Transcript: Maternal-Infant Bonding

MOD: Good afternoon everyone and welcome to the inaugural OppNet Symposium Series. I want to thank you all for coming and I also would like to thank the wonderful planning committee that put this symposium together for today. And particular to say thank you for coming to the first seminar on maternal-infant bonding. For those of you who are new to OppNet, I just want to remind you that Dr. Francis Collins, our director of the NIH, instituted and announced in November, I guess 2009, November 18, 2009, the establishment of the basic behavioral and social sciences opportunity network, hence called OppNet, for the purpose of looking trans-NIH in the areas of behavioral and social sciences to get a better handle on how we could work together more effectively to expand research in this area.

And that we could come together as an agency to actually provide more funding and more research for the behavioral and social sciences. We all recognize that this research is very important and that it’s necessary, it’s actually essential, to further our understanding of the fundamental mechanisms and patterns of behavioral and social functioning, particular how they relate to health, disease prevention, and of course the overall wellness of individuals. As most of us are aware, behavioral and social sciences interact with each other. The sciences interact with each other, they interact biologically with biological factors as well as physiological factors, and they interact with the environment.

It is essential that we get a better handle on any new or any creative approaches that will help us to reduce risky behaviors that will help us get a better handle on how we can prevent disease and again, promote wellness. Over the past several months, 13 or so to be exact, the various steering committees including the coordinating committee and the OppNet steering committee have worked, we all have worked very aggressively to assess really what’s going on at NIH now in the behavioral and social sciences, particularly looking at our portfolio and getting a sense of what’s existing there where we are funding, but most importantly, where are the gaps?

Where are the challenges? And where are there areas in which we as an agency and cross institutes can collectively meet some of the unmet needs as they relate to the behavioral and social sciences? And most importantly, how we can find better ways to collaborate and coordinate our research efforts. The importance of human and animal models in all of this discussion was very, very critical to our working groups and their thinking. So what we have planned to do throughout this symposium series is really to have the researchers and investigators who are involved in the presentations to give us a sense of animal models and human models.

And what are the ways that we can look at these two approaches to behavioral and social sciences such that we can highlight differences, how we can look at similarities between what we find in the two models, and in certain cases, actually be able to discern whether there are paradigms for the human situation. Hopefully when we hear our presenters today and get a better handle of what we’re hoping for them and hoping to hear from them as it relates to maternal and infant bonding, it will spark a lot of discussions. Not only in this seminar, but in future seminars as it relates to how these animal-human models can be used, what we know about them so far, because there are many of us who have worked in animal models and also worked in human models and we know there are paradigms in which we can integrate them, but most importantly, we want to know about future research opportunities that these two approaches to research can lend themselves.

We have three speakers today, and again I want to commend the planning group for their selection of a wonderful trio of speakers. And they will discuss, as I’ve described to you earlier, the various models. We’ll have one who perhaps will present on animal models, and another who will present on animal and human models, and then we might even have one person who’s going to present on just human models. But I think they’re all going to be interactive in a very special way. The symposium for today is really focused on that one sort of first-and most important perhaps-experience that we have in our lives when we’re first born. And that is the bond between the mother and child or the mother and infant.

That early bond is we want to focus on today. This seminar series will begin with Dr. Megan Gunnar. She will begin with an overview of how early attachment relationships help infants. And here, we’re talking about monkeys, we’re talking about rats, and we’re talking about humans. How these early interactions help these infants regulate their psychological responses to stress, and until such terms that they have matured sufficiently, these infants, such that they can self-regulate, we will be depending on that relationship between mothers to help regulate our responses to stress.

She will also speak on how early social experiences can influence these processes over a period of time and as they emerge. Next, Dr. Steve Suomi will discuss his rhesus monkey research. Steve will be talking about how different early attachment experiences affect development at multiple levels of analysis. So Steve will be talking about behavioral expression and analysis of behavioral expression. He’ll be speaking about emotional regulation and its expression, physiological responses; he’ll be speaking at the level of brain structure and function, and gene expression, and particularly looking at gene-environment interactions and their interplay.

He will conclude, I hope, with some indication of the research he’s doing in his new studies focusing on early mother interactions and issues associated with infant imitation. Dr. Jeanne Brooks-Gunn will complete the trio of presentations with her work, and she will be speaking about fragile families. And this work that she will be presenting to us today is actually from the fragile families study. And this is a prospective longitudinal study that focuses on sampling individuals from well-defined communities and well-defined geographical areas, well-defined human populations.

