Neural, cognitive, and evolutionary foundations of human altruism

13 Pages·2015·0.15 MB·English

AdvancedReview
Neural, cognitive, and
evolutionary foundations
of human altruism
*
Abigail A. Marsh
This article considers three forms of altruism from both a psychological and a
neural perspective, with an emphasis on homologies that can be observed across
species and potentially illuminate altruism’s evolutionary origins. Kin-based
altruism benefits biological relatives and, according to the theory of inclusive fit-
ness, is ultimately beneficial to the altruist from a genetic standpoint. Kin selec-
tion adequately explains some altruistic behavior, but it is not applicable to
muchhumanaltruism.Little isknownabouttheneuralprocessesthatsupportit,
but they may include cortical regions involved in processing autobiographical
memory and the identities of familiar others. Reciprocity-based altruism is per-
formed in expectation of future rewards and is supported by dopaminergic
cortico-striatal networks that guide behavior according to anticipated rewards.
Care-based altruism is aimed at improving the well-being of distressed and
vulnerable individuals and is closely linked to empathic concern. This form
of altruism is thought to rely on the subcortical neural systems that support
parental care, particularly structures densely populated with receptors for the
hormones oxytocin and vasopressin, including the amygdala, stria terminalis,
and striatum. The amygdala may be a particularly important convergence point
for care-based altruism because of its dual role in responding both to cues that
signal infantile vulnerability and those that signal distress. Research on altruism
continues to converge across disciplines, but more research linking molecular-
level neural processes to altruistic behavior in humans and other species
is needed, as is research on how various forms of altruism intersect. © 2015Wiley
Periodicals,Inc.
Howtocitethisarticle:
WIREsCogn Sci2016,7:59–71.doi: 10.1002/wcs.1377
INTRODUCTION in scope and specifics, even seemingly divergent beha-
viors performed for the good of others maybe unified
Altruism is a behavior that improves the welfare
bycommonpsychologicalandneuralmechanisms.
ofanotherindividualattheexpenseofthealtru-
Research on altruism has historically proceeded
ist.1,2Thisdefinitioncapturesawidevarietyofhuman along separate lines within distinct academic fields,
and nonhuman behaviors, from a heroic rescuer sav- including biology, psychology, and economics.3
ing a drowning child to a mother rat who sacrifices Increasinglyinterdisciplinaryworkacrossthesefields,
her own caloric resources to nurse her young. however, has led to a more unified understanding of
Although these behaviors may appear wildly different
the processes that support altruism. In particular,
attempts to reconcile psychological and biological
*Correspondenceto:[email protected] explanations for altruism have helped to connect the
ultimate biological forces that drive the emergence of
Department of Psychology, Georgetown University, Washington,
DC,USA altruistic behavior in a species and the proximate
psychological forces that motivate altruistic behavior
Conflictofinterest:Theauthorhasdeclarednoconflictsofinterest
forthisarticle. inanorganism(Box1).
Volume 7, January/February 2016 ©2015Wiley Periodicals, Inc. 59
AdvancedReview wires.wiley.com/cogsci
subsequent generations. The types of behaviors that
BOX1
most obviously promote survival and reproduction,
such as acquiring resources and fending off threats,
ISITREALLYALTRUISM? tend to be selfish.8 Altruistic behavior by definition
disadvantages altruists, who in sacrificing their own
Altruism remains a highly contentious psycho-
self-interest reduce their fitness relative to those who
logical phenomenon. It is common for people
actselfishly.
to argue that all altruistic behavior is actually
selfish, in part due to two persistent misunder- Nevertheless, an impressive array of altruistic
behaviors is found among social species.1,9 Ground
standings: about ultimate versus proximate
causes4 and about intended versus foreseen squirrels shriek alarm calls to their colony-mates
effects.5 Some who argue that all altruism is upon spotting a predator, thereby improving the
selfish believe this to be the case because the colony-mates’ odds of escaping while at the same
altruist’s genes may benefit from the behavior, time drawing the predator’s attention to them-
asinthecaseofkinselectionorcaringforone’s selves.10,11Vampirebatsregurgitatebloodforcolony-
offspring. All evolved behaviors must be mates who might otherwise risk starvation.12 Male
assumed to yield evolutionary benefits or they turkeys support other males’ courtship displays to
would not exist, and the same must be true for improve those males’ odds of successfully reprodu-
altruism. But genetic selfishness should not be cing.13 And of course humans display a wide variety
confused with psychological selfishness. An of altruistic behaviors.14–16 These behaviors, by
organism that acts intentionally to sacrifice its which an individual risks his or her own energy,
own fitness for another is altruistic, regardless safety, or fitness to improve other individuals’ wel-
of potential genetic advantages. This is because fareseemcontradictorytobasicbiologicalprinciples.
