Bringing basic research on early experience and stress neurobiology to bear on preventive interventions for neglected and maltreated children
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Development and Psychopathology 18 ~2006!, 651–677
Copyright © 2006 Cambridge University Press
Printed in the United States of America
DOI: 10.10170S0954579406060330
Bringing basic research on early experience
and stress neurobiology to bear on
preventive interventions for neglected
and maltreated children
MEGAN R. GUNNAR,a PHILIP A. FISHER,b and THE EARLY
EXPERIENCE, STRESS, AND PREVENTION NETWORK
a Institute
of Child Development, University of Minnesota; and
b Oregon Social Learning Center
Abstract
A major focus in developmental psychopathology is on understanding developmental mechanisms and, armed with
this information, intervening to improve children’s outcomes. Translational research attempts to bridge the distance
between understanding and intervention. In the collaborations that have formed the core of our research network on
early experience, stress, and prevention science, we have focused on translating basic research on early experiences
and stress neurobiology into preventive interventions for neglected and abused children. Our experiences in
attempting to move from bench to bedside have led us to recognize the many challenges that face translational
researchers. This review provides a brief synopsis of the animal model literature on early experience and stress
neurobiology from which we glean several key bridging issues. We then review what is currently known about the
impact of childhood neglect and abuse on stress neurobiology in human adults and children. Next, we describe how
this work has informed the evaluation of our preventive interventions with maltreated children. Finally, we discuss
several considerations that should facilitate a more complete integration of basic research on early experience and
stress neurobiology into preventive intervention strategies.
How do early care experiences shape develop- accumulate lists of all the ways in which early
ment, and to what extent can later experiences adversity may limit or impair children. Rather,
remediate the effects of early adverse care? the objectives have been to better understand
These questions have tremendous importance developmental mechanisms of early adversity
for our understanding of developmental psy- and, armed with this information, to intervene
chopathology ~Cicchetti, 1989!. The goals of to improve children’s outcomes. Unfortu-
research on early adversity have never been to nately, there is a large distance between em-
barking on the first objective and achieving
This paper reflects the work of the Early Experience, the second one. Bridging this distance is a
Stress and Prevention Science Network ~R21 MH65046!, primary goal of translational research ~Strau-
whose members are Mary Dozier, Philip Fisher, Nathan man & Merrill, 2004!.
Fox, Megan Gunnar, Seymour Levine, Charles Neal, Seth
In the collaborations that have formed
Pollak, Paul Plotsky, Mar Sanchez, and Delia Vazquez.
Preparation of this manuscript was supported by a Senior the core of our Early Experiences, Stress,
Scientist Award ~K05 MH66208! to Megan Gunnar, and and Prevention Science Research Network, we
by MH59780 and MH65046, NIMH, U.S. PHS; MH46690, have focused on one facet of this translational
NIMH and ORMH, U.S. PHS; and DA17592, NIDA, challenge: applying basic research on the de-
NIH, U.S. PHS to Philip Fisher.
velopmental neurobiology of stress to the un-
Address correspondence and reprint requests to: Megan
R. Gunnar, Institute of Child Development, 51 East River derstanding of how early neglect and abuse
Road, University of Minnesota, Minneapolis, MN; E-mail: shapes vulnerability to emotional and behav-
gunnar@umn.edu. ioral disorders and then using this information
651652 M. R. Gunnar, P. A. Fisher, and The Early Experience, Stress, and Prevention Network
to inform preventive interventions for mal- disorders ~e.g., Manly, Kim, Rogosch, & Cic-
treated children. Our experiences in attempt- chetti, 2001!. Although some children who
ing to move from bench to bedside have led us suffer maltreatment are resilient, becoming
to recognize the many challenges that face competent adults despite the odds ~Kinard,
translational researchers. These challenges will 1998!, many others suffer from a variety of
need to be overcome if we are to adequately clinical disorders in adulthood including
integrate basic neuroscience into developmen- substance abuse and affective and personality
tal psychopathology and effective preventive disorders ~e.g., Putnam, 2003!. Research ex-
treatments ~Cicchetti & Tucker, 1994!. In what amining why some individuals are resilient
follows, we will describe the paths we have has identified protective factors, but cannot
taken and some of the issues with which we wholly explain resilience among maltreated
have grappled. We wish that we could con- children ~e.g., Cicchetti & Rogosch, 1997; Ki-
clude by demonstrating that we have accom- nard, 1998!. Similarly, efforts to determine
plished the task we set ourselves, fully moving maltreatment factors associated with different
research on early experiences and stress neuro- types of disorders has examined such charac-
biology into effective preventive interven- teristics as the timing, dose, duration, and type
tions for children. However, at this stage, what of maltreatment ~e.g., Manly et al., 2001!.
we have to offer is a progress report on where These studies indicate that more severe, mul-
we are and the obstacles we and others have tiple, and prolonged maltreatment results in
encountered thus far. increased symptoms; however, a coherent
We begin this review by briefly summariz- and consistent model of the differential ef-
ing the data on behavioral outcomes of ne- fects of specific subtypes of maltreatment on
glected and abused children and the major specific symptom clusters or mental health
psychological orientations that have been ap- disorders does not appear to have emerged.
plied to explicate why these outcomes occur. Interestingly, one challenge to identifying
Next, we very briefly synopsize the vast liter- associations between the type and timing of
ature, spanning the last half-century, on early maltreatment and disordered outcomes may
experience and stress neurobiology using ro- be that outcomes vary over the course of de-
dent models. From this work we have distilled velopment. Externalizing problems appear to
key issues that provide a framework for our predominate during childhood, while sub-
translational efforts. We then review what is stance abuse and affective disorders are ob-
currently known about the impact of child- served in adulthood ~e.g., Tieman, van der
hood neglect and abuse on stress neurobiol- Ende, & Verhulst, 2005!. The developmental
ogy in human adults and children. Based on trajectories associated with these varied out-
similarities in primate and human develop- comes are at present not understood, due in
ment, we extend this review to nonhuman part to the complexity and cost of prospective
primate studies. Finally, we discuss how this longitudinal designs.
