pe rs pect ive s - university of...

PERSPECTIVES

Model systemsTo study the neural basis of attachment, weneed model systems with three features.First, we need to be able to observe a clearonset of the behaviour to identify factorsthat initiate or inhibit the formation ofattachment bonds. Second, attachmentbehaviour must be, by definition, selectiveand enduring. Selectivity distinguishesattachment from generalized social interac-tion. Duration distinguishes a bond from atransient preference. And finally, we need tobe able to measure and manipulate thesebehaviours. Studies of social affiliation andattachment have spanned gene targeting inCaenorhabditis elegans2 to psychodynamicapproaches in humans3. In mammals, ele-gant studies of non-human primates, par-ticularly of monogamous species includingthe New World callitrichids who showbiparental care4 and the Old World titimonkey in which pair bonds are expressedby tail twining5, have described the behav-ioural features of social attachment. Thereis, as yet, no indication of neural systemsthat are involved in pair bond formation inthese species. With the exception of phar-macological studies in maternal monkeys6

and recent human imaging studies7,8, inves-tigations of neural systems that are impor-tant for attachment have so far used non-primate mammals. Investigations of themolecular and cellular mechanisms thatunderlie these behaviours have focused on:first, infant attachment to a parent; second,maternal behaviour in species with selectivecare of their young; and last, partner prefer-ence formation in species with long-term,selective bonds.

Infant–mother attachmentThe study of neural mechanisms that under-lie infant attachment has progressed furthestin species with precocial offspring. Within adiscrete developmental window, the newlyhatched chick shows visual imprinting, anenduring selectivity for following the mother(or a mother-like object) that can be quanti-fied with great accuracy. Imprinting in thechick is not a single process but consists of atleast three largely independent processes thatare relevant to all other forms of attachment.First, there is the approach response, which isassociated with increased arousal and inhibi-tion of avoidance. This is followed by theacquisition or learning stage when chicksform a long-term memory for the imprintedstimulus, a stimulus that is partially pre-specified9. Finally, marking the end of thesensitive period for learning, there is a rever-sal of the approach–avoidance bias as chicksbegin to avoid new objects while continuingto follow the familiar imprinted object.

A region within the intermediate medialpart of the hyperstriatum ventrale (IMHV)of the chick brain is critical for acquisitionand early consolidation of the memory of animprinted visual stimulus10. A related region,the mediorostral neostriatum11, respondsselectively to imprinted auditory stimuli12,13.The learning phase of imprinting in thechick, whether it is visual or auditory, involvesearly and persistent enhancement of pre-synaptic release of amino acids, as well aschanges in postsynaptic ultrastructure withinspecific cortical regions10. It is not clear how,or if, these changes differ from other forms ofvisual or auditory learning in chicks or whichsystems modulate the approach–avoidancechanges that are necessary for imprinting.

The study of infant attachment in mam-mals has not identified a specific neural cir-cuit or predominant neurochemical system.In rat pups, a great number of neurochemicalsystems increase or decrease the number ofultrasonic distress calls14,but the response toseparation might be neurochemically distinctfrom the process of attachment. We know

It is difficult to think of any behaviouralprocess that is more intrinsically importantto us than attachment. Feeding, sleepingand locomotion are all necessary forsurvival, but humans are, as BaruchSpinoza famously noted, “a social animal”and it is our social attachments that we livefor. Over the past decade, studies in a rangeof vertebrates, including humans, havebegun to address the neural basis ofattachment at a molecular, cellular andsystems level. This review describes someof the important insights from this work.

Social attachment is not only intrinsicallyimportant, it is intrinsically difficult to study.One of the early pioneers in this area, HarryHarlow, described the different behaviouralprocesses that are involved in the formation ofparent–infant, filial and pair (male–female)bonds1. Each of these involves multi-sensoryprocessing (predominantly olfactory inrodents and visual in primates) and complexmotor responses (for example, proximityseeking, nurturing responses and defensivebehaviours). Attachment also requires severalcognitive processes that link sensory inputs tomotor outputs, including attention,memory,social recognition, and,perhaps most charac-teristically, motivation. In non-human ani-mals, this motivational aspect of attachmentcan be assessed as proximity seeking, a socialpreference or a separation response. Inhumans, the ultimate form of this motivationis what we experience as ‘love’. Recent studieshave begun to reveal neural mechanisms forsocial recognition,nurturing behaviour and,most importantly, the development of specificsocial preferences.

NATURE REVIEWS | NEUROSCIENCE VOLUME 2 | FEBRUARY 2001 | 129

The neurobiology of attachment

Thomas R. Insel and Larry J. Young

OP I N ION

© 2001 Macmillan Magazines Ltd

130 | FEBRUARY 2001 | VOLUME 2 www.nature.com/reviews/neuro

P E R S P E C T I V E S

retrieve and defend young, and develop aspecies-typical arched back posture for nurs-ing20.Many of the same neuroendocrine fac-tors that are associated with pregnancy, par-turition and lactation are also critical in thistransition from avoidance to nurturingbehaviour. Indeed, the changes in oestrogenand progesterone that accompany gestationand delivery are sufficient for inducingmaternal behaviour in virgin females21.However, we still know relatively little aboutwhere or how these gonadal steroids affectneuronal activity to facilitate maternalbehaviour.

Although there is no defined nucleus orcircuit for motherhood in the rat brain, lesionstudies, patterns of immediate-early geneexpression and pharmacological manipula-tions have all implicated the medial preopticarea (MPOA), the overlying bed nucleus of thestria terminalis (BNST) and selected projec-tion sites, including the lateral habenula andthe ventral tegmental area (VTA)20,22,23. Theabundant literature on the MPOA has still notdefined how this structure changes with theonset of maternal behaviour,but there is con-siderable evidence from lesion studies that theMPOA is particularly critical for the integra-tion of perioral sensory cues that permit nestbuilding and retrieval of pups24.

Neuropeptides, such as prolactin and oxy-tocin, which are involved in lactation, havebeen implicated as central neuroendocrinemediators of rat maternal care. In rats, pro-lactin administration facilitates maternalbehaviour in a steroid-primed non-pregnantrat and treatments that decrease prolactin lev-els inhibit maternal care25. A mouse with aprolactin receptor null mutation showsimpaired maternal retrieval and nest build-ing26.Oxytocin, a uniquely mammalian neu-ropeptide, also facilitates the onset of maternalbehaviour27,but is not required once maternalbehaviour is established28. These results haveindicated that oxytocin might be importantfor the transition from avoidance to approachof the young. The facilitation of maternalbehaviour by oxytocin might be mediatedeither through the VTA, the MPOA or fromwithin the olfactory bulb, as injections of aselective oxytocin antagonist into each of thesesites can block the onset of maternal care29,30.

It is important to realize that in rats, as inmany mammals, the onset of maternal behav-iour involves overcoming a natural avoidanceof neonates and, specifically neonate odours.Thus, in rat females, the initiation of mater-nal care is facilitated by lesions that reduceolfactory processing31. Oxytocin that isreleased centrally at parturition32 seems todecrease the firing rate of mitral and granule

mals, we know remarkably little about howsocial attachment, as opposed to generalenvironmental enrichment18, modifies thedeveloping brain.