any behavior can simultaneously be driven by
ultimate (genetic) and proximate (psychologi-
cal) causes.4 Another argument is that inten-
tionally altruistic acts are really selfish because KIN SELECTION
the altruist ultimately derives satisfaction from
Two models of altruistic behavior emerged to resolve
them.Thus, the altruist somehow benefits from
this seeming conundrum in the mid-twentieth cen-
the act; but this argument confuses foreseen
tury. The first is kin selection,17,18 which hinges on
outcomes with intended outcomes.5 A person
the concept of inclusive fitness. Inclusive fitness dic-
who gives change to a needy stranger or res-
cues a drowning child undoubtedly finds the tates that an altruistic behavior can evolve ifits bene-
ficiaries are closely related enough to the altruist to
behavior rewarding, just as successfully per-
compensate the altruist for the risk he or she engen-
forming any goal-directed behavior is reward-
ing. The fact that hedonic reward can be ders. So an altruistic act’s benefit to a full sibling
foreseen doesnot mean that achieving hedonic must be more than twice as great as the risk it poses
rewardwasthegoalofanaction.Theresultsof to the altruist. If the beneficiary is a half sibling, the
empirical studies generally contradict the idea benefit must be more than four times as great as the
that hedonic reward is the proximate goal of risk.19 This model most clearly explains altruistic
altruisticbehavior.6 behaviors like provisioning and defense of the colony
among eusocial species such as honeybees and mole
rats,amongwhichmembersofacolonyaretheprog-
But altruism remains one of the central myster- eny of a single queen and so share strong genetic
ies of the biological and social sciences. As Darwin7 overlap.20,21
commented,‘Itisextremelydoubtfulwhethertheoff- What appears to be preferential altruism
spring of the more sympathetic and benevolent par- towardkinisalsocommonlyobservedamongspecies
ents, orofthose whichwere themost faithful totheir that are not eusocial, including humans.10,11,13,22 In
comrades, would be reared in greater number than the laboratory, human participants report a prefer-
the children of selfish and treacherous parents of the ence for helping kin over non-kin in response to
same tribe. He who was ready to sacrifice his life, as hypothetical scenarios, especially those that involve
many a savage has been, rather than betray his com- life-or-death situations.23 Real-world life-or-death
rades, would often leave no offspring to inherit his situations may also favor altruism toward kin. Living
noble nature’ (p. 157). Organisms whose behavior organ donation requires donors to undergo signifi-
promotes their successful survival and reproduction cant inconvenience, discomfort, and a small risk of
pass along their genes in greater proportions to serious injury or death to provide a life-saving organ
60 ©2015Wiley Periodicals, Inc. Volume 7, January/February 2016
WIREsCognitiveScience Neural,cognitive,andevolutionaryfoundationsofhumanaltruism
like a kidney to another person.24 Perhaps unsurpris- likely to perform costly or altruistic behaviors for
ingly, the vast majority of living kidney donations others who have helped them previously, or who
(as many as 65%) are made to biologically related they anticipate will help them in the future. As a
individuals, with far fewer donations made to indivi- result, reciprocal altruism is most commonly
dualswhoarenotbiologicallyrelated.25Humans’and observed in closed systems of individuals who
other species’ tendency to exhibit preferential treat- encounter one another frequently or comprise a
menttowardindividualswhoaresimilartothemselves mutually dependent social group.12 This form of
may also reflect preferential altruism toward indivi- altruism is pervasive among human societies, and
dualswithstrongergeneticoverlap.18,26–29 may have been essential to the survival of early
Proximally, kin selection must rely on special- hunter-gatherers.35
ized mechanisms for identifying and detecting biolog-
ical kin.22 These mechanisms may include those
required to evaluate whether a target was observed Reward Expectancy
inassociationwithone’sownmotherduringtheperi- Neurocognitively, reciprocal altruism appears to be
natal period, or whether the target was frequently driven by reward expectancy, which is mediated in
present when the perceiver was receiving parental part by activity in the striatum and ventromedial pre-
care, both of which are markers of siblinghood. Both frontal cortex.36 In the ventral striatum, the release
variables predict self-reported altruistic behaviors of the neurotransmitter dopamine corresponds to
toward siblings, including the number of favors pre- enhanced expectation of reward.37 Information
viously performed for the sibling and willingness to about rewards expected and received is communi-
donate a kidney to the sibling in the future.22 At the cated between the striatum and ventromedial pre-
neural level, the mechanisms that may be involved in frontal cortex to guide decision-making.38–40
identifying kin are not well understood, but coordi- Changesinrewardexpectancyarisingfromthisproc-
nated activity among cortical regions involved in ess may in some cases be conscious and explicit, but
social recognition (such as the fusiform face area) inmanycases(perhapsmost)theyarenot.
and multi-modal autobiographical memory (such as The role of these regions has been demon-
theprecuneus)maybeinvolved.30,31 stratedinneuroimagingstudiesusinganiteratedPris-
Kin selection is in some ways a problematic oner’s Dilemma paradigm to model reciprocal
model for describing human altruism, how- altruism.41,42 In this paradigm, players maximize
ever.18,19,32 On the one hand, a large proportion of their own gain within a round if they defect on their
human helping behavior across cultures is clearly partner. But because defecting is typically met with
aimed at benefiting close genetic relatives, including retaliation, cooperation improves players’ outcomes
offspring, siblings, and parents. On the other hand, it across multiple rounds by increasing the odds that
is difficult to calculate whether such behaviors yield their partner will cooperate in the next round.
theadaptiveoutcomes required for truekin selection. Mutual cooperation leads to the highest payoffs over
This is in part because humans, unlike eusocial spe- multiple iterations of the game.43 The paradigm
cies, are much less strongly genetically related to models reciprocal altruism because players achieve
other members of their social groups, and they do the greatest mutual long-term gains by sacrificing
not relinquish reproduction to a single colony mem- their own short-term gains in any single round. In
ber. And of course their behavior is more flexible, keeping with the striatum’s role in tracking expected
cognitively complex, and influenced by culture.16 As reward, activation in this region is lowest during
a result, it is often difficult to disentangle the various rounds in which a player defects. This is despite the
ultimate and proximate motives that drive helping fact that defecting results in higher gains within that
behaviors towardkin,whichcaninclude habit,social round. The ventral striatum is more active when the
learning, conformity to cultural norms, and indirect playercooperates(eveniftheplayer’spartner defects).