basic work has informed the evaluation of our Until recently, theoretical perspectives on
preventive interventions with neglected and maltreatment have been solely psycho-
maltreated children, and considerations that logical, reflecting two dominant lenses: the
may move the field forward to more fully in- developmental–organizational perspective, and
tegrate an understanding of stress neurobiol- the social learning theory perspective. From a
ogy into preventive intervention strategies. developmental–organizational perspective,
maltreatment disturbs resolution of stage-
salient developmental issues. If the experi-
Early Neglect and Abuse: Increased Risk ence occurs early in development, attachment
of Behavioral and Emotional Problems relationships are likely to be disturbed ~Page,
1999!. Studies have revealed that exposures
Decades of research have provided unequivo- to frightening or overwhelming behaviors from
cal evidence that childhood maltreatment in- the caregiver are associated with the develop-
creases the risk of psychological and behavioral ment of a disorganized0disoriented ~Type D!Stress neurobiology and prevention science 653
attachment, characterized by freezing and dis- may affect brain development and heighten
sociative behavior and heightened risk for risk of psychological disorders ~see for re-
numerous poor outcomes in childhood and be- views, Bremner & Vermetten, 2001; De Bel-
yond ~Lyons-Ruth, 2003!. lis, 2005; Heim, Owen, Plotsky, & Nemeroff,
Through the lens of social learning theory, 1997; Shea, Walsh, Macmillan, & Steiner,
learning principles account for behaviors ex- 2005; Teicher, Andersen, Polcarri, Anderson,
hibited by neglected and abused children ~Reid & Navalta, 2002!. Neurobiological models may
& Kavanagh, 1985!. Behaviors on the part of hold keys to mechanisms through which psy-
the child that are less likely to elicit abusive chosocial interventions operate and to expla-
treatment may be adaptive in the context of nations of individual differences in response
abuse, but when generalized to other settings to treatment. These neurobiological models are
may disrupt the child’s relationships ~e.g., with largely based on animal studies of the effects
teachers, peers, and so on!. Indeed, poor peer of early life adversity on stress neurobiology
relations associated with a tendency to attribute and brain development. The largest body of
hostile intent to others has been noted for mal- such early experience research has been con-
treated children ~Price & Glad, 2003!. Mal- ducted on rodents, notably the rat.
adaptive social behavior may lead children into Over the past 50 years, rodent early expe-
antisocial peer groups, increasing the risk of rience studies have shown that early parental
conduct disorders, and substance abuse ~Pat- care profoundly influences brain develop-
terson, DeBaryshe, & Ramsey, 1989!. Mal- ment, regulates gene expression, and shapes
treated children have been consistently shown the neural systems that in humans are in-
to be at high risk for externalizing behavior, volved in vulnerability to affective disorders
conduct problems and substance abuse ~Ege- in response to later stressful life events ~e.g.,
land, Yates, Appleyard, & van Dulmen, 2002!. Levine, 2005a!. Recent studies also indicate
To date, preventive interventions for mal- that interventions in the postinfancy period
treated children also have been based almost may help ameliorate some, but not all, of the
exclusively on the two psychological perspec- impacts of early inadequate parental care ~Fran-
tives describe above. Early interventions based cis, Diorio, Plotsky, & Meaney, 2002!. From
on developmental–organizational perspec- the beginning, this rodent research was con-
tives have focused on attachment, attempting ducted to inform the understanding of human
to increase parental sensitivity, and responsive- development and psychopathology. Below we
ness to foster more secure parent–child rela- review the rodent literature and arrive at a
tionships and alter the child’s inner working core list of critical bridging themes. Thorough
models ~Cicchetti, 2005; Dozier, 2003!. Inter- understanding of these themes, however, re-
ventions based on social learning theory have quires a brief description of the neurobiology
attempted to avert child behavior problems by of stress.
training parents to use clear, consistent, non-
hostile guidance and discipline techniques
The Neurobiology of Stress
~Fisher, Burraston, & Pears, 2005!. The effec-
tiveness of these psychosocial interventions is Stress neurobiology is organized at three lev-
typically evaluated at the group level; how- els ~see Figure 1!: a corticolimbic level that or-
ever, there are always differences in effective- chestrates responses to anticipated threat, a
ness at the level of the individual. These hypothalamic–brainstem level that coordinates
psychosocial models often fall short of fully central and peripheral responses in response to
explicating why some individuals respond and corticolimbic input ~and also in response to less
others do not. processed input regarding threats to homeosta-
sis!, and a neural to adrenal gland level that ef-
fects increases in stress-sensitive hormones
Early Experience and Stress Neurobiology
~glucocorticoids and epinephrine; Herman &
Recently, researchers have turned to neurobi- Cullinan, 1997!. Corticotropin-releasing fac-
ological models to explain how maltreatment tor ~CRF! is a neuroactive peptide that oper-654
Figure 1. The three levels of neurobiological organization of the stress system that are responsive to psychological stressors. The corticolimbic level of
organization involves the anterior cingulate ~ACC! and orbital frontal cortex ~OFC!, which relay information to subcortitical structures involved in the stress
response. The ACC and OFC are reciprocally interconnected with each other and with the amygdala, which has connections with the hippocampus and BNST.