Mother–infant attachmentMammalian maternal behaviour is extremelydiverse19. At one extreme are the minimallymaternal eutherian species such as tree shrewsand rabbits that spend only a few minuteseach day in contact with their young. At theother extreme are species, including manyprimates, that seem ‘promiscuously parental’,showing maternal behaviour throughout thelife cycle. Between, there are many species forwhom maternal care is restricted to the post-partum period, providing a clear onset andoffset, with an opportunity for the study ofneural mechanisms of this behaviour.

Maternal motivation — the rat. The Norwayrat is a species in which females either activelyavoid or attack pups until just before deliv-ery, at which time they begin to build a nest,

more about how infants learn to identifytheir mothers. Unlike imprinting in chicks,maternal recognition is largely an olfactoryprocess in rat pups, involving noradrenaline-mediated pathways for olfactory learning15.The neuropeptide oxytocin has a curiouseffect on olfactory learning in rat pups.Pupscan be rapidly conditioned to stimuli thatare associated with maternal odours ormaternal care16.Oxytocin facilitates learningin pups when the association is to socialcues, such as the mother, but it fails to alterlearning that is associated with non-socialstimuli17. Conversely, an oxytocin antagonistdelays this form of conditioning, indicatingthat this neuropeptide might be importantfor forming associations that are specificallyrelated to the mother17. In contrast to thestudies of imprinting in chicks, so far thereare no comparable reports that identify aspecific cortical region for attachment in theneonatal mammalian brain. Although socialexperience is critical for the normal develop-ment of the brain and behaviour in mam-

Functional significance not yet determined

Vaginocervical stimulationduring parturition

Decreases agression and aversion to newborn lambs

Noradrenaline

Oxytocin

OxytocinParaventricular nucleus

Oxytocin-mediated neurons

Spinal cord

Brain stem

A1, A2, A6:noradrenergic neurons

Olfactory bulbMedial preoptic area

Cingulatecortex

Lateral septal nuclei

Rat only: enhancesmaternal nuturing

behaviour

Ventral tegmental area

Post-partum oestrus

Mediobasalhypothalamus

Figure 1| Oxytocin and maternal behaviour in the sheep brain. Within 2 hours of parturition the ewedevelops a selective, permanent bond to her lamb. One neurobiological model for this process43 positsthat afferent stimulation through the spinal cord from vaginocervical dilation during parturition increasesthe activity of noradrenaline cells in the brainstem (A1, A2 and A6), which project to the paraventricularnucleus (PVN) in the hypothalamus, as well as to the olfactory bulb. Stimulation of oxytocin cells in thePVN facilitates maternal behaviour through coordinated effects on several regions in which oxytocinincreases GABA (γ-aminobutyric acid) and noradrenaline release. Oxytocin in the olfactory bulb and medialpreoptic area reduces aggressive or aversive responses to newborn lambs. Oxytocin in the mediobasalhypothalamus inhibits post-partum oestrus. On the basis of data from rats, oxytocin in the ventraltegmental area might facilitate the onset of maternal nurturing behaviours, although this has yet to beshown in the sheep brain. Projections to the lateral septum and cingulate cortex have not yet been studiedfor their functional significance. (Figure modified from REF. 43 © (1997) with permission from ElsevierScience, and includes release data based on microdialysis studies and behaviour data based onretrodialysis of oxytocin into regions noted.)



bulb) is reduced during parturition and virtu-ally absent after VCS in inexperienced ewescompared with experienced females43,45. Afterdevelopment of a bond, even within a fewhours,VCS can stimulate oxytocin release andinduce maternal acceptance for the rest of theewe’s life. This permanent effect of experienceon a neurochemical and behavioural responseto a simple sensory input is reminiscent ofimprinting in chicks. But how does experienceresult in a permanent reorganization of theresponse to lambs? And how does this accep-tance become selective, such that a single lambis accepted but all others are rejected? Ananswer to this latter question requires aninvestigation of olfactory learning.

Olfactory learning. In contrast to declarativememory or various types of hippocampallearning that have been studied in rats andmonkeys46, the form of permanent ‘imprint-ing’ that occurs in sheep seems to involve pri-marily a reorganization of the olfactory bulb47.Early in the post-partum period, the mother’sselective bond with her lamb can be shown bytwo marked physiological changes in additionto her behaviour. Recordings in the mitral celllayer of the bulb during the first weeks post-partum reveal a 60% increase in the numberof cells that respond preferentially to lambodours, compared with those made duringlate pregnancy47.Of this group of cells,~30%respond specifically to the ewe’s own lamb andnot to other lambs47. In addition to this physio-logical correlate of selectivity, within 24 hoursof parturition, the ewe’s own lamb odour elic-its a robust increase in extracellular concentra-tions of glutamate and GABA (γ-amino-butyric acid) within the olfactory bulb47. Thiseffect seems to be selective, as neither gluta-mate nor GABA is released in response to

cells in the bulb33, thereby decreasing olfacto-ry processing and facilitating approachbehaviour. This might be a critical step forthe initial acceptance of pups (in addition tothe effects of the peptide on maternal moti-vation) and it might be mediated throughthe VTA. The onset of maternal care in therat requires at least two potentially neurobio-logically independent events: first, inhibitionof the avoidance/attack response to pupodours; and second, subsequent initiation ofnurturing responses to pups.

Selective maternal care — sheep. A post-par-tum rat is virtually a maternal machine, car-ing for any pups placed in her nest. Ratmaternal care may be intense, reflecting‘attachment’ to a generic pup, but it is notselective. Sheep, which are highly selective,have proven more rigorous models ofattachment, because the post-partum ewerejects any lamb that is not her own. In addi-tion to overcoming avoidance and initiatingnurturing behaviour as observed with rats,the ewe must learn who is her lamb. Being aherd animal and seasonal breeder, she mustlearn this individual recognition quickly andaccurately.

As with rats, gonadal steroids are impor-tant for priming the onset of maternal behav-iour in sheep. But, in contrast to rats,prolactindoes not seem critical for sheep maternalbehaviour34.Much of our understanding ofthe onset and selectivity of maternal care insheep has been based on the curious observa-tion that vaginocervical stimulation (VCS)can induce almost immediate maternalbehaviour in a steroid-primed non-pregnantewe35. Furthermore, VCS will induce accep-tance of an alien lamb even two to three daysafter she has bonded with her own lamb36.Epidural anaesthesia blocks these effects ofVCS, indicating that central feedback mightbe essential34.

How does the process of birth or VCSinduce acceptance of a lamb? VCS and birthare both potent stimuli for release of the neu-ropeptide oxytocin within the central nervoussystem (CNS)37 and, as with rats, oxytocinseems to be important for the onset of mater-nal care.Oxytocin can facilitate acceptance ofan alien lamb, even in a non-pregnant ewewithin 30 seconds of intra-cerebral ventricu-lar (icv) injection38,39. In post-partum femalesin which maternal behaviour has been pre-vented by epidural anaesthesia, icv adminis-tration of oxytocin can induce a maternalresponse40. Although the location of theeffects of oxytocin is not entirely clear, oxy-tocin mRNA and oxytocin receptor mRNAare increased regionally in the sheep brain

post-partum41,42. A series of site-specific injec-tion studies by Kendrick and colleagues pointto a distributed network of effects with someregions that are important for selectivity andothers that are related to nurturance43 (FIG. 1).