self-benefit.33 In addition, kin selection theory does This is consistent with the idea that activation in this
not explain altruism that benefits distantly related or region does not correspond to rewards already
unrelatedindividuals,asmuchhumanaltruismdoes. received, because cooperating when one’s partner
defects yields the lowest payoff of all. This region is
preferentially active following cooperation because
RECIPROCAL ALTRUISM
cooperatingtypicallyresultsinhigherpayoffsinfuture
An alternate model for explaining altruistic behavior rounds. In addition, the strength of striatal activation
toward non-kin is reciprocal altruism.34 The premise in any given round predicts the likelihood that the
of reciprocal altruism is that individuals will be more player will continue cooperating in future rounds,
Volume 7, January/February 2016 ©2015Wiley Periodicals, Inc. 61
AdvancedReview wires.wiley.com/cogsci
suggestingthatactivationinthisregionpositivelyrein- bear altricial, or highly dependent, young. The sur-
forces cooperation. This may be because it increases vival of mammalian offspring requires that their
the immediate reward value of mutual cooperation or parents (usually mothers) provide them with
becauseitprovidesalearningsignal.41,44 constant care and remain vigilant to cues that their
well-being may be threatened.52 The emergence some
Serotonin
220 million years ago of the earliest mammals who
These processes may be modulated by the neuro- were motivated to provide their offspring with such
transmitter serotonin, low levels of which decrease care has been described as a ‘star hour,’ or a trans-
the value of delayed rewards relative to immediate formational moment in the evolution of vertebrate
rewards.45 As a result, pharmacologically lowering behavior, because it enabled the possibility of care-
serotonin levels decreases cooperation in the Prison- basedaltruism.53
er’s Dilemma, whereas elevating serotonin levels Parental care is such an ordinary phenomenon
increases cooperation.46,47 The effects of serotonin that we often fail to think of it as altruism.54 But it
on reciprocal altruism may be mediated by the ven- clearly meets the definition, which is a behavior that
tromedial prefrontal cortex, a region in which activ- improves the welfare of another individual at the
ity is modulated by serotonin.48 Depleting serotonin expense of the altruist. Parental care may involve,
can result in similar patterns of neurocognitive dependingonthespecies,providingyoungwithfood,
impairment, like increased impulsiveness, as are cleaning them, warming them, retrieving them to the
observed in people with ventromedial prefrontal nest, and defending them from predators.2 These
lesions.49,50 Serotonergic modulation of activity in behaviors often require significant sacrifice to par-
the ventromedial prefrontal cortex may promote ents’ own safety, resources, and reproductive oppor-
reciprocal altruism by increasing the subjective value tunities. Caring for offspring genetically benefits
of the long-term rewards that result from mutual parents, of course. But among social mammals,
cooperation.36,47 Together, these findings reinforce parental care is often and enthusiastically extended
the view that reciprocal altruism is a fundamentally to distantly related or unrelated infants as well, a
utilitarian phenomenon that requires the ability to behaviorknownasalloparenting.6
use—either explicitly or implicitly—representations
of delayed rewards to inhibit the tendency to acquire
Alloparenting
smallershort-termrewards.
Alloparenting is observed in species ranging from
Striatally mediated reward-expectancy seems
rodents to primates to humans.52 Rats, for example,
unlikelytoaccountforallinstancesofaltruism, how-
will exhibit equally intense devotion to and defense
ever, particularly altruism for which reciprocity is
of their own pups and unrelated pups in their nest.55
not expected. Human altruism often takes the form
Mother deer respond with comparable behaviors to
of care or protection provided in response to the dis-
the distress vocalizations not only of unrelated infant
tress of vulnerable individuals. Examples range from
deer, but infant marmots, seals, and even humans.56
sharing low-cost resources with strangers in need,
Over a century ago, McDougall57 suggested that
such as giving directions or spare change, to lifesav-
altruistic responses to distress in one’s own offspring
ing heroic rescues, such as the famous case of the
‘Subway Superman’ Wesley Autrey, who saved a might also emerge (in somewhat weaker form) in
response to distress in other’s offspring and even
young man named Cameron Hollopeter from certain
death on the New York City Subway tracks.51 Altru- other adults, providing a basis for altruism more
generally:
ism in response to vulnerability or distress in stran-
gers is notutilitarian, and isunlikely tobeperformed
in anticipation of reciprocity, as those who are very Tender emotion and the protective impulse are, no
vulnerable are generally not capable of performing doubt, evoked more readily and intensely by one’s
altruistic acts in return.34 Instead, altruistic care in ownoffspring,becauseaboutthemastronglyorgan-
response to the distress of vulnerable individuals is ized and complex sentiment grows up. But the dis-
thought to emerge from systems that evolved to sup- tress of any child will evoke this response in a very
portparentalcare.2,16 intense degree in those in whom the instinct is
strong…Byafurtherextensionof thesamekindthe
emotion may be evoked by the sight of any very
CARE-BASED ALTRUISM young animal, especially if in distress… In a similar
direct fashion the distress of any adult (towards
Intensive care of vulnerable offspring is an whom we harbour no hostile sentiment) evokes the
essential behavior in mammals, which typically emotion…(p.61,63).