The hypothalamic–brainstem level of organization involve the hippocampus and brainstem structures such as the locus coeruleus, which releases NE to brain
areas involved in alerting. The BNST provides pathways into the PVN of the hypothalamus, which produces corticotrophin-releasing hormone ~CRH! and
arginine vasopressin ~AVP!, whereas the hippocampus and regions in the medial frontal cortex ~e.g., ACC! maintain feedback control on the paraventricular
nucleus ~PVN!. Considering the neural to adrenal level of analysis, nuclei in the lateral hypothalamus activate highly interconnected nuclei in the brainstem,
including the parabrachial nuclei, that regulate the sympathetic ~NE and epinephrine, EPI! and parasympathetic ~acetylcholine, Ach! nervous systems via
pathways traveling through the spinal cord to preganglionic nuclei or to target organs ~e.g., the adrenal medulla!. The production of CRH and AVP by the PVN
regulates activity of the HPA axis and the production of glucocorticoids ~GCs! as depicted more fully in Figure 2. Adapted from Gunnar and Davis ~2003!.Stress neurobiology and prevention science 655 ates at the first two levels to coordinate behav- The second CRF pathway involves CRF- ioral, emotional, autonomic, and endocrine fac- producing neurons in the central nucleus of ets of stress and defensive responding the amygdala ~CeA; Van Bockstaele, Colago, ~Heinrichs & Koob, 2004!. Two loosely cou- & Valentino, 1998!. The CeA receives infor- pled CRF pathways are involved. One, termed mation through multiple pathways, many of the hypothalamic CRF pathway, involves CRF- which involve cortical systems that support producing neurons in the paraventricular nu- the integration of past and present experi- clei of the hypothalamus ~PVN !, which ences, thereby allowing the anticipation of regulates activity of the hypothalamic– threat ~see Figure 1!. Psychosocial stressors pituitary–adrenocortical ~HPA! axis ~see Fig- operate through these pathways converging ure 2!. CRF secreted from these neurons travels on the CeA to orchestrate behavioral, auto- through a small blood connection to the ante- nomic, and neuroendocrine reactions in antici- rior pituitary where it, along with other co- pation of threat ~Heinrichs & Koob, 2004!. secretagogues ~e.g., arginine vasopressin, Efferent CeA–CRF pathways project indi- AVP!, stimulates the production and release rectly to PVN–CRF neurons through the bed of adrenocorticotropic hormone ~ACTH!. nucleus of the stria terminalis ~BNST! to stim- ACTH released into general circulation stim- ulate the HPA axis, and directly to the locus ulates cells in cortex of the adrenal glands to coeruleus to stimulate release of norepineph- produce and release glucocorticoids. rine ~NE! into the terminal fields of the as- Glucocorticoids ~predominantly cortisol in cending noradrenergic system, supporting primates and corticosterone in rodents! are cognitive arousal and focusing, and behav- steroid hormones that affect almost every or- ioral and emotional components of fight0fight0 gan and tissue of the body ~Sapolsky, Romero, freeze responses ~Van Bockstaele et al., 1998!. & Munck, 2000!. In the brain, glucocorticoids Bidirectional connections of the amygdala and operate through two types of receptors with regions in the medial prefrontal cortex ~mPFC; distinct functions ~de Kloet, Vreugdenhil, Oitzl, orbital frontal cortex or OFC and anterior cin- & Joels, 1998!. At basal levels glucocorti- gulate cortex or ACC! support the modulation coids operate predominantly through Type I of behavior and cognition in relation to expec- or mineralocorticoid receptors to maintain the tations of rewards and punishments. These cor- sensitivity of neurons to their neurotransmit- ticolimbic pathways influences the balance ters and maintain the capacity of the brain to between responding based on rapid, habitual, respond to conditions that potentially threaten emotionally charged modes of acting and more the organism’s viability. At elevated or stress nuanced, considered, and dispassionate modes levels, glucocorticoids operate through Type II ~Sullivan & Gratton, 2002!. Notably, chronic or glucocorticoid receptors ~GR! to counter- intravenous infusions of glucocorticoids tend act the impact of other stress processes, sculpt to upregulate CeA–CRF activity, biasing func- neural systems to retain information about tioning toward rapid, emotion-charged fight0 threats to well-being, and return the organism flight0freeze responses, while at the same time to prestress levels of functioning. GR-mediated downregulating PVN–CRF, resulting in nor- neurochemical events, if well timed, acute, mal to hyporesponsiveness of the HPA axis and contained, support organism via- ~Schulkin, McEwen, & Gold, 1994!. In addi- bility; however, these events are potentially tion, particularly during development, CRF damaging if too frequent or too prolonged and glucocorticoids interact with other neuro- ~Sapolsky et al., 2000!. Multiple mechanisms peptides ~e.g., oxytocin! and neurotransmitter operate in response to elevated glucocorti- systems ~e.g., serotonin and dopamine!, result- coids to suppress or downregulate the HPA ing in widespread influences of frequent stress system including negative feedback mecha- level activation of the CeA–CRF and PVN– nisms that regulate acute responses and geno- CRF systems on emotional and cognitive de- mic alterations at various levels of the system velopment ~Roceri et al., 2004!. that affect the responsiveness following pro- Genetic variability appears to modify these longed or frequent activation. effects. As only one example, individuals car-
656 M. R. Gunnar, P. A. Fisher, and The Early Experience, Stress, and Prevention Network
Figure 2. The brain structures that are central to the activation and inhibition of the HPA endocrine
stress response: ~A! the main brain areas that participate in the regulation of the HPA axis and ~B! a
schematic representation of the HPA endocrine stress response initiated by the release of CRH0AVP
from the medial parvocellular region of the paraventricular nucleus ~mpPVN! in the hypothalamus.
Glucocorticoids inhibit this system acting at the level of the pituitary, hypothalamus, and hippocampus.
GABA, gamma aminobutyric acid; CRH, corticotropin-releasing hormone. From “Stress Neurobiology
and Developmental Psychopathology,” by M. Gunnar and D. M. Vazquez. In D. Cicchetti and D. Cohen
~Eds.!, Developmental Psychopathology: Developmental Neuroscience ~2nd ed., Vol. 2!, 2006, New
York: Wiley. Copyright 2006 John Wiley & Sons, Inc. Reprinted with permission.
rying at least one short allele of the serotonin show larger cortisol responses to psycho-
transporter gene polymorphism exhibit larger social stressors ~Barr et al., 2004; Sanchez,
amygdala responses to threat stimuli ~Hariri Noble, et al., 2005!. They are also at higher
et al., 2002!. In addition, these individuals risk for depression following early experi-Stress neurobiology and prevention science 657
ences of abuse and neglect ~Caspi et al., 2003; However, comparing brain development across
Kaufman et al., 2004!. mammalian species is difficult. Although the
general pattern of brain development is prob-
Key Translational Issues ably comparable across mammals, specifics
likely vary and specifics are required to trans-
Models derived from rodent studies of ad- late mechanisms of early experience effects.
