Neural control of maternal and paternal behaviors

See allHide authors and affiliations

Science  15 Aug 2014:
Vol. 345, Issue 6198, pp. 765-770
DOI: 10.1126/science.1253291


Parental care, including feeding and protection of young, is essential for the survival as well as mental and physical well-being of the offspring. A large variety of parental behaviors has been described across species and sexes, raising fascinating questions about how animals identify the young and how brain circuits drive and modulate parental displays in males and females. Recent studies have begun to uncover a striking antagonistic interplay between brain systems underlying parental care and infant-directed aggression in both males and females, as well as a large range of intrinsic and environmentally driven neural modulation and plasticity. Improved understanding of the neural control of parental interactions in animals should provide novel insights into the complex issue of human parental care in both health and disease.

Parental behavior aims at caring for conspecific young and increasing their survival. Among oviparous animals, parenting can include behaviors such as egg-laying site selection, nest building, burrowing, egg attending, and brooding and carrying the young; among viviparous animals, it can include food provisioning, nursing, defense of offspring, and even teaching of skills. Parenting occurs in a surprisingly large variety of vertebrates and invertebrates, including insects, arachnids, mollusks, fishes, amphibians, reptiles, birds, and mammals. In mammals, mothers commonly take the primary responsibility of parental care, whereas fathers often ignore or even attack the young. However, in many species, direct engagement of fathers has been observed; in some species, fathers participate equally or even exclusively in parental duties (Fig. 1).

Fig. 1 Paternal care can be observed in many different taxa.

(A) Giant water bug (Abedus herberti). (B) Los Tayos rocket frog (Hyloxalus nexipus). (C) Silverback mountain gorilla (Gorilla beringei beringei) father with infants. [Photos by Ivan Phillipsen (A), Adam Stuckert (B), Lubert Stryer (C)]

Nurturing and affiliative behavior toward infants is sensitive to physiological and environmental factors such as stress and hormone levels. In humans, the quality of parental care is affected by stress and mental illnesses such as postpartum depression (PPD), which affects more than 10% of mothers in the United States (1). How is the diversity of parental behavior generated in males and females, across different species, and in various physiological or pathological conditions? Recent studies have begun to uncover the nature and function of circuits underlying parental interactions with young. Here, we review data suggesting the existence of highly conserved and antagonistic circuits controlling affiliative and aggressive behavior toward offspring, respectively. Circuits underlying these opposing behaviors are present in both male and female brains irrespective of the normal expression of parenting displays and are modulated by intrinsic and environmental factors.

Diversity in parental care

Parental care has evolved repeatedly across vertebrate and invertebrate taxa (2). The involvement of males and/or females in the care for offspring varies across taxa and even between populations within a species (Fig. 2). In many systems, the parent that cares for offspring can be partially correlated with certainty in parentage and/or adult sex ratio. In mammals, male involvement is rare because internal fertilization ensures maternity but not paternity, and because only females lactate (3). In some rodents, canids, and primates, males assist and invest substantially in the care of offspring (48), whereas closely related species are exclusively maternal (912). For example, prairie voles and California mice are biparental, with males showing all female-typical parental displays except nursing (4, 5), but closely related species in the same genus—such as the montane vole, meadow vole, or deer mouse—are female uniparental (912). Cross-fostering experiments showed that meadow vole males reared by biparental prairie voles exhibited significantly more paternal care to their offspring than in-fostered counterparts (13). This result demonstrates the influence of early social environment, in addition to genetic differences between congeneric species, on parental behavior.

Fig. 2 Evolution of diverse and distinct parental cares strategies across the animal kingdom.

Examples of different parental care strategies are shown across vertebrates and invertebrates. Male uniparental care is lacking only in the mammalian and reptilian lineages, although there are male-biased parental care systems in few canids and primates. (A) Common deer mouse (Peromyscus maniculatus). (B) Oldfield mouse (Peromyscus polionotus). (C) Kori bustard (Ardeotis kori). (D) Adélie penguin (Pygoscelis adeliae). (E) Pheasant-tailed jacana (Hydrophasianus chirurgus). (F) Water python (Liasis fuscus). (G) Black rock skink (Egernia saxatilis). (H) Diablito frog (Oophaga sylvatica). (I) Mimic poison frog (Ranitomeya imitator). (J) Dyeing poison frog (Dendrobates tinctorius). (K) Burton’s mouthbrooder (Astatotilapia burtoni). (L) Convict cichlid (Amatitlania nigrofasciata). (M) Three-spined stickleback (Gasterosteus aculeatus). (N) Golden brown stink bug (Anchises parvulus). (O) Burying beetle (Nicrophorus vespilloides). (P) Giant water bug (Abedus herberti). [Photos by Andrés Bendesky (A and B), Tamas Szekely (C), Oliver Kruger (D), Ghulam Rasool (E), Zachary Stahlschmidt (F), Alan Couch (G), Elicio E. Tapia (H), Evan Twomey (I), Lauren A. O’Connell (J), Rayna Harris (K), Bryan J. Matthews (L), Dwight Kuhn (M), Peter Chew at (N), Allen Moore (O), Michael Bogan (P)]