62 ©2015Wiley Periodicals, Inc. Volume 7, January/February 2016
WIREsCognitiveScience Neural,cognitive,andevolutionaryfoundationsofhumanaltruism
Similar themes have been elaborated more as looking babyish, round, submissive, and weak.
recently by many investigators.2,16,52,58–61 Why This is true even when digital manipulations prevent
would care be extended beyond direct offspring? the expressions from being explicitly recognized as
Care of vulnerable offspring may represent such a conveying fear.70 The babyish appearance of fearful
powerful impulse, and one that is so essential to the expressions—with their widened eyes, raised brows,
survival of social mammals, as to generate altruistic and a flattened brow ridge—may stimulate associa-
responding to anything resembling an infant in dis- tions between individuals expressing fear and human
tress. Supporting this link, the degree of alloparental infants.59,70 As a result, fearful expressions may elicit
care provided by adults of a given primate species is parental responses via similar mechanisms as do
the single best predictor of altruistic behavior actualinfants.
betweenadultsofthatspecies.6
Like many other primates, humans have lived
in small, mutually interdependent groups for most of Empathic Concern
their evolutionary existence.52 Within these groups, a Proximally, the mental state that promotes altruistic
tendency for adults to provide nurturant care for the care in response to vulnerability and distress is
offspring of siblings, the offspring of unrelated group empathic concern, sometimes called simply ‘empathy’
members, and even needy or vulnerable adult group or, alternately, sympathy, concern, caring, or com-
members may have increased the fitness of all group passion.1,16 Empathic concern is typically defined as
members.62 This pattern is consistent with multilevel the other-oriented, tender state experienced in
selection theory, also known as group selection.3 response to another individual’s distress that pro-
According to multilevel selection theory, selection motes altruistic behaviors aimed at relieving the dis-
can occur at the level of the gene, individual, or tress.1,16 A wealth of research has established the
group. Under some circumstances, altruism could be relationship between empathic concern and altruism.
selected if it yielded group-level benefits that out- In one classic paradigm, study participants watched
weighed individual-level costs.62 Although multilevel an obviously distressed stranger receiving electrical
selection theories are sometimes viewed in opposition shocks, then had the opportunity to volunteer to
toother theories of altruism, like kin selection, recent receive the remaining shocks in her place. Partici-
analyses suggest that they may actually be strongly pants induced to feel more empathic concern were
compatible.63 more likely to volunteer to receive the shocks even
That humans do experience the urge to care for whenanalternativewasoffered.72
unrelated children, even those who are strangers to Empathic concern may be powerful enough to
them, can be observed in many contexts today, override other self-interested or normative motiva-
including the adoption and fostering of children, the tions. For example, inducing empathic concern in the
use of child-related imagery to boost charitable giv- laboratory can upend the pattern of responses typi-
ing, even the urge to care for pets and dolls, particu- cally observed in the Prisoner’s Dilemma. During
larly those with infantile appearances.16 This urge is games in which participants’ learn of their partner’s
experienced by both sexes in our species, which is response before submitting their own response, parti-
unusualamongmammals,andmayreflecttheunusu- cipants will almost always (in upward of 95% of
allyintensive carerequiredbyhumaninfants.64 cases) defect if their partner defects first.73 But if par-
Care-based altruism is most effectively elicited ticipants are first induced to feel empathic concern
not by infantile features alone, but by infantile fea- for their partner, they will cooperate with her 45%
tures paired with indications of emotional dis- of the time even after she defects, and even though
tress.2,65 As it happens, the nonverbal cues used by cooperating conflicts with their material self-interest
many species, including humans, to signal distress as well as with social norms like fairness and distri-
tendtoincorporatefeaturesofinfants.Fearfulvocali- butivejustice.16
zations are typically high pitched, as are infant voca- It is important to distinguish empathic concern
lizations, and fear-related postural displays typically from other phenomena also commonly described as
make the individual appear smaller or otherwise ‘empathy,’ such as Theory of Mind, which is some-
more juvenile.58,66–68 Unlike most species, humans times called cognitive empathy, and empathic accu-
also rely heavily on facial expressive behavior for racy, which is sometimes called emotional
emotional communication.69 And human facial empathy.1,74,75 Cognitive empathy is the identifica-
expressions of fear also appear to serve to mimic tion of others’ high-level cognitive states, including
characteristics of infants.70,71 Fearful expressions are their focus of attention, intentions, and beliefs.76,77
perceived to look infantile, and perceivers rate them This is the form of empathy impaired by autism
Volume 7, January/February 2016 ©2015Wiley Periodicals, Inc. 63
AdvancedReview wires.wiley.com/cogsci
spectrumdisorders.76Cognitive empathyisrelianton time, her behavior toward pups radically transforms.
activity in higher-level perceptual and cognitive brain Now she not only does not avoid pups, but she will
areas including the medial prefrontal cortex and tem- make sacrifices to maintain contact with them. In
poroparietal junction, and it is not closely linked one study (described in detail by Preston2), rat
either to altruism or to antisocial behaviors like motherscontinuallypressedabarforhourswhenthe
aggression.75,78 On the other hand, emotional empa- barpressesallowed themtoaccessandretrieveasuc-
thy, which is the low-level representation or identifi- cession of unrelated rat pups.92 In another, new rat
cation of another’s emotional state, can be linked to mothers chose contact with rat pups over injections
altruism.15,79 Emotional empathy probably relies pri- ofapowerfullyenjoyabledrug(cocaine).93
marily on activity in the cortical and subcortical
structures responsible for generating the target emo-
tion. Hence empathy for pain involves activation in Oxytocin
dorsal anterior cortex and the anterior insula, which This shift in maternal orientation is driven primarily