verse early experience can be employed to For example, methylation of the GR gene
provide a biological level of explanation expressed in the hippocampus appears to be an
for the above noted association in children important mechanism through which early ex-
between early maltreatment and later height- periences in the rodent influence later stress
ened risk for emotional and behavioral disor- reactivity and vulnerability ~for review, see,
ders. At the most general level, the rodent Meaney & Szyf, 2005!. At birth, GR genes in
studies show that disturbances in early care the rodent hippocampus are heavily methyl-
exert a pervasive and lasting impact on the ated. This means that they are not available for
two neural pathways described above, and that transcription. The GR gene in the brain tran-
alterations in these systems has potential to scribes the proteins needed to produce the GR,
compromise subsequent development ~Meaney which, in turn, mediates many components of
& Szyf, 2005; Sanchez, Ladd, & Plotsky, the stress response, including negative feed-
2001!. However, if we are to move beyond back regulation of the HPA axis in response to
these broad generalities to inform specific in- psychosocial stressors. Maternal care in the in-
tervention strategies, we need to determine fant rodent determines patterns of GR gene de-
whether and how particular elements of the methylation with higher levels of maternal
rodent early experience models apply to hu- stimulation ~i.e., licking and grooming! being
man development. This, in turn, requires at- associated with more demethylation. Accord-
tention to the critical details of the rodent ingly, the offspring of high licking and groom-
findings. As we discuss in this section, we ing mothers have more operational GR genes
have identified four dimensions of the rodent in the brain and regulate stress more effec-
models that appear critical to building the trans- tively. These demethylations of the GR gene
lational bridge: occur most prominently during the first week
1. developmental timing of adversity, of life in the rat pup, consistent with evidence
2. the presence ~or absence! of a relative stress that manipulations that decrease maternal care
hyporesponsive period ~SHRP! during hu- have more profound effects if they are imposed
man development, beginning in the first rather than the second post-
3. parental or caregiver mediation of any natal week. Manipulations in the later peri-
SHRP, and pubertal period that reduce many of these early
4. the impact of caregiving that mediates the experiences effects do not affect GR methyl-
SHRP on the development of corticolim- ation, suggesting that this early experience ef-
bic stress response organization. fect is more or less permanent ~Francis et al.,
2002!. GR methylation can be affected in adult-
hood in the rat through pharmacological ma-
Timing
nipulations ~Weaver et al., 2005!. Importantly,
The effects of experiences on brain develop- pharmacological manipulations that demethyl-
ment depend on the maturity of the brain when ate the GR gene also result in changes in other
the events are experienced. Timing is critical. aspects of stress neurobiology, suggesting that
Rat pups are born young relative to human GR methylation is not only relatively perma-
infants; that is, birth occurs in the rodent when nent but also fairly critical to the evidence that
the brain is much less mature than the brain of early experiences in the rodent have lifelong
the full-term human infant. Indeed, the first consequences for adult stress vulnerability and
week of the rodent’s life are often equated resilience.
with development of the human infant during Translating the results of rodent GR gene
the last trimester of gestation ~Dobbing, 1981!. studies requires that we know when in human658 M. R. Gunnar, P. A. Fisher, and The Early Experience, Stress, and Prevention Network
development adult levels of central GR methyl- are observed in the first 2 weeks of the pup’s
ation are determined. As far as we know, this life. This roughly corresponds to a develop-
information simply is not available for hu- mental epoch in rodents that has been referred
mans or any other primate. Indeed, our level to as the relative stress hyporesponsive period
of knowledge of human developmental neuro- ~SHRP!. During the SHRP, the rat pup’s HPA
biology is sufficiently limited that in many axis ~the adrenal cortex in particular! shows
cases, even when basic studies provide a tar- very little reactivity when challenged by a va-
get for early experience effects, that knowl- riety of stressors ~Sapolsky & Meaney, 1986!.
edge does not help us pin point comparable The SHRP may have evolved to protect the
periods in the development of stress neurobi- rapidly developing brain from the impact of
ology in humans. Nonetheless, given the im- elevated glucocorticoids. Indeed, there is con-
portance of the first postnatal week in rodents siderable evidence that although basal activity
and its rough comparability to the last trimes- of the HPA axis is necessary for normal devel-
ter of gestation in humans, it is prudent to opment of the central nervous system, ele-
broaden our early experience-translational win- vated levels of glucocorticoids and CRH during
dow to include the prenatal period. this period in the rodent result in significant
Certainly there is abundant evidence that apoptosis ~cell death! and alter the develop-
many children who suffer adverse care during ment of brain regions that play critical roles in
postnatal development often are the products learning, memory, and stress resilience ~Sapol-
of high risk or stressed pregnancies ~Thomp- sky & Meaney, 1986!. In building a transla-
son et al., 1994!. There is increasing evidence tional bridge, therefore, we need to know
that maternal stress and anxiety during preg- whether there is a period in human develop-
nancy are associated with lower birth weight ment that is functionally comparable to the
infants, in part through increased maternal HPA SHRP in the rodent. That is, is there a period
activity, upregulation of placental CRH pro- during human development when it is diffi-
duction, with resulting decreases in birth weight cult to produce elevations in cortisol to stress-
and length of gestation ~Wadhwa, 2005!. Lower ors, and if so, when? Disturbances in care
birth weight is a known risk factor for poor de- during this time would be hypothesized to have
velopmental outcomes, many of which over- the greatest impact on the development of stress
lap with problems observed for neglected and neurobiology.
abused children ~e.g., Indredavikm, Vik, Hey- Evidence is accumulating that in human
erdahl, Kulseng, & Brubakk, 2005!. There is children there may be a roughly comparable
also increasing evidence that lower birth weight period that emerges in infancy and extends
is associated in adults with increased risk of throughout most of childhood ~see for review,
metabolic syndrome ~i.e., high blood pressure, Gunnar, 2003!. At birth, cortisol elevates
high cholesterol, abdominal fat, type II diabe- readily to a wide variety of stressors from
tes! and that elevated and poorly regulated glu- noninvasive stressors like a physical examina-
cocorticoids may mediate this risk ~Phillips tion to invasive stressors such as a heel lance.
et al., 2000!. Thus, as we proceed with our work The HPA system appears to remain highly re-
on postnatal maltreatment, we should keep in sponsive for several months postbirth, exhib-
mind that although the rodent findings appear iting elevations to physical examinations and
promising in explaining the impact of post- to childhood inoculations. However, over the
natal neglect and abuse in human children, some course of the first year it becomes increas-
of the mechanisms might actually translate more ingly difficult to produce elevations in corti-
directly to the impact of adversity on prenatal sol to acute stressors, including physical
human development. examinations, brief separations ~e.g., 3–5 min!,
inoculations, approach by strangers, and other
events that are capable of eliciting increases
Relative stress hyporesponsive period
in heart rate and behavioral distress.