Male involvement in offspring care is common in many taxa other than mammals. In teleost fish species, males provide care more often than females, including nest building and egg attendance (14). In the well-known case of the three-spined stickleback, males set up the territory, build nests, and defend their offspring (15). In birds, 90% of the species are biparental, with both parents sharing the responsibilities of building a nest, incubating eggs, and defending and feeding the young (16). The sex ratio of individuals available to mate in a bird population largely determines which parent cares for offspring. For example, male shorebirds are more likely to care for offspring in populations where males are more abundant than females (17).

Amphibians display striking diversity in parental care. Many species of anurans and salamanders display care for offspring beyond egg laying, with roughly 50 independent evolutionary transitions to parental care (18). These behaviors include preparation of foam nests (19), egg guarding, transport of offspring piggyback style (20), and egg incubation in dorsal pouches, vocal sacs, or the stomach. South American poison frogs (Dendrobatidae) show particularly striking diversity in which sex cares for offspring within closely related species (21).

Such natural diversity in parental care strategies across large evolutionary distances—as well as in related species and in individuals within a species—suggests the existence of conserved neural pathways underlying parental care that are differentially regulated in males and females, and in different species.

Sensory cues that drive parental interactions

Neuroethologists have long recognized intriguing differences in the nature and complexity of signals driving parental behaviors. In fish and birds, social behaviors including care of young were often seen to be triggered by simple cues (22). The domestic hen, for example, comes to the immediate rescue of a chick after hearing its distress call, but the sight of a struggling chick without sounds leaves the fowl indifferent (23). In turn, the reliance on simple visual signals in some species of birds generates the so-called “supernormal” stimulus effect, in which artificial eggs with exaggerated features (such as higher-contrast pigmentation or giant sizes) are even more effective at eliciting parental behavior than natural eggs (22).

In mammals, multiple sensory modalities have been shown to trigger maternal responses. Early studies in rats (24) found that blind, anosmic, or anaptic lactating females each retrieve pups in a fashion not significantly different from controls. However, the combination of anosmia and tactile deprivation results in more pronounced defects in retrieving than does the loss of either sensory system alone, and the defects are even more severe when all three sensory inputs were eliminated. Different sensory modalities appear to often synergize with each other, as they perform critical roles in different steps of the parental response.

In rodents, low-frequency wriggling calls emitted by pups when they struggle in the nest induce licking, change of suckling position, and nest building by the mother (25). By contrast, ultrasonic vocalizations produced by pups lost outside the nest trigger immediate search for the pups and retrieval to the nest (26), with retrieval occurring even if the ultrasonic vocalizations have ceased. Intriguingly, mouse fathers can be induced to display fast pup-retrieving behavior by 38-kHz ultrasonic vocalization from their female partners (27).

Chemosensory cues are extensively used to elicit or inhibit parental care according to the sex and physiological status of the animal. Many amphibians, fish, birds, and insects, such as ants or the burrowing bug, were shown to use olfactory cues to recognize offspring (2831). In many mammals, the vomeronasal pathway, in conjunction with the olfactory system for some species, inhibits parental behavior and drives pup-mediated aggression in virgin animals, whereas olfactory cues are often seen facilitating the care of offspring in parents or primed animals (29). Virgin rats initially find foreign pups aversive but exhibit parental care after continuous exposure to the pups (30). Surgical removal of the vomeronasal organ (VNO) reduces infanticidal behavior and induces faster paternal behavior (31); however, olfactory cues emitted by pups appear to facilitate parental care (29, 32). Recent studies in mice confirmed that surgical or genetic VNO loss of function leads to a marked reduction in pup-directed aggression and to the emergence of parental care in virgin males (33, 34). In humans, one study documented a much higher rating of infant body odors by postpartum mothers than by nulliparous women (35), and odors have been proposed as important cues in early interactions between mothers and infants (36, 37).