She will look too at behavioral interventions, biological issues, genetic data, as well as looking at these entities at multiple levels. She will also present some of her initial findings regarding gene and environmental influences and looking at candidate genes and then referencing how these candidate genes and this research might have an impact on public policy. Her presentation will also consider parent-child interactions, and the thing that’s very interesting about Dr. Brooks-Gunn’s work is that she is really also interested in the families of today and the family structures of today. So she’ll be looking at how children interact with their parents in a community setting or in a home setting in which the child is born to a married couple.

She’ll be looking at the influences of the child and the parent in cohabitating parents, and then she’ll be looking at single parents and the role that a single parent-child relationship brings to child development. So you can see we’ll have a very full day. The speakers have said that they’ll take burning questions right after their presentations, but when we really get into the content of this overall discussion of maternal-infant bonding, I’m going to ask them all to come to the table and we’re going to moderate a very exciting and active and open discussion around the entire topic.

So if there are no other questions or concerns from the planning committee, I suggest we get started and I’d like to introduce our first speaker. Our first speaker, as I mentioned to you earlier, is Dr. Megan Gunnar. Dr. Gunnar is the Regents professor and distinguished McKnight University professor of child development at the Institute of Child Development at the University of Minnesota. She received her PhD in psychology at Stanford and completed a postdoctoral fellowship in developmental cycle endocrinology at Stanford Medical School. Professor Gunnar has spent her career studying how infants and young children respond to potentially stressful situations. With her students-and she’s a big proponent of students and training and postdoctoral development-with her research team, she has documented the powerful role relationships play in regulating stress in young children.

Her work has shown that children in full daycare show elevations in stress hormones over and throughout the day. In the last decade, much of her work has been dedicated to understanding the impact of early deprivation and neglect on children’s brain and their behavioral development. And this work has mainly been done through the study of children adopted from orphanages and children in the foster care system. Dr. Gunnar is a member of the National Scientific Council on the Developing Child. She has also served on the Board of Zero to three and was the National Academy’s reviewer of early brain and behavioral development that produced a very important document entitled Neurons to Neighborhoods.

At the University of Minnesota she directs the International Adoption Project and is the co-director of the Center for Neurobehavioral Development. Nationally, she directs a multi-university center that’s funded by the National Institute of Mental Health, and this work is on early experience, stress, and neural development. Dr. Gunnar.

MG: Well, I want to thank the organizers. This is a wonderful thing to be here for your first OppNet colloquium, and to be here with my friend Brook and Steve. So, knowing there is potentially a broad spectrum of researchers out there, I know some of you are developmentalists, but many of you maybe know what a child looks like. I’m going to start at the 50,000 foot level to get us thinking about what development is in those early years in life. So one of the things about the human-and actually many mammalian-brains is we know that genes are important, not the basic structure of the brain, they’re involved in determining the timing, the opening and closing of sensitive periods.

But without experience, without stimulation, the brain does not develop normally, right? We’re all on that page? Combine that with the fact that the human is one of the most helpless of all mammals at birth. We are born extremely immature motorically. The baby cannot lift its head up, get a hand to mouth, crawl, walk, sit up, grasp, get things to its mouth, for many months. So coupled with the fact that we need stimulation, the human infant is unable to provide itself with the stimulation it needs for normal neurodevelopment for many, many months. Of course in a relationship context, the baby receives rich stimulation: tactile, auditory, proprioceptive, linguistic input, in that kind of relationship context with both parents and with other important social relationships in that child’s early life.

So in a relationship context, there’s plenty of stimulation that the baby needs to organize the development of the brain, and the infant comes equipped with very powerful social signals that in responsive relationships allow the child to communicate its need to have negative adverse stimulation removed or fixed and very soon, within a month or two, develops the social signals that are capable of keeping people providing interesting stimulation, and when they cease to be interesting, we get a signal on the left and then we work very hard to provide the interesting stuff which turns that signal into the signal on the right.

So in responsive contexts, this extremely motorically helpless organism runs entire households and is able to regulate the stimulation it needs, pacing it at the level that neural systems are developing. It is what becomes interesting; working on that next thing. Critically important, then, for the human brain to have responsive continuous relationships in order to develop normally. We also know from some brilliant work that really began 20, 30 years ago from Myron Hofer, who was studying rodents, mother and infant pups, that embedded in relationships are the homeostatic regulatory components of the infant’s immature regulatory systems.