In the rodent, the most profound effects pro- Although it seems fairly clear that the hu-
duced by systematically altering parental care man stress hyporesponsive period emergesStress neurobiology and prevention science 659
gradually over the first year, it is not as clear children when they are in full-day out-of-
how long it extends. Recent studies suggest home child care may provide some insight
that by puberty laboratory stressor tasks pro- ~for review, see Gunnar & Donzella, 2002!.
duce elevations in cortisol with increasing re- Although child care is not equivalent to the
sponse levels observed over the pubertal separation paradigms examined in the animal
transition ~reviewed in Gunnar & Vazquez, studies, child care often involves a reduction
2006!. Similar to the rodent, where corticoste- in individualized care. Comparable to the ro-
rone levels rise as the animal emerges out of dent findings, cortisol levels are not elevated
the SHRP, a rise in basal levels of cortisol has in the first few hours of the child care day;
been reported in human children around this however, by late afternoon, levels are higher
point in development ~see for review, Gunnar than are noted at home on nonchild care days.
& Vazquez, 2006!. An increase in basal activ- Increases over home levels are largest for tod-
ity of the axis and a corresponding increase in dlers, but are still significant in cross-sectional
HPA reactivity to stressors around the puber- studies until children are 5 or 6 years old.
tal transition would be consistent with pu- Similar to the rodent findings, children with
berty marking the close of a relative stress child care providers who engage them with
hyporesponsive period in human develop- focused attention and responsive stimulation
ment. If the HPA stress hyporesponsive period do not exhibit these elevated levels. Further
demarcates the period when stress neurobiol- evidence for caregiver mediation of the hu-
ogy is open to be shaped by experience, then man functional equivalent of the SHRP comes
in humans that period may extend throughout from studies of attachment in toddlers during
childhood. a fear-eliciting stress paradigm ~Nachmias,
Gunnar, Mangelsdorf, Parritz, & Buss, 1996!.
Toddlers in secure attachment relationships
Caregiving Mediation
show no elevations in cortisol, whereas tod-
In the rodent, the evidence is fairly conclusive dlers in insecure attachment relationships ex-
that the SHRP is maintained by maternal care. hibit significant cortisol elevations to events
Specifically, maternal licking and grooming that produce fearful behavior. If patterns of
and milk into the gut are the stimuli that buffer caregiving are important in maintaining the
the rat pup’s HPA axis ~Rosenfeld, Suchecki, relative hyporesponsiveness of the stress sys-
& Levine, 1992!. If these stimuli are removed tem early in development, then identifying the
for a number of hours, the buffering mecha- ingredients of stress buffering caregiving will
nism is disturbed and large increases in corti- be necessary to translate the basic research to
costerone, ACTH, and CRF can be observed. preventive interventions. Consistent with Do-
If there is a functionally equivalent SHRP in zier’s Attachment and Biobehavioral Catch-up
humans, are there comparable social stimuli intervention ~Dozier, Peloso, Sepulveda, et al.,
that help to maintain the HPA system in its in press!, we have examined whether the
relatively buffered state? In rodents, the role caregiver’s sensitivity and responsiveness af-
of maternal stimulation in maintaining the fects the development of stress neurobiology.
SHRP was identified by removing maternal Parents who were low in sensitivity and re-
stimulation for periods of time and then, for sponsiveness during medical exams when their
some pups, replacing elements of maternal care children were 2, 4, and 6 months of age had
during separation. Brief periods ~up to several children who as toddlers exhibited larger cor-
hours! of separation do not induce elevations tisol responses to childhood immunizations
in corticosterone or ACTH in rat pups, but ~Gunnar, Broderson, Nachmias, Buss, & Rig-
after 6- to 12-hr elevations are observed. Ar- atuso, 1996!. Recently, Hane and Fox ~in press!
tificially providing licking and grooming and extended these findings to include measures
milk maintain the SHRP ~Suchecki, Rosen- of frontal EEG asymmetry. Right frontal EEG
feld, & Levine, 1993!. Needless to say, there asymmetry is associated with withdrawal emo-
are challenges in examining parallel phenom- tions ~e.g., fear, sadness! and risk for anxiety
ena in humans; however, cortisol levels for and depression ~Davidson, 2002!. Hane and660 M. R. Gunnar, P. A. Fisher, and The Early Experience, Stress, and Prevention Network
Fox ~in press! found that mothers who were and prolonged glucocorticoid infusions re-
low responsive had infants who exhibited more model the dendrites of the mPFC ~Brown, Hen-
right frontal EEG asymmetry and were also ning, & Wellman, 2005!. As noted earlier, the
more fearful, whereas the infants of high re- mPFC plays a significant role in regulating
sponsive mothers exhibited a left frontal EEG behavioral, endocrine, and autonomic re-
pattern and were more bold. Studies with rhe- sponses to stressors ~Sullivan & Gratton, 2002!.
sus monkeys have demonstrated that greater Relatively few studies of maltreated chil-
right frontal EEG patterns are not only asso- dren have employed neurocognitive tests that
ciated with greater behavioral fearfulness, but can identify specific neurological deficits ~Pol-
also with higher cortisol reactions to psycho- lak, 2005!. General cognitive impairments have
social challenge ~Kalin, Larson, Shelton, & been associated with neglect, with abused chil-
Davidson, 1998!. Taken together, these find- dren sometimes performing more compe-
ings suggest that in early childhood, at least, tently than neglected children ~Pears & Fisher,
caregiver sensitivity and responsiveness may 2005!. Among abused children with posttrau-
play the role that maternal licking and groom- matic stress disorder ~PTSD!, both general def-
ing in rodents does to maintain a relatively icits and deficits in executive functions have
buffered or hyporesponsive neuroendocrine been noted ~De Bellis, 2005!. These deficits
stress system. are similar to ones reported for nonhuman pri-
mates and rats reared under conditions of so-
cial isolation ~see for review, Sanchez et al.,
Corticolimbic level of stress organization
2001!. Social, as opposed to stimulus, depri-
Although early experience research in rats fo- vation may underlie these effects, as only so-
cused initially on development of the HPA cial deprivation in rats has been shown to
axis ~Levine, 2005b!, recently there has also influence functions associated with the mPFC
been interest in the impact of early experi- ~Schrijver, Pallier, Brown, & Wurbel, 2004!.