are involvedinaffective andmotivational components by changes in the oxytocin system during and after
of experiencing pain,80 but empathy for fear relies on parturition. Toward the end of pregnancy and the
activation in the amygdala, which is the structure onset of lactation, the hypothalamic neurons that
mostdirectlyinvolvedintheexperienceoffear.81 manufacture oxytocin proliferate, as do oxytocin
It must be emphasized that emotional empathy receptors in brain areas such as the amygdala, bed
is not synonymous with empathic concern.81,82 Emo- nucleus of the stria terminalis, and septum.94,95 Both
tional empathy refers to understanding another’s the density of receptors and the amount of oxytocin
emotional state, not caring about that state—these produced appear to support the provision of mater-
are distinct processes. But some forms of emotional nal care. The density of oxytocin receptors in certain
empathy are important precursors for empathic con- brain regions strongly predicts variation in the qual-
cern. This is most demonstrably true regarding emo- ity of maternal care at baseline.86 And when oxyto-
tional empathy for fear, as individuals who are highly cin receptors are blocked chemically, maternal
sensitive to others’ fear—who can, for example, accu- behaviors drop off.96 Genetic manipulations that
rately identify it from others’ facial expressions—also eliminate oxytocin receptors also reduce maternal
tendtobehighly altruistic.15,79,83 Conversely, psycho- behaviors,97 but circulating levels of oxytocin are
paths, who lack empathic concern, also lack emo- important as well.98 Oxytocin injected directly into
tional empathy for others’ fear—but not for emotions the brains of virgin rats causes them to spontane-
likeangerordisgust.84,85 ously exhibit the full suite of maternal behaviors in
response to unrelated pups, including retrieving pups
to the nest and crouching over them, which simulta-
The Neural Basis of Care-Based Altruism neously provides pups with warmth, protection, and
At the neural level, care-based altruism appears to be nursing access.99,100 Intranasal administration of
driven by mechanisms that overlap heavily with the oxytocin is also effective, and can generate maternal
mechanisms driving parental care.2 These are com- care and alloparental care for unrelated young in a
prised primarily of subcortical structures containing variety of species.101–103 Exposure to infant-related
receptors for the neurohypophyseal hormone oxyto- sensorystimulicanalsoinitselftriggerincreasedpro-
cin.86,87Thishormone isproduced onlyin mammals, duction of oxytocin,104 which may explain why sim-
and its emergence is closely tied to the emergence of ple continuous exposure to infants typically results in
mammalian parental care.88 It is by no means the increasinglevels ofcare over time,even amongvirgin
only hormone involved in parental care (others femalerats.105
include vasopressin and prolactin, as well as gonadal Where does oxytocin act in the brain to pro-
hormones that prime the effects of neurohypophyseal duce these effects? Oxytocinergic messages are pro-
hormones87,89,90), but oxytocin is the hormone most jected from the paraventricular and supraoptic
robustlyassociatedwithparentalcare.91 nucleus of the hypothalamus to other subcortical
Parental care is frequently studied in rats structures that include the amygdala, bed nucleus of
because of how dramatically pregnancy changes the striaterminalis, striatum, and brainstem.106,107In
female rats’ responses to rat pups. Female rats who both humans and nonhuman animals, the amygdala
have not previously been pregnant find rat pups appears to be a central locus of oxytocin’s
actively aversive. They go out of their way to avoid effects.86,106,107 The amygdala is richly populated
them or may even cannibalize them.87 But around with oxytocin receptors, particularly in the central
the time a rat is ready to bear young for the first nucleus.86 This is the primary output nucleus of
64 ©2015Wiley Periodicals, Inc. Volume 7, January/February 2016
WIREsCognitiveScience Neural,cognitive,andevolutionaryfoundationsofhumanaltruism
the amygdala, which coordinates fear-related auto- also a critical region for perceiving others’ distress—
nomic and behavioral responses via its projections particularly fear-related cues like facial and vocal
to the hypothalamus and periaqueductal gray, expressions.121,122 Individuals with complete bilat-
respectively.108 eral amygdala lesions are impaired in recognizing
Within the central nucleus, oxytocin is facial and vocal expressions of fear, as are psycho-
expressed primarily in the lateral division, and pathic individuals who exhibit both amygdala dys-
increases in oxytocin production increase neuronal function and minimal empathic concern.84,85,123 This
excitabilityinthisregion.109Thismayseemparadox- may reflect the fact that the amygdala is the structure
ical, as oxytocin has anxiolytic effects.110,111 In neu- most directly involvedin the experience offear and it
roimaging studies oxytocin reduces overall amygdala may therefore be essential for empathic simulation of
activity in response to distress cues like fearful facial others’fear.81
expressions or infant cries.112–114 But neurons in the This may be why higher-than-average levels of
lateral division of the central nucleus include ‘off’ fear responsiveness and amygdala reactivity are asso-
neurons that are preferentially active in response to ciated with heightened altruistic responding
appetitive stimuli (and only appetitive stimuli) such (although excessive self-focused anxiety is associated
as cocaine or morphine.115 These neurons function with reduced altruistic behavior124). Rats bred for
to downregulate activity in the medial division of the high levels of anxiety provide better care to their
central amygdala.109 The role of these ‘off’ neurons pups than do low-anxiety rats,107 and this enhanced
may be to inhibit fear responding in the context of maternal care is associated with increases in the
appetitivestimuli. amount of oxytocin released in hypothalamus and
Infants themselves are powerfully appetitive sti- the central nucleus of the amygdala.98 High levels
muli.116 The characteristic features of an infant’s of fear-sensitivity in humans also are associated with
body and face have been conserved across species increased altruistic behavior as well as enhanced
precisely because they are so appealing.117,118 In amygdala reactivity.79,125–127 Higher fear sensitivity
humans, the appeal of infant faces has been demon- may reflect a more responsive amygdala that is
strated using a lever task in which neutral infant better equipped to respond empathically to others’
faces elicited behavioral approach, whereas neutral distress.