Injections of oxytocin into the MPOA orolfactory bulb reduce rejection of an alienlamb, much as described in rats, but theseinjections in sheep are not sufficient forinducing nurturing behaviour. Injections

near the oxytocin synthesizing cells of theparaventricular nucleus of the hypothalamus(PVN) induce the entire complement ofsheep maternal behaviours,possibly becauseautoreceptors on these cells mediate a shortpositive-feedback loop to increase oxytocincell firing and coordinate oxytocin release inseveral terminal fields44.

Although considerable data might indicatethe importance of oxytocin release in the CNSfor the initiation of maternal care, the effectsof oxytocin interact with experience, as theyare more evident in females with a history ofparental care. Oxytocin can induce lambacceptance in an inexperienced ewe in whichperidural anaesthesia blocks maternal behav-iour40, but oxytocin release (in the olfactory


Box 1 | Mechanisms for olfactory learning

Mitral cells contain glutamate and are closely regulated by inhibitory synapses with granule cellsand periglomerular cells. According to the model developed by Kendrick and co-workers, duringvaginocervical stimulation (VCS) or birth, increased noradrenaline-mediated stimulation leadsto decreased release of GABA (γ-aminobutyric acid) from granule cells, disinhibiting mitralcells47.With incoming olfactory signals from the lamb in the presence of this disinhibited state,mitral cells increase their release of glutamate, which activates ionotropic autoreceptors andprovides a short loop positive feedback on mitral cell activation. This model is consistent withthe results of in vivo microdiaysis experiments: extracellular concentrations of both glutamateand GABA within the bulb increase in response to lamb odours, with the ratio favouringglutamate.With concurrent granule cell and mitral cell activation, there is an increase in nitricoxide (NO) generation48. Mitral cells (but not granule cells) have the guanylyl cyclase subunitsthat are necessary for generation of cyclic GMP. The increase in cGMP in mitral cells facilitatesthe release of glutamate. Blockade of either neuronal nitric-oxide synthase or guanylyl cyclaseactivity blocks the increase in glutamate and GABA release and also prevents formation of thememory of the ewe’s own lamb48. The same treatments have no effect 24 hours post-partum,after the ewe has learned her lamb’s odour.

“… recent studies withchicks, rats, sheep, voles andnow humans have begun toreveal some importantcandidates for theneurobiology of socialattachment”.




changes have now been observed with olfac-tory conditioning in mice49. In sheep, thesechanges seem to be under the influence ofgonadal steroids50, but it is otherwise not yetclear how the process in sheep is distinctfrom olfactory learning that does not involvean enduring, selective attachment to theyoung. Specifically, we do not know how, insheep, the ‘imprinting’ of the lamb’s odourpermanently opens a sensory gate that signalsacceptance to the ewe, whereas rejectionremains her response to all other lambs. Thepermanent experience-dependent changes inthe post-partum ewe offer an excellentopportunity for identifying the cellularmechanisms for long-term memory in eitherthe olfactory bulb or the higher order stagesof olfactory processing.

Studies in sheep might not only yieldimportant clues for the cellular mechanismsof long-term memory, they might also holdgreat promise for revealing the neural mecha-nisms of attachment. The key will be to linkthe process of olfactory learning to the moti-vation for maternal care. Somehow theprocess of birth (or VCS) leads to willingnessto interact with a lamb and then within a fewhours, a permanent change in the behaviour-al,physiological and neurochemical responsesto this and no other lamb. The investigationof molecular and cellular changes in the ewe’sbrain during this initial period of interactionwith the lamb, like the process of imprintingin the chick, should reveal critical clues to theneurobiology of attachment.

Adult–adult pair bond formationAbout 5% of mammals are monogamous andbiparental51,52.With the development of quan-titative, operational definitions of variousbehavioural aspects of pair bonding, the voles(microtine rodents) have proven to be excel-lent model species for molecular and cellularstudies of complex social behaviours53. TwoNorth American species have been comparedextensively for neural differences:prairie volesthat are monogamous and montane voles thatfail to form social bonds.

Oxytocin and vasopressin. As oxytocin hasbeen implicated in maternal behaviour in ratsand sheep, it seems plausible that oxytocin orits close relative vasopressin might also beinvolved in adult–adult attachment.Prairievoles normally form pair bonds aftermating53. As copulation (or VCS) releases oxy-tocin and vasopressin54, one possibility is thatthese neuropeptides are involved in theprocess of pair bond formation after mating.Indeed, all of the major behavioural aspects ofmonogamy can be facilitated in the prairie

and act through the main olfactory bulb34.The main event seems to be at the granulecell–mitral cell synapses in the bulb (BOX 1).

The ewe learns the identity of her lambthrough a nitric-oxide-dependent processthat occurs within a few hours of birth,resulting in an increase in the ratio of excita-tory to inhibitory tone in the granulecell–mitral cell synapses in the olfactorybulb48. But this process might not be uniqueto the post-partum ewe, as analogous

lamb odours before birth or in response tounfamiliar lambs post-partum. These changesare reminiscent of the enhancement ofamino-acid release that is observed afterimprinting in chicks.

How does this reorganization of the olfac-tory bulb take place? We know that, as is thecase for olfactory learning in rat pups, affer-ent projections from noradrenaline cells inthe brainstem are critical. We also know thatthe signals from the lamb are largely volatile

PLC

NAcc

PrairieMontanea

PrairieMontane

NAccCombined

CSF

PLC

OTA

CP

Partnerchamber

Neutralchamber

Strangerchamber

0

40

80

120

100

80

60

40

20

0

Tiim

e sp

ent i

n ch

ambe

r (m

ins)

Con

tact

tim

e (m

ins)

b

c

**

Figure 2 | Oxytocin and social attachment in the monogamous prairie vole female. a | Prairie andmontane voles have different distributions of oxytocin receptors in the brain, particularly in the nucleusaccumbens (NAcc) and the prelimbic cortex (PLC; arrow). b | In the laboratory, pair bonding is assayedusing a partner-preference test. After mating or cohabitation with a ‘partner,’ the experimental vole isplace in a three-chambered arena in which the partner is tethered in one chamber (green) and anequivalent non-familiar vole or ‘stranger’ is tethered in another chamber (yellow). The experimental animalhas free access to both chambers. After mating, female prairie voles spend more time in the partner’schamber (green bar) than in the neutral or stranger’s chamber (yellow bar), indicating a partnerpreference. By contrast, mated montane voles show no preference for either the partner or the stranger,indicating the absence of a mating-induced social attachment. c | Partner preference formation isblocked in the mating female prairie vole by infusions of oxytocin receptor antagonist (OTA) into the NAccas well as the PLC, whereas cerebrospinal fluid (CSF) into either area or OTA infused into the caudateputamen (CP) had no effect on partner preference formation. OTA had no effect on mating. (Panel cmodified from REF. 65).