ences on development of the mPFC and stress- Notably, few studies of maltreated children or
mediating corticolimbic circuits. Repeated nonhuman primates have observed problems
separations in rats that result in increased vul- in memory processes associated with hippo-
nerability to stress also affect the develop- campal functioning, problems frequently noted
ment of the mPFC. Rat pups exposed to in rodent studies of maternal deprivation and
disturbances in maternal care exhibit lower attributed to disturbances in regulation of the
levels of neurotrophins that support neural plas- HPA axis. Of particular note for abused and
ticity, with decreases most pronounced in the neglected children is evidence that early ad-
prefrontal cortex ~PFC; Roceri et al., 2004!. verse care increases the risk of inattention and
Maternally deprived pups also exhibit defi- overactivity ~Kreppner, O’Connor, & Rutter,
cient attention, particularly problems with set 2001!. These problems are associated with dis-
shifting tasks that are dependent on the mPFC turbances in frontostriatal circuitry ~Casey
~Lovic & Fleming, 2004!. In addition, they et al., 1997!. Problems on tasks subserved by
display alterations in the responsiveness of frontostriatal circuits are noted for children
mesocortical dopamine neurons to stress and neglected early in life even after, through im-
psychostimulants ~Brake, Zhang, Diorio, provements in care, the children’s general cog-
Meaney, & Gratton, 2004!. As with work on nitive functioning has returned to the normal
the HPA axis, these effects appear to be asso- range ~Bruce, Tarullo, & Gunnar, 2005!. Im-
ciated with maternal behavior, particularly ma- aging studies also tend to support impacts on
ternal licking and grooming. The co-occurrence the development of the PFC in maltreated chil-
of problems in mPFC function and stress re- dren. In addition to overall reductions in brain
sponsiveness likely reflect interactions be- volume, reduced white matter in the PFC and
tween neuroendocrine stress systems and corpus collosum have been noted ~see for re-
frontal functioning throughout development. view, De Bellis, 2005!. In at least one study,
The PFC is also a target of glucocorticoids, particularly marked disturbances were re-
and there is evidence that both chronic stress ported in the mPFC. Imaging studies of so-Stress neurobiology and prevention science 661
Figure 3. The average amplitude ~ mV! of feedback negativity across correct and incorrect trials by
group: means and standard errors. Adapted from Fisher et al. ~2006!. ERPs assessed at Cz to a flanker
task for children in regular foster care ~RFC!, therapeutic foster care ~TFC!, and community compari-
son ~CC! conditions. Therapeutic foster care was the Early Intervention Foster Care Program ~Fisher
et al., 2005!.
cially deprived rhesus infants have reported ner & Petersen, 1990!, and is a component of
similar findings, particularly with regard to the mPFC network with extensive reciprocal
reduction in white matter volume in the fron- connections to the amygdala and hippocam-
tal and parietal cortices and corpus collosum pus and outflow pathways to the HPA and
~see for review, Sanchez et al., 2001!. sympathetic nervous system ~SNS!. Using a
What is not clear is whether disturbances flanker task, slowing of responses following
in mPFC functioning are produced by similar an error and an early negative component in
mechanisms to those producing disturbances the ERP in response to error feedback were
in behavioral, autonomic, and neuroendocrine examined in children who had been randomly
responses to stressors. What is apparent from assigned to treatment versus regular foster care.
the above studies is that translational research They were compared to a similar group of
should incorporate neurocognitive assess- lower income nonmaltreated children. The in-
ments of specific prefrontal regions into as- tervention involved supporting the foster
sessment protocols ~Pollak, 2005!. Consistent parent’s ability to manage the behavior prob-
with this suggestion are data from a small lems of their charges without hostility and in
pilot study using event-related potentials ways that allowed supportive relationships to
~ERPs! to examine the impact of a foster care be established between the foster parents and
intervention on activity of the ACC ~Fisher, child. Children who had lived in treatment
Martin, Bruce, & Fox, 2006; see also Dozier’s foster care, like the comparison children, ex-
Attachment and Biobehavioral Catch-up Inter- hibited a significant error feedback-related neg-
vention Protocol, Dozier, Peloso, Sepulveda, ativity in their ERPs ~see Figure 3!. This was
et al., in press!. The ACC plays a critical role not noted among the regular foster care chil-
in effortful control of attention and action ~Pos- dren. These data suggest that treatment foster662 M. R. Gunnar, P. A. Fisher, and The Early Experience, Stress, and Prevention Network
care had helped normalize this aspect of mPFC Association of Adverse Early Care With
functioning in these maltreated, foster care Stress Responding in Adulthood
children.
The rodent model demonstrates impacts of
early adverse care on stress neurobiology as-
Summary sessed in adulthood. If the rodent model is to
Four key bridging themes were identified: tim- help guide intervention work, then adverse
ing, the SHRP, caregiver mediation of the early care in humans should be associated with
SHRP, and impacts of caregiving on prefron- heightened adult stress responding. However,
tal functioning and corticolimbic levels of human studies are invariably difficult to inter-
stress organization. The timing of events rel- pret because they must be based on experi-
ative to brain development in the rodent and ments of nature and thus lack the rigor of the
human at birth suggests that the prenatal pe- animal research. Therefore, as a bridge to hu-
riod in humans should be included as a focus man development, we will first turn to nonhu-
of translational research on adversity and stress man primate studies.
neurobiology. In addition, consideration of data
on relative stress hyporesponsiveness sug- Nonhuman primate outcomes
gests that we should extend the lens beyond
infancy to include most if not all of childhood. In nonhuman primate research ~see for re-
Evidence that the long-term effects of adverse view, Levine, 2005b!, researchers have exam-
caregiving in the rodent reflect specific as- ined a number of different adverse early life
pects of parental care that buffer the HPA axis experiences, ranging from separating the in-
during early development encourages attempts fant from the mother ~and sometimes also the
to identify elements of parental or caregiver social group! to rearing the infant under con-
behavior that may serve similar buffering func- ditions in which it receives little of the care
tions during human development. Currently, typical of the species ~i.e., rearing on cloth
the best translation of these rodent findings surrogates or only with other infant mon-
appears to be in measures of parental sensitiv- keys!. Typically, outcome measures were ob-
ity and responsiveness; however, the lens on tained when the animals were older, but still
parenting behavior may widen as more stud- juveniles. These adverse early care experi-
ies are conducted that examine caregiving ences have been shown to impact behavior,
and the regulation of stress neurobiology in increasing fearfulness, reducing exploration
children. Finally, animal studies are begin- of novel environments, and decreasing social
ning to note the impact of early parental care status. Notably, relatively few consistent long-
on the development of prefrontal systems, in term impacts on activity of the HPA axis have
addition to their earlier focus on limbic– been reported. A few studies have shown long-
hypothalamic functions. These data are con- term increases in the production of ACTH or
sistent with studies with maltreated children cortisol; however, most studies have revealed
that have noted significant impacts on execu- normal to blunted activity of the HPA system.