A fascinating example of multisensory interaction comes from the mother-infant bonding in sheep (38). Olfactory cues are responsible for both inhibiting maternal responsiveness of ewes before parturition and for attraction to amniotic fluid immediately after parturition (39). Shortly after pregnant ewes give birth, a selective bonding between the mother and the infants rapidly develops, such that ewes nurse only their own offspring and behave aggressively toward alien young. However, artificial vaginocervical stimulation that mimics the expulsion of the lamb resets the ewe’s olfactory preference toward an alien lamb, likely through oxytocin release.

The sensing of infant cues is remarkably enhanced in parents, and parturition and maternal care have been associated with multisensory facilitation and extensive cortical plasticity. Recordings of ultrasonic calls played to lactating female rodents showed that searching behavior is facilitated by pup vocalizations in the presence of olfactory cues (40, 41). Neuron responses and population dynamics in the auditory cortex undergo significant changes in mothers relative to virgin female mice, likely facilitating the representation of pup vocalizations and enhancing their behavioral relevance (4245). Moreover, a significant modulation of sound-evoked responses by pup odors has been shown in the primary auditory cortex of lactating female mice shortly after parturition, with neurons from lactating mothers displaying more sensitivity to sounds than virgins (44). In the olfactory system, mitral cells in the olfactory bulb of female ewes have been shown to undergo marked changes in sensory responses to lamb versus food odors after parturition (46, 47). Moreover, in vivo time-lapse imaging of adult newly born granule cells in mice showed an enhanced integration of these neurons into the olfactory circuit of lactating mothers (46).

Switching between parental care and aggression

Infant-directed aggression is prevalent in animals that are not rearing offspring (such as virgin animals) and in sexually mature stranger males (48), and it is often seen to switch into affiliative behavior after birth of offspring or habituation to the young. Thus, adult animals may display parental care or aggression according to their physiological and environmental state, and the regulation of affiliative versus agonistic behavior circuits raises an important and fascinating question in the study of parental interactions.

In laboratory mice, infanticide is commonly observed in virgin males (49). Males stop committing infanticide and become paternal toward pups in a transient period after mating with a female, starting at the approximate time of birth and continuing until the weaning of pups (Fig. 3) (50). The coincidence of the suppression of infanticide in males and the birth of their own pups likely provides an adaptive mechanism that prevents a male mouse from killing its own pups, but successfully eliminates pups sired by competing males. Parental males and females, however, do not appear to differ in the overall incidence of retrieving, nest building, licking, and huddling over the pups (51). Wild-caught female mice are typically infanticidal, and they follow a similar transition to parental care associated with parturition and lactation, with a surprising elevation of infanticide throughout pregnancy (52, 53). The drastic difference between laboratory and wild female mice suggests that infanticide was selected out by colony breeding in females.

Fig. 3 Pup-directed behavior of males at different days after mating [replotted from (50), table 2].

Adult male mice of the CF-1 strain were mated with females, randomly assigned into groups, and tested at different days after mating. Control virgin males are plotted at day 0. After a significant increase in pup-directed aggression at day 4, there is a transient suppression of attack and increase in paternal care in the males from day 12 to day 50, which approximately corresponds to the birth of and the weaning of their own pups. This experiment illustrates a remarkable influence of mating on male parental behavior.

Time-dependent synaptic or transcriptional change triggered by mating, as well as the chemical cues released by females during pregnancy (54), have been hypothesized to drive the radical behavior shift from infanticide to parental behavior (50, 5557). The timing and mechanism of the mating-induced behavioral switch in mice has been assessed by two recent studies (33, 34). After pup exposure, neurons in the vomeronasal pathway appear more strongly activated in virgin males than in fathers (33), and impairment of VNO sensing results in decreased pup-directed aggression and induction of parental care (33, 34). These results raise the intriguing possibility that the transition of attack to parenting could be due to a time-dependent reduction of vomeronasal activation by pup cues in males.

The intriguing temporal switch in offspring recognition associated with mating is not restricted to rodents; it has been observed and characterized in a variety of species including isopods, burying beetles, African cichlids, and birds (58).

Neural circuits underlying parental interactions

Much of our knowledge about neural circuits underlying parental behavior comes from studies in rats, with recent insights provided by genetic studies in mice. In contrast to lactating females, which are highly maternal, virgin male and female rats usually avoid physical contact with foreign pups. Nonetheless, after continuous exposure to pups, virgin males and females approach, interact with them, and eventually exhibit parental care, in a process termed “sensitization” (30).