One additional aspect of the vole researchhas investigated the molecular mechanism forthe species differences in the neuroanatomicaldistribution of receptors,potentially a molec-ular basis for monogamy. The coding regionsfor both the oxytocin and vasopressin (V1a)receptors are essentially identical betweenmonogamous and non-monogamous voles.However, there are marked species differencesin the 5′ flanking region of the vasopressin V1areceptor gene with an ~460 bp microsatelliteinsertion into the prairie vole receptor genethat is not evident in the montane vole V1areceptor gene70 (FIG. 3). As promoter sequencesor related cis regulatory regions might beimportant determinants of tissue-specific geneexpression, these microsatellites could con-tribute to the species differences in V1a recep-tor expression. In a transgenic mouse with 2.2kb of the 5′ flanking region (including themicrosatellite) along with the coding regionand 2.4 kb of the 3′ flanking region of theprairie vole V1a receptor, the V1a receptor wasexpressed in a prairie vole-like pattern withinthe CNS70. Supporting the importance ofreceptor location for function, this transgenicmouse responded to arginine vasopressin(AVP) with increased affiliation, similar to the

vole by central injections of either oxytocin orvasopressin, even in voles that do not have theopportunity to mate55,56 (FIG. 2). Conversely,these behaviours are inhibited by either oxy-tocin or vasopressin antagonists given toprairie voles just before mating56,57. Theantagonists do not alter mating behaviour perse,but seem to prevent the partner preferencethat normally occurs with mating in prairievoles (FIGS 2,3). Thus, in monogamous prairievoles, oxytocin and vasopressin seem to benecessary and sufficient for pair bond forma-tion. Neither peptide has notable effects onsocial behaviour in the non-monogamousmontane voles58,59.

Are these pharmacological responses tooxytocin or vasopressin relevant to the behav-ioural differences between the vole species?Although there are no evident species differ-ences in the expression of these neuro-peptides60, there are regional differences in thereceptors for both peptides, assessed by eitherreceptor binding61,62 or receptor mRNA59,63. Inthe monogamous prairie vole, in which oxy-tocin and vasopressin facilitate partner prefer-ence formation, receptors are expressed athigh levels in the nucleus accumbens andrelated regions (for example, oxytocin recep-tors in the prelimbic cortex and vasopressinreceptors in the ventral pallidum) that areassociated with reinforcement and condition-ing (FIGS 2,3). Montane voles have fewdetectable receptors for either oxytocin orvasopressin in these regions, but have highlevels of receptor binding for both neuropep-tides in the lateral septum. In the prairie vole,blockade of oxytocin receptors in the nucleusaccumbens inhibits partner preference for-mation64 (FIG. 2) and viral vector-inducedoverexpression of the vasopressin V1a recep-tor in the ventral pallidum facilitates partnerpreference formation65, suggesting that recep-tors in these regions might be critical for pairbond formation. Indeed, vasopressin recep-tors in the ventral pallidum are present notonly in prairie voles but also in monogamousmice and primates, whereas they are absent inthis region in related rodent and primatespecies that do not form pair bonds66.

A simple model posits the release of oxy-tocin and vasopressin with mating leading tothe activation of reinforcement circuits inmonogamous species that form pair bonds.In non-monogamous species, oxytocin andvasopressin activate unrelated circuits with-out the conditioned response that is essentialfor attachment. Interestingly, there is anincrease in oxytocin receptors in montanevoles post-partum, associated with the onsetof nurturing behaviour towards pups61.Recent studies have described the role of the

nucleus accumbens and specifically D2dopamine receptors in this region for partnerpreference formation in prairie voles67,68. D2agonists facilitate and D2 antagonists inhibitpartner preference formation, whether givensystemically or directly into the nucleusaccumbens. It seems likely that the neuropep-tides (or mating) might be activating amesolimbic circuit implicated in the reinforc-ing effects of psychostimulants. It is not yetclear,however,how the neuropeptides interactwith dopamine or, at the systems level, howeither neuropeptides or dopamine influencepartner preference formation.On the basis ofstudies in rats, Everitt and colleagues haverecently suggested that the effects of dopamineon reinforcement might be mediated byincreasing the gain on glutamate-containingafferents to the nucleus accumbens69. It seemslikely that the neurobiology of partner prefer-ence formation in monogamous species willresemble the neurobiology of other forms ofconditioning, such as place-preference forma-tion in non-monogamous species.However,in monogamous species there might be aselective predisposition to condition to socialstimuli, in part,due to the role of oxytocin orvasopressin neurotransmission.


Pairie

VP

Montane

LS

Prairie

Prairie+ 1623+ 1

Frame shift

Repetitiveexpansion

TruncatedLINE

Montane

Montane

0

40

80

120

Affi

rmat

ive

beha

viou

r (s

ec/5

min

)

CSF

UnmatedMated

AVP CSF V1a antag

30

50

70

90

% ti

me

with

par

tner

CSFAVP

a b

c

Figure 3 | Vasopressin V1a receptor and attachment in prairie vole male. a | Montane and prairievoles have different distributions of V1a receptor binding, with prairie voles having relatively high densities ofreceptors in the ventral pallidum (VP). (LS, lateral septum.) b | The species differences in receptordistribution are probably responsible for the species differences in the behavioural effects of argininevasopressin (AVP) and perhaps in the formation of social attachments. In a test of nonspecific affiliativebehaviour, intracerebroventricular AVP infusions increase social interest in the male prairie vole but not in themontane vole. Furthermore, AVP stimulates the formation of a partner preference in the absence of matingand V1a receptor antagonists prevent partner preference formation after extensive mating bouts. (CSF,cerebrospinal fluid.) c | These species differences in receptor distribution might be the result of speciesdifferences in gene structure. The prairie vole V1a receptor gene has been duplicated with one copy beingdownstream of a retrotransposon element (LINE) and it also has a frameshift mutation in the coding region.Both copies have a large complex repetitive expansion (red) just over 700 bp from the transcription startsite. Either the gene duplication or the promoter expansion could contribute to the evolution of theexpression pattern. (Figure modified with permission from REF. 70 © (1999) Macmillan Magazines Ltd.)




neural mechanisms of attachment behaviourhave been largely conserved. In a generalsense, neuropeptides that are modulators offast neurotransmitters,have discrete CNS dis-tributions, and that are regulated by highlyplastic receptors seem especially suited asmediators of attachment.

Approach behaviour requires overcominga natural avoidance of offspring or strangersalong with the initiation of pro-social, prox-imity-seeking behaviours. Paradoxically,olfactory lesions seem to facilitate the initia-tion of maternal acceptance in rats andsheep, presumably by permitting females toovercome a natural aversion to offspringodours. We know less about the neu-roanatomical circuit for releasing pro-socialbehaviours, but affiliative behaviours inrodents and primates are facilitated by opi-ates76,77 and oxytocin78,79. Receptors for oxy-tocin are increased in the hypothalamus andbed nucleus of the stria terminalis by thephysiological changes in oestrogen and prog-esterone that occur at the end of gestation80,providing a mechanism by which pathwaysfor affiliation can be entrained to parturition.