tive functions, and imaging studies that have Moving to higher levels of the stress sys-
noted effects on development of the PFC. These tem, however, we find more evidence of hy-
data emphasize the value of including neuro- perresponsiveness following early adversity.
cognitive and imaging measures in preventive Specifically, there is evidence of increased
intervention work on early experiences and amygdala reactivity ~e.g., increased startle re-
stress in children ~see also Cicchetti, 1996; sponses! several years after repeated, unpre-
Curtis & Cicchetti, 2003!. We turn now to a dictable separations early in life ~Sanchez,
more focused examination of the extent to Noble, et al., 2005!. Disturbances in maternal
which human studies provide evidence that care in Bonnet macaques produces increased
early maltreatment has lasting impacts on the CRF in spinal fluid, sensitization of the NE
neurobiology of stress, and in particular on system, and behavioral sensitization to fear
activity of the HPA axis. stimuli, while resulting in normal to low cor-Stress neurobiology and prevention science 663
tisol levels ~e.g., Rosenblum et al., 2002!. Ex- ports of childhood experiences, and often fail
treme disturbances in early care ~e.g., isolation to control for current life stress ~see review by
rearing! have also been shown to produce sub- Heim, Plotsky, & Nemeroff, 2004!. Nonethe-
tle changes in the primate hippocampus, which less, these studies provide some useful guid-
may reduce the resilience of the hippocampus ance. One of the clearest results from these
and increase the risk of hippocampal atrophy studies is that the association between child-
in response to later insults ~Siegel et al., 1993!. hood abuse and the neurobiology of stress in
There is also increasing evidence that early adulthood is a function of whether or not the
adverse care in primates influences the devel- abuse is associated with adult psychopathol-
opment of the PFC. Thus, monkeys reared in ogy. We will, thus, consider outcomes for adults
isolation exhibit problems with tasks that in- with and without psychological disorders pur-
volve ventromedial regions of the PFC ~see suant to their childhood maltreatment.
review, Sanchez et al., 2001!. As noted, me- Turning first to individuals without psycho-
dial and ventral regions of the PFC and ACC pathology, by definition, these individuals are
have rich bidirectional connections with the resilient ~Kinard, 1998!. Across the various
amygdala, and appear to be involved in regu- studies of resilient adults, we find evidence of
lating behavioral, autonomic, and neuro- reduced activity of stress neurobiology. The
endocrine responses to psychosocial stressors CRF challenge test produces elevations in
~Sullivan & Gratton, 2002!. Glucocorticoid ACTH and cortisol. The magnitude of the
overexposure impairs these regions, poten- ACTH response is inversely proportional to
tially reducing stress and emotion regulatory the pituitary’s chronic or traitlike exposure to
competence. Overall, the nonhuman primate CRF ~see for review, Heim et al., 2004!.
data are consistent with the idea that early Chronic, high CRF exposure from the hypo-
adverse care will have its largest impact on thalamic CRF system results in downregula-
systems that are still developing. The HPA tion of CRF receptors in the pituitary. Resilient
axis is relatively mature at birth in most pri- adult survivors of child abuse produce larger
mates, and thus may be relatively protected ACTH responses in the CRF challenge test
from permanent changes introduced through than do healthy adults with nonabusive child-
variations in experience ~e.g., Levine, 2005b!. hood ~Heim et al., 2004!. This suggests an
In contrast, higher levels of the stress system atypically low chronic CRF drive in these in-
~e.g., amygdala, mPFC! that mature for lon- dividuals. Resilient adult survivors also pro-
ger periods after birth in primates may be more duce larger ACTH responses to a psychosocial
significantly affected. These primate data sug- stressor than do healthy adults with nonabu-
gest that in studying human postnatal expo- sive childhood; however, their cortisol and car-
sure to adverse care we may need to focus on diac responses are normal to blunted ~e.g.,
the extrahypothalamic CRF system, fearful, Girdler et al., 2003!. This suggests reduced
anxious behavior orchestrated by circuits in- sensitivity of the adrenal to ACTH. This has
volving the amygdala, and potentially the de- been confirmed in ACTH challenge tests ~see
velopment of circuits in the PFC that are for review, Heim et al., 2004!. Thus, individ-
involved in modulating, containing, and ter- uals ~notably women! who experience sig-
minating fear and stress responses. nificant childhood maltreatment but do not
develop mental disorders show low neuroendo-
crine responsiveness to stressors, with evi-
Human adults
dence of traitlike low levels of hypothalamic
The animal models of early adverse care em- CRF.
phasize deprivation or neglect. In contrast, all Differences of opinion exist in whether to
of the adult human studies in this area have view this low level of activity as a risk factor.
examined the sequelae of physical and sexual Heim et al. ~2004! suggest that sensitization
abuse. Furthermore, most of these studies have of the pituitary and counterregulation at the
methodological limitations in that they focus level of the adrenal in these women may pre-
only on women, deal with retrospective re- dispose them to hypersecrete CRF, result-664 M. R. Gunnar, P. A. Fisher, and The Early Experience, Stress, and Prevention Network
ing in depression and CRF receptor down- of the sensitivity of the adrenal. Here, com-
regulation if they experience significant life pared to adults with depression without
stressors in adulthood. On the other hand, it childhood abuse who tend to escape from dexa-
is possible that low CRF and HPA reactivity methasone suppression, depressed adults with
may have preceded rather than followed their histories of child abuse tend ~like adults with
early life exposures, protecting these individ- PTSD! to supersuppress in response to the
uals from developing emotional disorders in low-dose dexamethasone test ~see for review,
response to their adverse childhood experi- Heim et al., 2004!. However, this supersup-
ences. Future studies involving longitudi- pression may reflect unmeasured PTSD in
nal work may be able to sort out these two these depressed, abused women ~Rinne et al.,
alternatives. 2002!. Overall, studies using pharmacological
Depression and PTSD are two of the se- challenges suggest that childhood abuse plus
quelae of childhood maltreatment that have either major depression or PSTD in adulthood
been explored most frequently in studies of is associated with supersuppression of the HPA
stress neurobiology. PTSD and depression ex- axis combined with hyperactivity of central
amined without reference to childhood abuse CRF.
appear to share hyperactivity of the central The results of studies using pharmacologi-
CRF system at hypothalamic and0or extra- cal challenge tests do not mirror results ob-
hypothalamic levels ~Bremner et al., 1997; tained when psychosocial stressors are used.
