressive behavior per se. As proposed by the challenge hypothesis, males’ androgens levels are strongly influenced by the presence of male competitors and females 6,7.
In teleost fishes, females tend to prefer males that have nesting-care habits 32; therefore, androgens in males will be coordinated by the presence of females rather than by competitors’ threats 7. Nevertheless, in our study, males were always allocated into individual aquaria, facing another same-sex neighbor in order to avoid social isolation stress (the same was done for females). Thus, a possible hierarchy could have been achieved through aquaria facing walls during the days of the experiment. However, given that there were no females at sight and males’ (and females’) territories were never invaded or transposed by a competitor, this probably provided the necessary stability for the lack of abrupt changes in T levels observed in our study.
The second androgen tested herein was the one that was differently expressed between sexes: 11KT. Males presented higher plasma levels of 11KT, which was reflected in the higher 11KT/T ratio also observed for males. The 11KT is the most potent androgen in teleost fishes; it is behind secondary sexual traits for males and females and is related to neurogenesis (see 33, a study with Mozambique tilapia, Oreochromis mossambicus), a process directly related to cognition 34. Despite that, there was no difference in cognitive and social performances between sexes, and males and females performed equally in both behavioral tests.
The pearl cichlid is a monogamous species in which androgen levels are usually similar between sexes, particularly when males are not defending territories (male-male challenges) and searching reproductive partners (male-female interactions), as during the breeding season 6,7. Goymann et al. (2019) state that males providing nest-tending are more attractive to females and, therefore, there will be no conflict between mating and parenting that causes increases in T levels, since males can attract sexual partners and care for the youngsters at the same time 7. Interestingly, Scaia et al. (2018) observed in their study that aggression in females of the Neotropical cichlid fish Cichlasoma dimerus, rather than being modulated by androgens, were determined by initial levels of estrogens, with female winners showing higher levels of E2 than losers 35. Therefore, these results show that other sex steroids, not measured in our study, may modulate female social behaviors.
Glucocorticoids, although primarily linked to stress responses, modulate social behavior and cognitive performance, such as learning and memory 21,36. For instance, social animals, such as the pearl cichlid, present high levels of cortisol when isolated, which will help the individual to cope with such a stressful situation (e.g., as an energy provisioner, glucocorticoids may allow the social animal to look for conspecifics). In our study, males and females showed no differences in cortisol levels, evidencing the same predisposition to do the tests, without any perceived stress biasing our results. Moreover, we cannot consider that the cortisol levels presented by our subjects were detrimental, since subjects did not present significant weight loss throughout the days of experimentation and were prone to do all the behavioral tests we proposed by swimming, exploring, and feeding in an apparently healthy way.
Females have brain and behavioral factors mainly linked to cortisol, whereas males present several associations with testosterone
Although they presented no differences when compared to each other, female and male pearl cichlids showed remarkable correlations regarding hormone levels, behavioral performance, and brain volumes. Cortisol levels were interestingly correlated to cognitive function and to telencephalon size in females, yet not in males. During the inhibitory control test, higher levels of cortisol were related to females that took fewer trials to learn the detour-reaching task in the training phase. These females also spent less time trying to cross the glass of the transparent barrier in the test phase (i.e., better memory recall and inhibitory control), showing that higher cortisol levels seem to facilitate cognitive performance. As mentioned, higher cortisol concentrations were also related to bigger telencephalons, an important brain area related to social behavior and spatial learning 37. Cortisol levels also had a tendency of positive correlation with bigger diencephalons. The diencephalon controls several pathways linked to social organization 38, such as gonadotropin and corticotropin-releasing factors; it is, therefore, logical that bigger diencephalons might be related to higher levels of cortisol.
Long exposure to higher levels of cortisol is known to cause damage in brain cells 20,22, reflecting in lower cognitive performances 21. Nevertheless, glucocorticoids can be related to an improvement in environmental perception and attention as well, which increases memorization and learning 39,40 — in our study, spatial learning demanded by the detour-reaching paradigm. Therefore, even though both sexes did not present different levels of cortisol when compared to each other, female pearl cichlids seem to be positively responsive to cortisol. Males, on the other hand, did not show any relation between cortisol and the other tested variables, revealing that this sex is somehow less susceptible to cortisol effects on aggression and cognitive performance, as well as on brain size. Such lower susceptibility may be associated to the cichlid’s role in natural environments, in which males are more territorial and aggressive than females before and during brood care 41.
Pearl cichlid females also showed that the more T was metabolized into 11KT, the less prone they were to attack their mirror image; in other words, the 11KT/T ratio of conversion was negatively correlated to aggressiveness in females. However, males presenting a higher 11KT/T ratio of conversion showed a positive correlation with higher levels of aggression, which is the opposite from what we observed for females. Brain structure volumes also presented different correlations in each sex, revealing, again, the evolutive organizational particularities driving different evolutionary pressures on males’ and females’ reproductive roles.