adult faces did not.59 In this study, participants view- Supporting this, extraordinarily altruistic adults
ing infant faces could pull a lever toward themselves are more sensitive than average to others’ fear. A
(approach) more quickly than they could push it recent study investigated neural and cognitive corre-
away from themselves (avoidance), a response likely lates of extraordinary altruism by assessing neural
driven by dopaminergic reward systems in the stria- structure and functioning in apopulation ofaltruistic
tum that are highly responsive to infant faces, partic- kidney donors, all of whom had donated a kidney to
ularly those considered classically ‘babyish.’116 a stranger.15 Relative to controls, altruists exhibited
Interestingly, similar approach-based behavioral astrongeramygdala response tofearful facialexpres-
responses are observed in response to adults’ fearful sions, and were relatively better at identifying fearful
expressions, corroborating the babyish qualities of expressions as well. Their amygdalae were also 8%
theseexpressions.59,119 larger than those of controls (in contrast to psycho-
Infants’ appeal also increases as oxytocin levels paths, who have relatively small amygdalae).128
rise. Infant faces are rated as more appealing follow- Importantly, however, altruists’ amygdalae were not
ing a dose of intranasal oxytocin, whereas the same generallymoreresponsivethancontrols—theiramyg-
is not true for adult faces, which are actually seen as dala response to angry expressions was actually
less appealing following oxytocin.120 Increased oxy- reduced. Their increased sensitivity was only in
tocin levels may potentiate the appetitive qualities of responsetofear.
sensory stimuli related to infants, whether it be the Together, these findings suggest that when indi-
smell, appearance, or vocalizations of actual infants, cations of vulnerability and distress are paired—
orstimuli thatmimicthesefeaturesofinfants.86 whether in the form of actual infant distress cries or
infantile-looking fearful expressions—the informa-
tionconvergesintheamygdala,perhapsinthelateral
Convergent Roles of the Amygdala division of the central nucleus, stimulating oxytocin-
The two key features of care-based altruism, then— mediatedinhibitionofbehavioralandautonomicfear
responses to distress and responses to infant cues— responding via the central nucleus’s efferent projec-
appear to converge in the amygdala. In addition to tions, and potentiating motivated altruistic
its role in supporting parental care, the amygdala is care.86,106,114,129 The provision of care appears to be
Volume 7, January/February 2016 ©2015Wiley Periodicals, Inc. 65
AdvancedReview wires.wiley.com/cogsci
coordinated via connections between the central altruism. Under normal circumstances, the two forms
amygdalaandregionsthatincludethehypothalamus, of altruism may be complementary, such that altru-
periaqueductal gray, stria terminalis, and dopaminer- ism is more likely the more closely related and infan-
gic cortico-striatal networks.2,95 Oxytocin activity tile a target is. This would result, appropriately, in a
in regions of sensory cortex may also play a role person’s own infant receiving more intense care than
in responding to infant-related sensory cues like dis- her older child, but the older child receiving more
tresscries.130 intense care than unrelated peers. But tensions
between these systems could arise as well. Respond-
ing to the distress of an unrelated individual to
the disadvantage of one’s own kin is an obvious
UNRESOLVED QUESTIONS
example, and one that arises in the real world,
Although research on altruism continues to converge particularly in the case of so-called ‘pathological
across disciplines, investigations of the various forms altruism.’ Wesley Autrey, for example, left his own
of altruism remain somewhat siloed. A major as-yet young daughters alone on a subway platform to
unresolved question concerns how altruistic beha- save a stranger.51 Any risky, altruistic act on behalf
viors based on kin, cooperation, and care intersect, of a stranger runs a similar risk of advantaging a
astheysurelydo. stranger over family if harm does befall the altruist
Although the study of reciprocal altruism has and causes family members distress. This may
focusedprimarilyontheroleofdopaminergicstriatal explain why such behavior is sometimes described
reward systems, the study of care-based altruism has aspathological.134
focused primarily on the role of oxytocinergic limbic Because multiple altruistic motivations almost
structures. It is likely, however, that these regions certainly co-occur with somefrequency,better under-
interact to support both forms of altruistic behavior. standingofthe interactions amongthem wouldlikely
Care systems may modulate the tendency to cooper- help to illuminate complex human behaviors as well
ate in reciprocal altruism paradigms and real-world as the neural and psychological basis of each form of
settings. Care may moderate the tendency to cooper- altruisminisolation.
ate as a first move in the absence of any baseline
information about one’s partner, given oxytocin’s
role in promoting trust.131 Care may also modulate Conclusion
retaliatory behavior following defection. If a partner Altruism is a central organizing principle among
is portrayed as vulnerable and distressed, retaliation group-living mammals, and there are few species for
plummets, suggesting that when the urge to provide which this is more evident than humans. Altruistic
care for a vulnerable individual is triggered, it may behavior is ubiquitous in human social interactions,
overwhelm the urge to maximize personal gain.73 whether it be kin-based, cooperation-based, or care-
This is consistent with evidence of mutual inhibition based,orsomecombinationofthesetypes.Kin-based
between striatal networks that promote approach altruisminvolvesaltruismtowardbiologicalrelatives,
and reward learning and limbic networks that pro- whose survival and reproduction will ultimately ben-
moteavoidanceandfear-learning.132 efit the genes of the altruist. This form of altruism
But the interaction between systems that sup- likelyreliesonneurocognitivesystemsthatenablethe
portcare-basedaltruismandreciprocalaltruismmust recognition of siblings and other close genetic rela-
be more complex than this, as striatal reward tives. Cooperation-based altruism involves sacrificing
mechanisms are also thought to support motivated immediate gains to benefit another in anticipation of
care-based behavior once the urge to provide altruis- future reward, and is supported by dopaminergic
tic care has been generated.2 These interactions may striato-cortical pathways, which are also likely mod-
occur at multiple levels. Subcortically, lesions to the erated by serotonin. Care-based altruism co-opts sys-
hypothalamic neurons that project to the dopamine- tems that originally developed to support parental
producing ventral tegmental area inhibit maternal care. In social species, these systems also support
behavior.133 But dopaminergic and oxytocinergic alloparental care, or care for others’ young. Allopar-
pathways also converge in the medial prefrontal cor- enting likely represents thedirect origin ofcare-based
tex, and this convergence may play an important altruism, which is elicited by signs of vulnerability
role in reinforcing the rewarding aspects of parental anddistress.Thisformofcareissupportedbyoxyto-
care.106 cinergic limbic structures like the amygdala.