For species that form selective bonds,learning must be rapid but enduring, analo-gous to single-trial learning. In chick imprint-ing, this process is associated with enhancedrelease of glutamate in select cortical regions.In the case of maternal behaviour in sheep,selectivity is associated with enhanced excita-tory amino-acid release in the olfactory bulb(although also see REF. 81), secondary to cyclicGMP generation within mitral cells. In prairievoles, the circuitry for selectivity is notknown, but on the basis of studies withknockout mice, the medial amygdala mightbe critical for individual recognition. It is notyet clear how this process in any of thesespecies that form selective bonds differs fromthe plasticity in other regions of the CNS thatare associated with other forms of learning.

Finally, the process of investing in a singleattachment partner probably involves thesame pathways that are required for otherforms of motivation.Oxytocin has emergedas one candidate from studies in rats, sheepand monogamous voles, but we do notunderstand fully how this neuropeptide,released during nursing and copulation, fitsinto a systems model for motivation (FIG. 1).Oxytocin is, at most, one element in a cas-cade. Endogenous opioids have been shownto influence affiliative and maternal behav-iours in sheep82 and primates83, possibly bystimulating oxytocin release84 or possibly byan independent effect on reinforcement.Oxytocin modulates release of monoamines,acetylcholine and GABA43. The available data

maternal behaviour, show a profound socialamnesia without other evident cognitivedeficits74. In contrast to the sheep studies ofolfactory memory described above, in theseknockout mice a social stimulus elicits nor-mal amounts of Fos activation in the olfacto-ry bulb,but fails to activate the medial amyg-dala and its projection sites, the bed nucleusof the stria terminalis and the MPOA. In themouse brain, the medial nucleus of theamygdala is enriched with oxytocin receptorsand in the knockout mouse, injections ofoxytocin into this region (but not into theolfactory bulb) restores social recognition75.Therefore, it remains possible that theabsence of partner preferences in prairievoles treated with oxytocin and vasopressinantagonists results from an inability to recog-nize the partner, not from an absence of pairbonding. Wang et al.67 have shown that D2dopamine receptor blockade given justbefore a preference test does not interferewith recognition of the partner but, givenjust after mating, a D2 dopamine antagonistinhibits consolidation of the memory for themate. Analagous studies have yet to be donewith oxytocin or AVP in prairie voles.

Unifying principlesWhat do the studies of infant, maternal andadult attachment have in common? Althoughthe data, so far,have been obtained in differ-ent species, the neurobiological tasks in eachform of attachment are the same: first,approach the parent, infant or partner; sec-ond, learn the identity of this individual; andthird, invest in this individual while rejectingall other individuals. These tasks might beaccomplished by different mechanisms at dif-ferent life stages, contingent on developmen-tal and gonadal status. However, a workinghypothesis is that not only the tasks but the

prairie vole response to AVP. In contrast tothese results with transgenic mice and prairievoles, AVP did not increase affiliation in wild-type mice without the prairie vole V1a recep-tor70. These results are consistent with thehypothesis that the species differences in pro-moter sequence are responsible for the speciesdifferences in receptor distribution,but theydo not rule out other factors, such as environ-mental influences that would be differentbetween the species.

Neuropeptides. The studies with voles providea model by which changes in gene structurecould alter regional receptor expression withprofound effects on the functional response toendogenous or exogenous ligands.Oxytocinand vasopressin are interesting candidatesbecause they could link the neuroendocrineresponse to copulation with the behaviouralconsequence of partner preference formationand ultimately pair bonding. In sheep, severalstudies have begun to define how the ewerecognizes her lamb71. In voles, this has notbeen studied to the same extent. Althoughpheromones are important for reproductivefunction in prairie voles72 , we know very littleabout how the prairie vole learns the identityof its partner.

Furthermore, it remains unclear whetheroxytocin and vasopressin primarily increasethe preference for the partner because theyconfer reinforcing properties onto the mateor whether they simply facilitate recall.Oxytocin has been implicated in the reinforc-ing effects of psychostimulants (cocaine)73,but we know of no evidence that either oxy-tocin or vasopressin is, by itself, reinforcing.Conversely, there is considerable evidencethat implicates both oxytocin and vasopressinin social recognition or social memory.Oxytocin knockout mice, capable of full

Box 2 | Of human bonding

Are animal studies of attachment relevant to human love? In the human brain, oxytocin receptorsare concentrated in several dopamine-rich regions, especially the substantia nigra and globuspallidus, as well as the preoptic area90.Whereas this pattern is consistent with a monogamousbrain, the receptors are not found in the ventral striatum or ventral pallidum, areas in whicheither oxytocin or vasopressin V1a receptors are abundant in monogamous voles and monkeys91.There is no evidence, at this time, that these pathways are involved in human attachment.

A recent functional magnetic resonance imaging (fMRI) study of adults looking at pictures oftheir partners, as opposed to close non-romantic friends, found bilateral activation in theanterior cingulate (Brodmann’s area 24), medial insula (Brodmann’s area 14) as well as caudateand putamen7. The pattern of cortical activation was distinct from previous studies of facerecognition, visual attention, sexual arousal or other emotional states,but resembled preliminaryresults from an fMRI study of new mothers listening to infant cries8. Both studies of humanattachment show marked overlap between the pattern of activation when looking or hearing aloved one and a previous report of activation during cocaine-induced euphoria92. It seems likelythat pathways that mediate the hedonic properties of psychostimulants evolved as neural systemsfor social attachment.



30. Yu, G.-Z., Kaba, H., Okutani, F., Takahashi, S. & Higuchi,T. The olfactory bulb: A critical site of action for oxytocinin the induction of maternal behaviour in the rat.Neuroscience 72, 1083–1088 (1996).

31. Fleming, A. S. & Rosenblatt, J. S. Olfactory regulation ofmaternal behavior in rats: II. Effects of peripherallyinduced anosmia and lesions of the lateral olfactory tractin pup-induced virgins. J. Comp. Physiol. Psychol. 86,233–246 (1974).

32. Landgraf, R., Neumann, I., Russell, J. & Pittman, Q.Push-pull perfusion and microdialysis studies of centraloxytocin and vasopressin release in freely moving ratsduring pregnancy, parturition, and lactation. NY Acad.Sci. 652, 326–339 (1992).

33. Yu, G. -Z. et al. The action of oxytocin originating in thehypothalamic paraventricular nucleus on the mitral andgranule cells in the rat main olfactory bulb. Neuroscience72, 1073–1082 (1996).

34. Levy, F., Kendrick, K., Keverne, E., Porter, R. & Romeyer,A. in Advances in Study of Behavior (eds Rosenblatt, J. &Snowdon, C.) 385–422 (Academic, San Diego, 1996).

35. Keverne, E. B., Levy, F., Poindron, P. & Lindsay, D. R.Vaginal stimulation: An important determinant of maternalbonding in sheep. Science 219, 81–83 (1983).

36. Kendrick, K. M., Levy, F. & Keverne, E. B. Importance ofvaginocervical stimulation for the formation of maternalbonding in primiparous and multiparous parturient ewes.Physiol. Behav. 50, 595–600 (1991).