In females, 11KT and 11KT/T were negatively associated with the volume of the brain stem. This brain area is known to be linked to motor coordination and autonomic activity in vertebrates, such as cardiovascular, respiratory, and sexual function 42, as well as to the lateral line and to auditory information 43. The correlation observed herein could mean that this area of the brain helps fish to better interpret its environment. The brain stem has a lot of T neurons 42; it may also be the case for 11KT, although it has not been investigated yet.
As observed in tetrapods, the brain stem of teleost fish is also related to the auditory-vocal integration and presents the same auditory pathways 44. Teleosts also use acoustic information for sexual selection 45,46. In cichlids, females show higher levels of T and 11KT when they are close to spawn47,48. In G. brasiliensis, a lower ratio of conversion of T into 11KT linked to a bigger brain stem and followed by the same tendency regarding 11KT concentrations lead us to speculate that evolutionary pressure would select females with bigger brain stems which would balance the lower 11KT production. This could be the reason why a correlation between this androgen level with the brain stem size is observed in females, whereas in males it does not happen.
During the aggressiveness test, females showed a tendency to be less prone to fight when there was a higher 11KT/T ratio of conversion. Herein, we assessed aggression by measuring mirror-elicited fights. This methodology is pervaded by some concerns, and some researchers state that brain measures and hormonal responses are different when individuals fight with real vs virtual opponents 49. Nevertheless, it is still an efficient method to measure aggression and avoid physical injuries in cichlids, thus improving animals’ welfare and survival 50–52.
In this study, aggressiveness was inversely correlated with T and positively correlated with 11KT/T, meaning that individuals with higher T were less aggressive. On the other hand, 11KT seems to increase aggression. T levels in males, for instance, were related to decreased aggression (in both attacks and displays), a correlation not observed for females. We did not measure hierarchical status; thus, we cannot speculate about dominance or subordination driving changes in androgen levels. However, T is an androgen commonly related to high rates of aggression 6,12,25 yet we observed the contrary in this study — a negative correlation. Nevertheless, as mentioned, T seems to be more related to the maintenance of an individual’s status 31, and the lack of an intruder throughout the experimentation combined with a stable pairing to a same neighbor conspecific might be pointing to this status stability, given that blood sampling for hormonal analysis happened only on the last day.
In males, androgens also seem to be detrimental to cognitive performance, as individuals with higher levels of androgens took more time to conclude the detour-reaching task, with the trend of less correct responses as well. In the polygamous African cichlid Nile tilapia, males and females play very distinct roles during the reproductive season, with males being more aggressive and territorial, and leaving after reproducing with the females, which will care alone for the offspring inside their mouth 53. Using a similar approach, Brandão et al. (2019) observed that these different roles in the reproductive repertoire of Nile tilapia seem to lead to different cognitive demands in inhibitory control. In that study, females were quicker to learn the detour-reaching task than males 27, differently from the monogamous pearl cichlid, whose males and females obtained similar results in the same task, in spite of the apparent different action of T levels on cognitive performance. Moreover, the only relation regarding the brain volume in males was between the telencephalon and T levels, showing that the bigger that brain area, the lower the concentration of T in the plasma. Besides its role in spatial learning, the telencephalon is also related to the coordination of social abilities 38, and a negative correlation to a hormone linked to aggression corroborates this idea.
In male pearl cichlids, the higher the cortisol, the lower the 11KT levels. This is in accordance with other studies showing this negative correlation between cortisol and androgen levels 54,55. In teleost fish, this effect can be seen in the direct effect of cortisol on the production of 11KT, by affecting the producing capacity of the testis, and consequently, decreasing 11KT levels in the plasma, since cortisol is converted in the testis into 11-keto-derivatives 56. It is important to point that we only observed this negative correlation between cortisol and 11KT (and not with T) for males. We also collected the hormones only at the end of the experiment and we did not observe if there was a synchronized increase or decrease in cortisol/11KT during the experiment. Nevertheless, the correlation between these two steroids, already observed in other teleost fish, seems to exist in the pearl cichlid as well.
Surprisingly, the cerebellum and the diencephalon were not correlated to any hormone we tested herein, either in males or females. Both areas are known to be involved in several spatial-learning processes and coordination of social behaviors, respectively 57. This lack of correlation shows us that, although extensively found in the literature, the volume of brain areas might not answer all questions regarding correlations between brain and other physiological measures, as is the case of the hormones tested herein. Thus, other comparative methods (e.g., molecular analysis or quantification of the types of brain cells composing specific brain areas) could be more efficient to detect the associations among hormones, behavior, and brain in teleost fish.
In summary, our correlational study aggregated, for the first time, the two most important units of individual’s fitness: animals’ behavior (measured by aggressiveness and cognitive tests) and the physiology coordinating it (measured by hormone levels and brain size). Males and females of G. brasiliensis, although showing very similar roles during parental care, evidenced in similar aggressiveness and inhibitory control, show different patterns of brain and hormonal correlations. Nevertheless, further studies concerning more specific anatomical components — quantifying the number of different types of brain cells, for instance — are necessary in order to elucidate how brain and behavior work together in such socially complex individuals.