Less empirical information is available regard- Althoughtheseformsofaltruismsurelyinteract,little
ing interactions between kin-based and care-based is known at this point about how these interactions
66 ©2015Wiley Periodicals, Inc. Volume 7, January/February 2016
WIREsCognitiveScience Neural,cognitive,andevolutionaryfoundationsofhumanaltruism
might occur. However, a recent surge in academic promising signs that the understanding of the neural
interest in altruism and an increasing emphasis on and psychological bases of altruism will continue
cross-disciplinary explorations of the topic are progressingrapidly.
REFERENCES
1. de Waal FB. Putting the altruism back into altruism: 17. Haldane JBS. Population genetics. New Biol 1955,
the evolution of empathy. Annu Rev Psychol 2009, 18:34–51.
59:79–300.
18. HamiltonWD.Thegeneticalevolutionofsocialbeha-
2. Preston SD. The origins of altruism in offspring care. viour.I.JTheorBiol1964,7:1–16.
PsycholBull2013,139:1305–1341.
19. Nowak MA, Tarnita CE, Wilson EO. The evolution
3. Kurzban R, Burton-Chellew MN, West SA. The evo- ofeusociality.Nature2010,466:1057–1062.
lution of altruism in humans. Annu Rev Psychol
20. Ratnieks FL, Helantera H. The evolution of extreme
2015,66:575–599.
altruism and inequality in insect societies. Philos
4. Scott-Phillips TC, Dickins TE, West SA. Evolutionary TransRSocLondBBiolSci2009,364:3169–3179.
theory and the ultimate-proximate distinction in the
21. Jarvis JU. Eusociality in a mammal: cooperative
human behavioral sciences. Perspect Psychol Sci
2011,6:38–47. breeding in naked mole-rat colonies. Science 1981,
212:571–573.
5. Mangan J. An historical analysis of the principle of
doubleeffect.TheolStud1949,10:41–61. 22. Lieberman D, Tooby J, Cosmides L. The architecture
ofhumankindetection.Nature2007,445:727–731.
6. BatsonCD,ShawLL.Evidenceforaltruism:towarda
pluralismofprosocialmotives.PsycholInquiry1991, 23. Burnstein E, Crandall C, Kitayama S. Some neo-
2:107–122. Darwinian decision rules for altruism: weighing cues
for inclusive fitness as a function of the biological
7. Darwin C. The Descent of Man. London: John Mur-
importance of the decision. J Pers Soc Psychol 1994,
rayPublishers;1871.
67:773.
8. Dawkins R. The Selfish Gene. Oxford: Oxford Uni-
24. Tong A, Chapman JR, Wong G, Kanellis J,
versityPress;2006.
McCarthy G, Craig JC. The motivations and experi-
9. Burkart JM, Allon O, Amici F, Fichtel C,
ences of living kidney donors: a thematic synthesis.
Finkenwirth C, Heschl A, Huber J, Isler K, AmJKidneyDis2012,60:15–26.
Kosonen ZK, Martins E, et al. The evolutionary ori-
25. Matas AJ, Smith JM, Skeans MA, Lamb KE,
ginofhumanhyper-cooperation.NatCommun2014,
Gustafson SK, Samana CJ, Stewart DE, Snyder JJ,
5:4747.
Israni AK, Kasiske BL. OPTN/SRTR 2011 annual
10. Sherman PW. The meaning of nepotism. Am Nat
datareport:kidney.AmJTransplant2013,13(Suppl
1980,116:604–606. 1):11–46.
11. Dunford C. Kin selection for ground squirrel alarm
calls.AmNat1977,111:782–785. 26. HamiltonWD.Thegeneticalevolutionofsocialbeha-
viour.II.JTheorBiol1964,7:17–52.
12. Carter GG, Wilkinson GS. Food sharing in vampire
27. RushtonJP,RussellRJH,WellsPA.Geneticsimilarity
bats: reciprocal help predicts donations more than
theory: beyond kin selection. Behav Genet 1984,
relatedness or harassment. Proc Biol Sci 2013,
14:179–193.
280:20122573.
28. Volk AA, Quinsey VL. Parental investment and
13. Krakauer AH. Kin selection and cooperative court-
shipinwildturkeys.Nature2005,434:69–72. resemblance: replications, refinements, and revisions.
EvolPsychol2007,5:1–14.
14. Rand DG, Epstein ZG. Risking your life without a
second thought: intuitive decision-making and 29. DeBruine LM, Jones BC, Little AC, Perrett DI. Social
extremealtruism.PLoSOne2014,9:e109687. perceptionoffacialresemblanceinhumans.ArchSex
Behav2008,37:64–77.
15. Marsh AA, Stoycos SA, Brethel-Haurwitz KM,
Robinson P, Cardinale EM. Neural and cognitive 30. Platek SM, Keenan JP, Mohamed FB. Sex differences
characteristics of extraordinary altruists. Proc Natl intheneuralcorrelatesofchildfacialresemblance:an
AcadSciUSA2014,111:15036–15041. event-related fMRI study. Neuroimage 2005,
25:1336–1344.
16. BatsonCD.Thenakedemperor:seekingamoreplau-
sible genetic basis for psychological altruism. Econ 31. Platek SM, Kemp SM. Is family special to the brain?
Philos2010,26:149–164. Anevent-relatedfMRIstudyoffamiliar,familial,and
Volume 7, January/February 2016 ©2015Wiley Periodicals, Inc. 67
AdvancedReview wires.wiley.com/cogsci
self-face recognition. Neuropsychologia 2009, 48. PuigMV,GulledgeAT.Serotoninandprefrontalcor-
47:849–858. tex function: neurons, networks, and circuits. Mol
32. Campbell DT. On the conflicts between biological Neurobiol2011,44:449–464.
and social evolution and between psychology and 49. Robbins TW. Chemical neuromodulation of frontal-
moraltradition.Zygon1976,11:167–208. executive functions in humans and other animals.