37. Kendrick, K. M., Keverne, E. B., Chapman, C. & Baldwin,B. A. Microdialysis measurement of oxytocin, aspartate,GABA and glutamate release from the olfactory bulb ofsheep during vaginocervical stimulation. Brain Res. 442,171–177 (1988).

38. Kendrick, K. M., Keverne, E. B. & Baldwin, B. A.Intracerebroventricular oxytocin stimulates maternalbehaviour in the sheep. Neuroendocrinology 46, 56–61(1987).

39. Keverne, E. & Kendrick, K. Oxytocin facilitation ofmaternal behavior in sheep. Ann. NY Acad. Sci. 652,83–101 (1992).

40. Levy, F., Kendrick, K., Keverne, E., Piketty, V. & Poindron,P. Intracerebral oxytocin is important for the onset ofmaternal behavior in inexperienced ewes delivered underperidural anesthesia. Behav. Neurosci. 106, 427–432(1992).

41. Broad, K., Kendrick, K., Sirinathsinghji, D. & Keverne, E.Changes in oxytocin immunoreactivity and mRNAexpression in the sheep brain during pregnancy,parturition and lactation and in response to oestrogenand progesterone. J. Neuroendocrinol. 5, 435–444(1993).

42. Broad, K. et al. Previous maternal experience potentiatesthe effect of parturition on oxytocin receptor mRNAexpression in the paraventricular nucleus. Eur. J.Neurosci. 11, 3725–3737 (1999).

43. Kendrick, K. et al. Neural control of maternal behaviourand olfactory recognition of offspring. Brain Res. Bull. 44,383–395 (1997).

44. Da Costa, A. P. C., Guevara-Guzman, R. G., Ohkura, S.,Goode, J. A. & Kendrick, K. M. The role of oxytocinrelease in the paraventricular nucleus in the control ofmaternal behaviour in the sheep. J. Neuroendocrinol. 8,163–177 (1996).

45. Levy, F., Kendrick, K. M., Goode, J. A., Guevara-Guzman, R. & Keverne, E. B. Oxytocin and vasopressinrelease in the olfactory bulb of parturient ewes: changeswith maternal experience and effects on acetylcholine, γ-aminobutyric acid, glutamate and noradrenalinerelease. Brain Res. 669, 197–206 (1995).

46. Eichenbaum, H. A cortical–hippocampal system fordeclarative memory. Nature Rev. Neurosci. 1, 41–50(2000).

47. Kendrick, K., Levy, F. & Keverne, E. Changes in thesensory processing of olfactory signals induced by birthin sheep. Science 256, 833–836 (1997).

48. Kendrick, K. et al. Formation of olfactory memoriesmediated by nitric oxide. Nature 388, 670–674 (1997).

49. Brennnan, P., Schellinck, H., De La Riva, C., Kendrick, K.& Keverne, E. Changes in neurotransmitter release in themain olfactory bulb following an olfactory conditioningprocedure. Neuroscience 87, 583–590 (1998).

50. Guevara-Guzman, R., Barrera-Mera, B., De La Riva, C. &Kendrick, K. Release of classical neurotransmitters andnitric oxide in the rat olfactory bulb, evoked byvaginocervical stimulation and potassium, varies with theoestrus cycle. Eur. J. Neurosci. 12, 80–88 (2000).

51. Kleiman, D. G. Monogamy in mammals. Q. Rev. Biol. 52,39–69 (1977).

52. Dewsbury, D. A. in American Zoology NebraskaSymposium on Motivation (ed. Leger, D. W.) 1–50(Nebraska Univ. Press, Lincoln, Nebraska, 1988).

might indicate an important link withmesolimbic dopamine pathways as oxytocineffects on rat maternal care might be mediat-ed through the VTA85 and several studies haveshown that mesolimbic dopamine pathwaysinfluence motivational aspects of maternalbehaviour86–88. As noted above,only monoga-mous voles that pair bond after mating havereceptors for oxytocin in the nucleus accum-bens and vasopressin in the ventral pallidum.But, there are still unanswered questions con-cerning whether these neuropeptides are spe-cific for social stimuli or how they interactwith other forms of reinforcement that aremediated through mesolimbic pathways. Theconnection between attachment and addic-tion has been suggested phenomenologicallyand might provide important clues for futureresearch89.

SummaryAttachment behaviour is both biologicallyimportant and technically difficult to study.The behaviour is complex and there arechanges in several cognitive and affective vari-ables to consider. Nevertheless, recent studieswith chicks, rats, sheep, voles and nowhumans (BOX 2) have begun to reveal someimportant candidates for the neurobiology ofsocial attachment. The neuropeptides oxy-tocin and vasopressin have yielded a modelthat links molecular, cellular and systemsapproaches. Dopamine pathways in the fore-brain, especially the nucleus accumbens andventral pallidum, seem to be important forcertain aspects of partner preference forma-tion. It seems likely that for attachment tooccur, these neuropeptides must link socialstimuli to dopamine pathways associated withreinforcement. It is also possible that neuralmechanisms that we associate with drugabuse and addiction might have evolved forsocial recognition, reward and euphoria —critical elements in the process of attachment.In the very near future, we can hope that dis-coveries of the molecular and cellular mecha-nisms of addiction might be applied to theneurobiology of attachment,providing a newunderstanding of one of our most complexand intriguing emotions.

Center for Behavioral Neuroscience,954 Gatewood Road Northeast, Emory

University, Atlanta, Georgia 30329, USA.Correspondence to T.R.I.

e-mail: [email protected]

1. Harlow, H. & Suomi, S. The nature of love — Simplified.Amer. Psychol. 25, 161–168 (1970).

2. de Bono, M. & Bargmann, C. Natural variation in aneuropeptide Y receptor homolog modifies socialbehavior and food response in C. elegans. Cell 94,679–689 (1998).

3. Bowlby, J. Attachment and Loss: Attachment 2nd ednVol. 1 (Basic Books, New York, 1982).

4. Snowdon, C. in Contemporary Issues in ComparativePsychology (ed. Dewsbury, D.) 225–251 (Sinauer,Sunderland, Massachusetts, 1990).

5. Mendoza, S. & Mason, W. Contrasting responses tointruders and involuntary separation by monogamousand polygynous New World monkeys. Physiol. Behav.38, 795–801 (1986).

6. Martel, F. L., Nevison, C. M., Rayment, F. D., Simpson, M. J. & Keverne, E. B. Opioid receptor blockade reducesmaternal affect and social grooming in rhesus monkeys.Psychoneuroendocrinology 18, 307–321 (1993).

7. Bartels, A. & Zeki, S. The neural basis of romantic love.NeuroReport 11, 3829–3834 (2000).

8. Lorberbaum, J. et al. Feasibility of using fMRI to studymothers responding to infant cries. Depression andAnxiety 10, 99–104 (1999).

9. Bolhuis, J. & Honey, R. Imprinting, learning anddevelopment: from behaviour to brain and back. TrendsNeurosci. 21, 306–311 (1998).

10. Horn, G. Visual imprinting and the neural mechanisms ofrecognition memory. Trends Neurosci. 21, 300–305(1998).

11. Metzger, M., Jiang, S. & Braun, K. Organization of thedorsalcaudal neostriatal complex: a retrograde andanterograde tracing study in the domestic chick withspecial emphasis on pathways relevant to imprinting. J. Comp. Neurol. 395, 380–404 (1998).