The changes in levels of female hormones such as estrogen, progesterone, and prolactin through pregnancy have long been implicated in the regulation of maternal behavior (59). Treatment of virgin females with a regimen of hormones mimicking this pattern facilitates the display of maternal behavior (60). Moreover, recent genetic studies have shown that the prolactin receptor is essential for the normal display of maternal behavior (61). Prolactin is also an important regulator of parental care in nonmammalian vertebrates, most notably in birds and teleost fish, where prolactin rises ­during egg laying/spawning and remains elevated throughout the duration of parental care (62, 63).

Male interaction with infants is also influenced by hormonal changes (64). In many vertebrate species where males are involved in offspring care, testosterone levels decrease during fatherhood, such as in humans, frogs, and fish (6567). The intrauterine position of mouse fetuses, and therefore the early exposure to different levels of sex hormones, has been proposed to influence pup-directed aggression in adulthood (68). In addition, progesterone receptor knockout virgin male mice were shown to exhibit little aggression but elevated parental care toward foster pups (69).

The contrast between caring by parents and aversion by virgin animals has led to searches for brain areas involved in the stimulation and inhibition of maternal behavior (70) (Fig. 4). Classical mapping experiments have demonstrated the essential role of several brain areas in the control of maternal behavior, including the medial preoptic area (MPOA) and the adjacent ventral bed nucleus of stria terminalis (vBNST) (Fig. 4A) (70, 71). A combination of immediate early gene (IEG) mapping and tracing further mapped the projection sites of the active MPOA/vBNST neurons (Fig. 4A) (72). In addition to the preoptic area, the lateral septum is also involved in the regulation of parental care (73), and both areas have been implicated by IEG studies in the paternal care of biparental rodents (74) and biparental cichlid fish (67). Electrical stimulation of the preoptic area in male bluegill sunfish also elicits paternal care (75). These results suggest that highly conserved circuits and neuroendocrine mechanisms may be repeatedly recruited to mediate similar social behaviors (2, 76). What specific information is carried by these brain regions and how they encode the various components of parental care remain to be determined.

Fig. 4 Schematic presentation of brain areas associated with parental care (A) and pup-directed avoidance and aggression (B).

Solid lines denote projections that are involved in the regulation of pup-directed behavior, as supported by direct evidence. Dashed lines denote known connections that exist between these areas and are potentially involved in the behavior. The lines and arrows simply denote origins and targets and do not represent actual axon path or excitatory inputs. Not all the known connections are shown. Abbreviations of brain areas: MOB, main olfactory bulb; PFC, prefrontal cortex; NAc, nucleus accumbens; VP, ventral pallidum; LSd, lateral septum, dorsal part; LSi, lateral septum, intermediate part; vBNST, bed nucleus of stria terminalis, ventral part; MPOA, medial preoptic area; PVNm, paraventricular nucleus, magnocellular part; BL, basolateral amygdala; Ce, central amygdala; VTA, ventral tegmental area; PAG, periaqueductal gray; Raphe, Raphe nucleus; LC, locus coeruleus; AOB, accessory olfactory bulb; LSv, lateral septum, ventral part; BNSTpr, bed nucleus of stria terminalis, principal nucleus; PVNp, paraventricular nucleus, parvocellular part; DMH, dorsomedial hypothalamic nucleus; VMH ventromedial hypothalamic nucleus; MeA, medial amygdala; PMd, premammillary nucleus, dorsal part.

A similar set of experiments uncovered a parallel neural system that inhibits maternal behavior, thus opposing the function of the parental pathways described above (Fig. 4B). In particular, the medial amygdala, which receives direct projection from the accessory olfactory bulb, was shown to mediate the suppression of maternal care and the initial avoidance responses in virgin female rats (77). A number of other brain areas, many of them interconnected and involved in defensive social encounters, were also shown to inhibit maternal responses (78, 79); such findings suggest that pup aversion may share common circuitry with defensive behavior.

From these studies, a hypothetical neural model of the control of parental behavior in rats has been proposed, according to which two competing pathways mediate active maternal responses and aversive behavior toward pups, respectively (80, 81). In male and most female virgin rats, the aversive circuit, primarily innervated by vomeronasal inputs, is dominant and suppresses parental care, whereas in postpartum and “sensitized” females, hormonal, neuromodulatory, and experience-dependent factors activate the facilitative circuit and silence the avoidance circuit. Uncovering how these two conflicting circuits are differentially modulated in different physiological or environmental conditions is therefore central to the understanding of the control of parental care in males and females of various species.