So Myron was able to show very nicely that when you remove the mother from the pup, within a few hours, of course you saw problems with temperature regulation, distress vocalizations, we could go on through the infant’s regulatory systems. But that if you supplied certain pieces of the mother back, you could regulate some but not all of these systems. So embedded in these interactions were the extension, the adult part of the immature regulatory system. And we think that is also true for humans as well. So embedded in these close relationships are temperature regulatory systems. The mother, the parent, becomes part of temperature regulation.

We know certainly part of the regulation of nutrient intake that Elliot Blass a few years ago demonstrated that feeding actually connects with opioid mediated pain regulatory systems in the infant’s brain. And of course, we have soothing interventions that calm down heart rate, and when the baby is under-aroused, we have stimulatory regulation that increases arousal, and this goes on 24/7. Well, thankfully not entirely 24/7 after a while. Day after day, month after month, as the infant’s immature regulatory systems are becoming increasingly more mature. So the external regulator is toning, we believe, the activity of these basic yet immature regulatory systems.

They’re developing in the context of this external regulatory process with one or two caregivers, and as the child’s regulatory systems mature, they take on characteristics that were part of this history of dyadic regulation or triadic regulation over time. Okay, that’s the basic underlying concept of what many of us are working on, is trying to understand from that 50,000 down to the molecular level of what’s actually happening in the course of these 24/7 boring everyday interactions that are a part of the baby’s first year or so and until the child is able to take over many of those regulatory functions themselves.

Today, to demonstrate this, I’m just going to focus in one important regulatory system: the hypothalamic pituitary adrenocortal system that produces corticotrophin releasing factor, which also has interactive effects, and ultimately produces glucocorticoids, cortisol in humans from the adrenal cortex. Glucocorticoids are gene transcription factors that produce, as Ron DeKloet who is well-known in this area, says slow but persistent effects on the development of neural architecture. Cortisol and corticotrophin releasing factor or hormone are critical in stress and adaptation, as their interaction plays important roles in physical and mental health throughout the lifespan.

And we have 60 years of preclinical data, minutely and beautifully analyzing in the rodent-and in the mouse now-how interactions between the mother, the dam and her pups, early in life, is shaping the reactivity and regulation of the hypothalamic-pituitary-adrenal axis in ways that carry forward into the adult activity of that system for the rodent. So we know also of course, I have to say, that cortisol is not the end all and be all of all stress regulatory hormones. The cortisol operates within a dynamic, biphasic, and nonlinear set of relationships with other stress mediators, and together the interaction of these things influences cognitive development, metabolic health, cardiovascular health, and immune function.

And the brain is the grand orchestrator of all of these interactions. Now my problem is wanting to take what we have come to understand in our preclinical models and apply it to human models. That is, I need noninvasive measures if I’m going to study babies from birth through three, four, five years of life. Which means that there is just a lot of this I can’t get my hands on. I can measure cortisol in spit, I can measure cardiovascular activity, et cetera, heart rate and parasympathetic, and I can tell you some of those stories today as well. But getting up into the grain and getting to many of these other factors, they’re not something I can do with health, developing children, because you wouldn’t bring your kids in if I was going to bleed them every three days, right? Okay.

If we’re really going to understand the development of these systems in humans and use our animal literature well, we need to do this in the concert with folks who are continuing to study models in smaller animals. We need a translation element here because our friend the rodent, its brain is born much younger than our brain. Although it’s also as motorically helpless as we are, but there are situations that differ. So as we come to understand what’s happening in our small animal models, we need a transitional object, as Freud would say, which for some of us is the non-human primate, and you’ll hear some of that work today.

And we need to have the researchers who are doing all of this work working in concert together, passing information back and forth so that we can guide, on the human side, the shaping of the way the experiments are being done by what we’re finding and we can get ideas for what we should be looking at from our friends who are working with these animal models. As part of the NIH roadmap—gosh I almost said roadblock, I bet you guys say that too, right? Roadmap; there have been a lot of efforts across the NIH to create these interdisciplinary teams of researchers. In the National Institute of Mental Health, one of the sets of centers that they’ve worked on establishing are these interdisciplinary developmental science centers, and as Yvonne said, I’m the director of one of them: Early Neurodevelopment Center.

Steve is actually one of the faculty in my center. And so I’m going to be bringing forward some of the results across that center when it’s appropriate today; not just my own research. Now I’m going to teach you a bit about what we have learned about the development of the HPA axis in humans. Much of the first part of this talk deals with the work I did early in my career, as we were just mapping things out, and then we’ll get more recent. So one of the striking things that became obvious to us very early on in doing this work is that the HPA axis becomes less responsive over the first months of life. I often used my friends in the medical community to give me stressors that were more intense than I could pose ethically in my lab.