12. Bredenkotter, M. & Braun, K. Changes of neuronalresponsiveness in the mediorostral neostriatum/hyperstriatum after auditory filial imprinting in thedomestic chick. Neuroscience 76, 355–365 (1997).

13. Gruss, M. & Braun, K. Stimulus evoked increase ofglutamate in the mediorostral neostriatum/hyperstriatumventrale of domestic chicks after filial imprinting: an in vivo microdialysis study. J. Neurochem. 66, 1167–1173(1996).

14. Winslow, J. T. & Insel, T. R. The infant rat separationparadigm: A novel test for novel anxiolytics. TrendsPharmacol. Sci. 12, 402–404 (1991).

15. Sullivan, R. & Wilson, D. The locus ceruleus,norepinephrine, and memory in newborns. Br. Res. Bull.35, 467–472 (1994).

16. Johnson, B. & Leon, M. Spatial distribution of [14C]2-deoxyglucose uptake in the glomerular layer of the ratolfactory bulb following early odor preference learning. J. Comp. Neurol. 376, 557–566 (1996).

17. Nelson, E. & Panksepp, J. Oxytocin and infant-motherbonding in rats. Behav. Neurosci. 110, 583–592 (1996).

18. van Praag, H., Kemperman, G. & Gage, F. Neuralconsequences of environmental enrichment. Nature Rev.Neurosci. 1, 191–198 (2000).

19. Rosenblatt, J. S. & Siegel, H. I. in Parental Care inMammals (eds Gubernick, D. J. & Klopfer, P. H.) 1–76(Plenum, New York, 1981).

20. Numan, M. in The Physiology of Reproduction(eds Knobil, E. & Neill, J.) 221–302 (Raven, New York,1994).

21. Bridges, R. S. A quantitative analysis of the roles ofdosage, sequence, and duration of estradiol andprogesterone exposure in the regulation of maternalbehavior in the rat. Endocrinology 114, 930–940 (1984).

22. Corodimas, K. P., Rosenblatt, J. S. & Morrell, J. I.Thehabenular complex mediates hormonal stimulation ofmaternal behavior in rats. Behav. Neurosci. 106,853–865 (1992).

23. Numan, M. & Sheehan, T. P. Neuroanatomical circuitry formammalian maternal behavior. Ann. NY Acad. Sci. 807,101–125 (1997).

24. Numan, M., Corodimas, K. P., Numan, M. J., Factor, E. M. & Piers, W. D. Axon-sparing lesions of the preopticarea and substantia innominata disrupt maternal behaviorin rats. Behav. Neurosci. 102, 381–395 (1988).

25. Bridges, R. S. & Ronsheim, P. M. Prolactin (PRL)regulation of maternal behavior in rats: Bromocriptinetreatment delays and PRL promotes the rapid onset ofbehavior. Endocrinology 126, 837–848 (1990).

26. Lucas, B., Ormandy, C., Binart, N., Bridges, R. & Kelly, P.Null mutation of the prolactin receptor gene produces adefect in maternal behavior. Endocrinology 139,4102–4107 (1998).

27. Pedersen, C. A., Ascher, J. A., Monroe, Y. L. & Prange, A.J. Jr Oxytocin induces maternal behaviour in virgin femalerats. Science 216, 648–649 (1982).

28. Insel, T. R. in Mammalian Parenting, Biochemical,Neurobiological, and Behavioral Determinants (edsKrasnegor, N. & Bridges, R.) 260–280 (Oxford Univ.Press, New York, 1990).

29. Pedersen, C. A., Caldwell, J. O., Walker, C., Ayers, G. &Mason, G. A. Oxytocin activates the postpartum onset ofrat maternal behavior in the ventral tegmental and medialpreoptic areas. Behav. Neurosci. 108, 1163–1171 (1994).





86. Giordano, A., Johnson, A. & Rosenblatt, J. Haloperidol-induced disruption of retrieval behavior and reversal withapomorphine in lactating rats. Physiol. Behav. 48,211–214 (1990).

87. Hansen, S., Harthon, C., Wallin, E., Lofberg, I. &Svensson, K. Mesotelencephalic dopamine system andand reproductive behavior in the female rat: Effects ofventral tegmental 6-OHDA lesions on maternal and sexualresponsiveness. Behav. Neurosci. 105, 588–598 (1991).

88. Keer, S. & Stern, J. Dopamine receptor blockade in thenucleus accumbens inhibits maternal retrieval andlicking, but enhances nursing behavior in lactating rats.Physiol. Behav. 67, 659–669 (1999).

89. Panksepp, J. Affective Neuroscience (Oxford Univ.Press, New York, 1998).

90. Loup, F., Tribollet, E., Dubois-Dauphin, M. & Dreifuss, J. J. Localization of high-affinity binding sites for oxytocinand vasopressin in the human brain. Anautoradiographic study. Brain Res. 555, 220–232 (1991).

91. Wang, Z. et al. Vasopressin in the forebrain of commonmarmosets (Callithrix jacchus): Studies with in situhybridization, immunocytochemistry and receptorautoradiography. Brain Res. 768, 147–156 (1997).

92. Breiter, H. et al. Acute effects of cocaine on human brainactivity and emotion. Neuron 19, 591–611 (1997).

AcknowledgementsThe authors are supported by NSF and NIMH. The authorsthank M. Davis for his comments on an early draft of the manu-script and Z. Wang for his contributions to many of the ideas andresults described.

77. Martel, F. L., Nevison, C. M., Simpson, M. J. & Keverne,E. B. Effects of opioid receptor blockade on the socialbehavior of rhesus monkeys living in large family groups.Dev. Psychobiol. 28, 71–84 (1995).

78. Insel, T. R. Oxytocin: a neuropeptide for affiliation —evidence from behavioral, receptor autoradiographic, andcomparative studies. Psychoneuroendocrinology 17,3–33 (1992).

79. Holman, S. D. & Goy, R. W. in Motherhood in Human andNon-Human Primates (eds Pryce, C. R. & Martin, R. D.)87–93 (Karger, Basel, 1994).

80. Insel, T. R. Regional changes in brain oxytocin receptorpost-partum: Time-course and relationship to maternalbehavior. J. Neuroendocrinol. 2, 539–545 (1990).

81. Ferreira, G. et al. Learning of olfactory cues is notnecessary for early lamb recognition by the mother. Phys.Behav. 69, 405–412 (2000).

82. Kendrick, K. Oxytocin, motherhood, and bonding. Exp.Physiol. 85, S111–S124 (2000).

83. Keverne, E. B., Martensz, N. & Tuite, B. B-endorphinconcentrations in CSF of monkeys are influenced bygrooming relationships. Psychoneuroendocrinology 14,155–161 (1989).

84. Bicknell, R. J. & Leng, G. Endogenous opiates regulateoxytocin but not vasopressin secretion from theneurohypophysis. Nature 298, 161–162 (1982).

85. Pedersen, C. A., Caldwell, J. D., Walker, C., Ayers, G. &Mason, G. A. Oxytocin activates the postpartum onset of rat maternal behavior in the ventral tegmental andmedial preoptic area. Behav. Neurosci. 108, 1163–1171(1994).