The ventral tegmental area (VTA) is a major dopaminergic area that is involved in reward and reinforcement learning. Pups are known to be a reinforcing stimulus to postpartum females (82), and MPOA lesions were found to disrupt the performance of an associative learning task using pups as positive reinforcement (83). Moreover, inactivation of VTA projections disrupts maternal behavior in postpartum rats (84), and depletion of dopamine in the ventral striatum or lesion of dopamine neurons in the VTA causes a persistent deficiency in pup retrieval (85, 86). These results suggest that the dopaminergic system helps initiate and maintain maternal behavior in rats, likely by engaging the MPOA (71).

Targeted disruption of the dopamine β-hydroxylase (Dbh) gene, which synthesizes noradrenaline and adrenaline, leads to severe defects in maternal behavior (87). Intriguingly, providing noradrenaline precursor at the time of parturition is sufficient to restore maternal behavior in Dbh mutant females and maintain maternal care toward their future litters; this finding suggests that noradrenaline is critical at birth for the formation of a stable behavioral memory, which in turn is responsible for the maintenance of maternal care (87, 88).

The role of the serotonergic system was recently demonstrated by the maternal defects of Pet-1 (an ETS transcription factor whose brain expression is limited to serotonin neurons) knockout mice, in which serotonergic gene expression and serotonin synthesis are greatly reduced (89). The MPOA and the BNST are innervated by serotonin-immunoreactive fibers (90), which suggests that the maternal deficiency may result from impaired serotonin inputs to these areas.

The highly conserved neuropeptide oxytocin is also an essential regulator of parental care across animals (91). Female mice deficient in oxytocin are unable to nurse, although they display largely normal maternal behavior (92). Studies using oxytocin receptor knockout females found no obvious deficits in their maternal care (93), but a recent reexamination of their behavior suggested that oxytocin is involved in the initiation but not the maintenance of maternal behavior (94). In addition to mammals, the function of oxytocin appears to also extend to other vertebrate systems including birds (95) and fish (67).

The brain regions involved in the control of parental behavior are highly heterogeneous structures, and newly designed molecular and genetic tools make it possible to identify and functionally manipulate precise subsets of neurons, thus enabling a deeper understanding of the associated behavior circuits.

A recent study uncovered a subset of MPOA neurons expressing the neuropeptide galanin that are specifically activated during male and female parenting (Fig. 5A) (34). Specific ablation of MPOA galanin neurons in virgin females, mothers, and fathers results in a marked impairment of parental responses and induced pup-directed aggression in virgin females (Fig. 5, B and C). In contrast, optogenetic activation of these neurons in virgin males suppresses pup-directed aggression and induces pup grooming (Fig. 5D). These results suggest a direct role of MPOA galanin neurons in activating parental responses and confirm the suspected reciprocal inhibition between circuits activating and repressing parental behavior. The identification of MPOA galanin neurons as an essential regulatory node of male and female parenting behavior provides a precious entry point for further dissection of behavior circuits underlying parental care and their modulation by social experience.

Fig. 5 MPOA Gal neurons serve as an essential regulatory node for parental care in both male and female mice.

(A) Co-labeling of c-fos and Gal in the MPOA of parenting females. (B) Cumulative percentages of virgin females that retrieved or attacked pups as a function of the percentage of remaining Gal cells after Gal cell ablation. Reference cell number (100%) is the average MPOA Gal cell number in the control group. (C) Cumulative percentages of fathers that retrieved pups as a function of remaining Gal cells after Gal cell ablation. (D) Behavior raster plots after optogenetic activation of Gal cells in virgin males interacting with pups. The control group consisted of littermates that do not express light-activated channelrhodopsin-2. Different behavior elements are color-coded and could occur simultaneously.

Emerging evidence suggests that highly conserved circuits and modulatory mechanisms may exist across species and in both male and female brains to regulate parental interactions with offspring. Remarkably, the natural behavior of adults toward infants emerges as the mutually exclusive output of two highly regulated circuits driving affiliative versus aversive responses. Future studies should exploit the natural diversity of parental systems across animal species to gain mechanistic insights into the regulation of parental behavior in physiologically and ecologically relevant contexts. Such results are likely to shed new light on the complexity of human parental behavior and its susceptibility to mental illness.

References and Notes

  1. Acknowledgments: We thank A. E. Autry and A. Bendesky for comments on the manuscript, and R. Hellmiss for assistance with figure artwork. Supported by the Howard Hughes Medical Institute (C.D., Z.W.), the National Institute on Deafness and Other Communication Disorders (C.D., Z.W.), and the William F. Milton Fund (L.A.O.). We apologize to colleagues whose work we could not cite because of space constraints.
View Abstract

Navigate This Article