Heim et al., 2004!. Chronic CRF drive on the Rather than the blunted ACTH response ob-
pituitary in both depression and PTSD leads served in CRF challenge tests, for both child-
to counterregulatory downregulation at the hood abuse survivors with major depression
level of the pituitary, resulting in blunted ACTH ~Heim et al., 2000! and those with PTSD
responses to CRF challenge tests. However, ~Bremner et al., 1997!, hyperresponsiveness
these disorders differ in the sensitivity of feed- of ACTH and cortisol have been noted. The
back regulation of the HPA axis. Depression dissociation between the hyporesponsiveness
among adults is associated with reduced sen- when the pituitary is pharmacologically chal-
sitivity of negative feedback mechanisms, re- lenged with CRF, and hyperresponsiveness
sulting in larger cortisol responses to stressors, when psychosocial challenges are imposed
higher basal levels especially late in the day, clearly implies upregulation of the corticolim-
and reduced suppression of the axis in re- bic stress and emotion circuits in response to
sponse to dexamethasone ~Heim et al., 2004!. childhood abuse. As with the resilient adult
In contrast, PTSD is associated with increased survivors, we have a chicken and egg prob-
sensitivity of negative feedback mechanisms, lem. We do not know whether these findings
resulting in blunted cortisol responses to stress- reflect the effects of child abuse or explain
ors, lower basal cortisol levels especially late why these individuals developed affective pa-
in the day, and enhanced suppression of the thology in response to their early maltreat-
axis in response to dexamethasone ~Yehuda, ment. Again, longitudinal work with maltreated
2000!. The question is whether affective dis- children is needed to untangle the direction of
orders pursuant to childhood maltreatment fol- effects.
low these neuroendocrine patterns.
First, considering depressed women with a
Developmental Studies of Stress
history of childhood abuse, their patterns of
and Maltreatment
ACTH and cortisol responses to CRF chal-
lenge is comparable to those observed in Alterations in stress neurobiology observed
depression and PTSD without early abuse under conditions of adverse care may reflect
~Bremner et al., 1997; Heim et al., 2004!. Spe- transient adjustments that will remit to normal
cifically, blunted ACTH and normal to blunted functioning once the child’s care improves.
cortisol responses have been noted. This find- Transient adaptations that permit mainte-
ing is consistent with chronic CRF drive. nance of viability in response to adverse care
ACTH stimulation tests permit examination reflect allostasis or the maintenance of stabil-Stress neurobiology and prevention science 665
ity ~homeostasis! through change ~McEwen, hypocortisolism is a likely consequence of
2000!. Even when these allostatic adjust- chronic stress ~Friese, Hesse, Hellhammer, &
ments remit under conditions of improved care, Hellhammer, 2005!. This view is consistent
the processes of neural development they in- with mounting evidence that children living
fluenced may impact the child’s subsequent under conditions of maltreatment tend to ex-
development. This can be considered a devel- hibit low early a.m. levels of cortisol and a
opmental version of the costs associated with relatively low pattern of cortisol production
allostasis ~McEwen, 2000!. In designing pre- over the day ~Gunnar & Vazquez, 2001!. In
ventive intervention research, thus, we need our research network, we have observed low
to consider both the patterning of stress re- early a.m. cortisol levels among young chil-
sponding observed for children during peri- dren living in an orphanage in Russia ~re-
ods of adverse care as well as patterns and viewed in Gunnar & Vazquez, 2001!, toddlers
sequelae noted for maltreated children follow- within a month of adoption from Russian and
ing improvements in their care. Chinese orphanages ~Bruce, Kroupina, Parker,
As in the adult work, more information is & Gunnar, 2000!, and both infants and pre-
available on children who have been physi- schoolers removed from their homes and
cally and sexually abused than on children placed in foster care ~Dozier, Pelsos, Gordon,
who have been neglected but not abused— et al., in press; Fisher, 2005!. Among the chil-
although many abused children also suffer ne- dren recently removed from conditions of ne-
glect ~De Bellis, 2005!. In addition, we have glect and abuse, between 35 and 40% exhibit
very few studies of children’s responses to these abnormally low a.m. levels ~see Fig-
psychosocial stressors and even fewer studies ure 4!. One of us ~Fisher! has examined the
that have used pharmacological probes. Most abuse histories that differentiate preschool-
of our information on children, therefore, aged children with low a.m. levels assessed
comes from measures of ambulatory cortisol within a month of foster placement. Neither
assessed at various points in the day at home physical nor sexual abuse predicted this pat-
or when the children are in group care set- tern; instead the best was “failure to provide,”
tings. Although some of the studies have spe- a measure of neglectful care. This finding is
cifically sampled children with depression or highly consist with evidence of low a.m. cor-
PTSD, others have selected children based tisol levels among institutionalized infants
solely on their maltreatment histories. Over- and toddlers, children whose care has been
all, the study of stress neurobiology in mal- described as “institutional neglect.”
treated children is a relatively new area, and As noted, it seems likely that these low a.m.
thus much is unknown or uncertain. Despite levels reflect a transient downregulation of
this, several themes appear to be emerging the axis in response to frequent or chronic
that may provide some guidance in preventive CRF drive and elevated glucocorticoids. Abuse
intervention designs. as well as neglect might have this effect. Two
studies with rhesus infants also tend to sup-
port this argument. Sanchez and colleagues
Low early a.m. cortisol may reflect ~McCormack, Maestripieri, Plotsky, & Sán-
ongoing neglect and abuse chez, 2003!, have studied the infants of rhesus
mothers who spontaneously abuse their off-
As described above, counterregulatory mech- spring. Abuse is most frequent in the first 2
anisms result in downregulation of hypotha- months of life, followed by rejection and ne-
lamic CRF in response to frequent elevations glect in subsequent months. During the first
in glucocorticoids ~Makino, Gold, & Schulkin, month of life, early a.m. cortisol levels are
1994! and downregulation of CRF receptors elevated in these infants relative to infants of
in the pituitary in response to chronic CRF supportive mothers; however, beginning in the
drive ~see for review, Heim et al., 2004!. Thus, second month a.m. levels are suppressed be-
although hyperresponsiveness of the axis is low those of typically reared rhesus infants. A
often viewed as synonymous with acute stress, similar pattern has been noted for rhesus in-You can also read