53. Carter, C., DeVries, A. & Getz, L. Physiological substratesof mammalian monogamy: The prairie vole model.Neurosci. Biobehav. Rev. 19, 303–314 (1995).

54. Witt, D. M. Oxytocin and rodent sociosexual responses:from behavior to gene expression. Neurosci. Biobehav.Rev. 19, 315–324 (1995).

55. Williams, J., Insel, T., Harbaugh, C. & Carter, C. Oxytocinadministered centrally facilitates formation of a partnerpreference in female prairie voles (Microtus ochrogaster).J. Neuroendocrinol. 6, 247–250 (1994).

56. Winslow, J. T., Hastings, N., Carter, C. S., Harbaugh, C. R. & Insel, T. R. A role for central vasopressin in pairbonding in monogamous prairie voles. Nature 365,545–548 (1993).

57. Insel, T. R. & Hulihan, T. J. A gender specific mechanismfor pair bonding: oxytocin and partner preferenceformation in monogamous voles. Behav. Neurosci. 109,782–789 (1995).

58. Winslow, J. T., Shapiro, L. E., Carter, C. S. & Insel, T. R.Oxytocin and complex social behaviors: speciescomparisons. Psychopharmacol. Bull. 29, 409–414(1993).

59. Young, L. J., Winslow, J. T., Nilsen, R. & Insel, T. R.Species differences in V1a receptor gene expression in monogamous and nonmonogamous voles:Behavioral consequences. Behav. Neurosci. 111,599–605 (1997).

60. Wang, Z. X., Zhou, L., Hulihan, T. J. & Insel, T. R.Immunoreactivity of central vasopressin and oxytocin pathways in microtine rodents: A quantitativecomparative study. J. Comp. Neurol. 366, 726–737(1996).

61. Insel, T. R. & Shapiro, L. E. Oxytocin receptor distributionreflects social organization in monogamous andpolygamous voles. Proc. Natl Acad. Sci. USA 89,5981–5985 (1992).

62. Insel, T. R., Wang, Z. & Ferris, C. F. Patterns of brainvasopressin receptor distribution associated with socialorganization in microtine rodents. J. Neurosci. 14,5381–5392 (1994).

63. Young, L. J., Huot, B., Nilsen, R., Wang, Z. & Insel, T. R.Species differences in central oxytocin receptor geneexpression: Comparative analysis of promotersequences. J. Neuroendocrinol. 8, 777–783 (1996).

64. Young, L. J., Gingrich, B. & Insel, T. R. Molecular andcellular mechanisms of monogamy. Horm. Behav. (in thepress).

65. Sharer, C. et al. Manipulation of vasopressin receptorexpression in adult brain using an AAV viral vector. Soc.Neurosci. Annu. Meeting 30, 373.11 (2000).

66. Young, L. J. Oxytocin and vasopressin receptors andspecies-typical social behaviors. Horm. Behav. 36,212–221 (1999).

67. Wang, Z. et al. Dopamine D2 receptor-mediatedregulation of partner preferences in female prairie voles: A mechanism for pair bonding. Behav. Neurosci. 113,602–611 (1999).

68. Gingrich, B., Liu, Y., Cascio, C., Wang, Z. & Insel, T. R.Dopamine D2 receptors in the nucleus accumbens areimportant for social attachment in female prairie voles.Behav. Neurosci. 114, 173–183 (2000).

69. Everitt, B. et al. Associative process in addiction andreward. Ann. NY Acad Sci. 877, 412–438 (1998).

70. Young, L., Nilsen, R., Waymire, K., MacGregor, G. &Insel, T. Increased affiliative response to vasopressin inmice expressing the V1a receptor from a monogamousvole. Nature 400, 766–768 (1999).

71. Levy, F., Kendrick, K. M., Keverne, E. B., Porter, R. H. &Romeyer, A. in Parental Care: Evolution, Mechanisms,and Adaptive Significance (eds Rosenblatt, J. S. &Snowdon, C. T.) 385–416 (Academic, San Diego, 1996).

72. Carter, C. S., Witt, D. M., Schneider, J., Harris, L. &Volkening, D. Male stimuli are necessary for female sexual behavior and uterine growth in prairie voles (Microtus ochrogaster). Horm. Behav. 21, 74–82(1987).

73. Kovacs, G. L. & Telegdy, G. in Oxytocin: Clinical andLaboratory Studies (eds Amico, J. & Robinson, A. G.)359–371 (Elsevier, New York, 1985).

74. Ferguson, J., Young, L., Hearn, E., Insel, T. & Winslow, J.Social amnesia in mice lacking the oxytocin gene. NatureGenet. 25, 284–288 (2000).

75. Ferguson, J., Aldag, J., Winslow, J., Insel, T. R. & Young, L. J. Neural activation in the socially amnesticmouse. Soc. Neurosci. Annu. Meeting 30, 373.8 (2000).

76. Herman, B. H. & Panksepp, J. Effects of morphine andnaloxone on separation distress and approachattachment: Evidence for opiate mediation of social affect. Pharmacol. Biochem. Behav. 9, 213–220(1978).

cognitive abilities and disabilities it is surpris-ing that textbooks in cognitive neuroscienceseldom mention individual differences1,2.

The very standard deviationBecause the species-universal and the individ-ual-variability perspectives ask different ques-tions they can arrive at different answers. Inmost species-universal research, the researchermanipulates something — creates lesions,administers drugs or sets specific tasks — andexamines the average effect of the manipula-tion on the study population. By contrast,rather than creating variance through manip-ulations, the individual-variability perspectivefocuses on the naturally occurring differencesbetween individuals that would be considerednoise in a species-universal approach.Research on individual differences is thereforecorrelational; it investigates factors that dohave an effect in the world outside the labora-tory. One of these factors is genetic back-ground. Indeed, although 99.9% of thehuman DNA is identical for all humanbeings, the 0.1% that differs — three millionbase pairs — is ultimately responsible for theubiquitous genetic influence found for allindividual traits, including cognitive abilitiesand disabilities3.

The g factor refers to the substantial overlapthat exists between individual differences indiverse cognitive processes in humans. Inthis article, I argue that a mouse model of gcould provide a powerful analytic tool forexploring cognitive processes that are linkedfunctionally by genes.

Despite its name, analysis of variance — themost widely used statistical tool in biomedicalscience — is actually an analysis of meaneffects in which individual differences are liter-ally called the error term. This error term, thevery standard deviation, is the topic of thisessay. Instead of treating differences betweenindividuals as error or noise, and averagingindividual data across groups as commonlyoccurs, I believe that the analysis of the indi-vidual differences themselves might providenew insight into how the brain works. Theanalysis of means and variances representdifferent perspectives. Neither perspective isright or wrong, just more or less useful for aparticular purpose.Means lead us into think-ing about universal or at least species-widephenomena, whereas variances lead towardsinter-individual differences within species —variations on species themes. Given the im-portance of individual differences for human

The genetics of g in human and mouse

Robert Plomin

OP I N ION


pe rs pect ive s - university of...

Documents