33. Griffin AS, West SA. Kin selection: fact and fiction. ExpBrainRes2000,133:130–138.
TrendsEcolEvol2002,17:15–21. 50. Walderhaug E et al. Lowering of serotonin by rapid
tryptophan depletion increases impulsiveness in nor-
34. Trivers RL. The evolution of reciprocal altruism. Q
RevBiol1971,46:35–57. mal individuals. Psychopharmacology (Berl) 2002,
164:385–391.
35. Cosmides L, Tooby J. Evolutionary psychology and
51. Buckley C. Man is rescued by stranger on subway
the generation of culture, part II: case study: a com-
tracks.TheNewYorkTimes,January3,2007.
putational theory of social exchange. Ethol Sociobiol
1989,10:51–97. 52. HrdySB.MothersandOthers:TheEvolutionaryOri-
ginsofMutualUnderstanding.Cambridge,MA:Har-
36. Rilling JK, Sanfey AG. The neuroscience of social
decision-making.AnnuRevPsychol2011,62:23–48. vardUniversityPress;2009.
53. Eibl-Eibesfeldt I. Love and Hate: the Natural History
37. Ikemoto S, Panksepp J. The role of nucleus accum-
ofBehaviorPatterns.NewYork:Aldine;1996.
bens dopamine in motivated behavior: a unifying
interpretation with special reference to reward-seek- 54. Cosmides L, Tooby J. Beyond intuition and instinct
ing.BrainResRev1999,31:6–41. blindness:towardanevolutionarilyrigorouscognitive
science.Cognition1994,50:41–77.
38. Kable JW, Glimcher PW. The neural correlates of
subjective value during intertemporal choice. Nat 55. Insel TR. Science, philosophy, and religion in dia-
Neurosci2007,10:1625–1633. logue. In: Post SG, Underwood G, Schloss JP, eds.
Altruism and Altruistic Love. New York: Oxford
39. Cardinal RN, Parkinson JA, Hall J, Everitt BJ. Emo-
UniversityPress;2002,254–263.
tionandmotivation:theroleoftheamygdala,ventral
striatum, and prefrontal cortex. Neurosci Biobehav 56. LingleS,RiedeT.Deermothersaresensitivetoinfant
Rev2002,26:321–352. distress vocalizations of diverse mammalian species.
AmerNat2014,184:510–522.
40. Knutson B, Cooper JC. Functional magnetic reso-
nanceimaging of rewardprediction. Curr OpinNeu- 57. McDougallW.AnIntroductiontoSocialPsychology.
rol2005,18:411–417. London:Methuen;1908.
41. Rilling J, Gutman D, Zeh T, Pagnoni G, Berns G, 58. LorenzK.OnAggression.London:Methuen;1966.
KiltsCA.Neuralbasisforsocialcooperation.Neuron 59. HammerJL,MarshAA.Whydofearfulfacialexpres-
2002,35:395–405. sions elicit behavioral approach? Evidence from a
42. Rilling JK, Sanfey AG, Aronson JA, Nystrom LE, combined approach-avoidance implicit association
CohenJD.OpposingBOLDresponsestoreciprocated test.Emotion2015,15:223–231.
and unreciprocated altruism in putative reward path- 60. Panksepp J, Panksepp JB. Toward a cross-species
ways.Neuroreport2004,15:2539–2543. understanding of empathy. Trends Neurosci 2013,
43. AxelrodR,HamiltonWD.Theevolutionofcoopera- 36:489–496.
tion.Science1981,211:1390–1396. 61. Swain JE, Konrath S, Brown SL, Finegood ED,
44. King-Casas B, Tomlin D, Anen C, Camerer CF, Akce LB, Dayton CJ, Ho SS. Parenting and beyond:
QuartzSR,MontaguePR.Getting toknowyou:rep- common neurocircuits underlying parental and altru-
utationandtrustinatwo-personeconomicexchange. isticcaregiving.ParentSciPract2012,12:115–123.
Science2005,308:78–83. 62. Wilson DS, Wilson EO. Rethinking the theoretical
45. Schweighofer N, Bertin M, Shishida K, Okamoto Y, foundation of sociobiology. Q Rev Biol 2007,
Tanaka SC, Yamawaki S, Doya K. Low-serotonin 82:327–348.
levelsincreasedelayedrewarddiscountinginhumans. 63. Lehmann L, Keller L, West S, Roze D. Group selec-
JNeurosci2008,28:4528–4532. tion and kin selection: two concepts but one process.
46. Tse WS, Bond AJ. Serotonergic intervention affects ProcNatlAcadSciUSA2007,104:6736–6739.
both social dominance and affiliative behaviour. Psy- 64. Kaplan H, Hill K, Lancaster J, Hurtado AM. A the-
chopharmacology(Berl)2002,161:324–330. ory of human life history evolution: diet, intelligence,
andlongevity.EvolAnthropol2000,9:156–185.
47. Wood RM, Rilling JK, Sanfey AG, Bhagwagar Z,
Rogers RD. Effects of tryptophan depletion on the 65. Lishner DA, Oceja LV, Stocks EL, Zaspel K. The
performance of an iterated Prisoner’s Dilemma game effect of infant-like characteristics on empathic con-
in healthy adults. Neuropsychopharmacology 2006, cern for adults in need. Motiv Emot 2008,
31:1075–1084. 32:270–277.
68 ©2015Wiley Periodicals, Inc. Volume 7, January/February 2016