So these are longitudinal work; its babies experiencing their well baby exam with shots they get at two, four, six, and at my clinic, 15 months. At two months, and this is true at birth, and probably up to about three months, we have a very reactive system. At birth, if I undress, weigh, and measure a baby, I can elevate cortisol. You’ll look cross-eyed at the baby when you elevate cortisol at birth; it’s a very reactive system. It’s also a critical system for the maturation of lungs and tissues. So it remains very reactive. By four to six months, it’s less responsive. We think this is from the development of fast negative feedback systems. We will need our animal friends to actually pin that down because we can’t do to kids what I really need to do to tease that out.

And then there is this remarkable thing that happens, probably somewhere between six and 12 months, that in children who are being reared in their families, at least we see almost a complete buffering of response to being undressed, weighed, measured, poked, punched, and jabbed in the legs with needles. We think of these as somewhat stages of the development of reactivity. So I’m going to take you first to this initial stage here where the system is extremely responsive. And I’m actually going to draw from some research that isn’t my own. It’s new; we had done things earlier but it’s nice to be replicated. So this is a study by Esther Albers that was published just a few years ago. Babies like their baths, right?

But when you take a baby out of the bath, there’s a rapid drop in temperature and skin temperature. That is a sort of a biologic threat to the organism. The HPA axis should kick in, but the question is how rapidly does it come back to baseline? Albers not only assessed reactivity and regulation returning to baseline of the axis, but she had observed the mothers and infants and scored them on how sensitive and responsive the mother was to the child’s signals. She then grouped them into high and low responsive and she obtained results similar to things we’ve seen. Here, the system responds to that sudden drop in temperature. In high sensitive and responsive pairs, we see a very rapid return to baseline.

Not so much if the mother is less sensitive or responsive. This is a normal daily event. You can imagine that for those several months, many normal daily events; systems reacting and being regulated differentially by a function of the quality, the sensitivity and responsiveness of care. And the idea is a steady drip, over and over. This is being now moved into the way this system is going to need to operate. We have tracked over time, again using our well baby exams, whether high or low responsive care-giving in these first six months of life predict trajectories of regulation. These are T-scores, because of course everybody is basically coming down during this time, so it’s relative to others.

These are average to high responsive parents, and the system is becoming increasingly regulated. These are low responsive parents, and over time relative to these children are becoming more reactive, okay? Over time. That’s an old study of ours. This is a recent one by Clancy Blair studying children. Low income rural children, over 1,000 kids of scoring maternal sensitivity and responsiveness in the first year, six months of life, looking at response to emotional challenge tests, again and again, imposing them at 15 months. We see some patterns here but they’re not as clear as a function of high and low maternal responsiveness.

But now by the second year, the kids are filling out in terms of how responsive their history has been, how sensitive or responsive their care has been, and this is in low income. So we’re seeing the story play out in different settings. Now one of the real problems, it’s really annoying, of studying parents and children, is that they’re often genetically related, which makes it truly challenging to sort out how much of these patterns might be due to shared genes and what might be due to what we think is the operative factor, in this case sensitive and responsive care-giving. But of course, children are not only cared of by the people that they’re biologically related to.

Children spend a good time, in our culture, in the care of people that they’re not biologically related to. Childcare is a normal phenomenon now, for kids, and as Yvonne indicated, a number of years ago, we were stunned because I never expected this. I was studying temperament and I wanted to study kids in childcare so I’d see them interacting with one another. We were stunned to find that in full day out of home childcare, and this is true whether its center or family-based care, the youngest children, these are one and two year olds, show actually an increase in cortisol across the day. What you’re seeing here is our delta scores. The children’s mid-morning and mid-afternoon at childcare compared to their own levels on days where they don’t go to childcare.

Morning, not so different; this is not a separation response. But by mid-afternoon, three or four o’clock, we see marked elevations in the system over what they would be at home. Three and four year olds are a little better able to hold it together. And by five and six and seven and eight, you can be away from home all day and you don’t produce these elevations. We’ve been on the hunt for what aspects of the child or the environment are driving this, and certainly age is a huge one, right? As children develop the capacities to engage in better behavior regulation, they seem to be able to last longer away from home. But we’ve always strongly suspected that the quality of care is going to matter as well. And what we’ve looked at is the same thing that we’ve looked at with parent-children.

The sensitivity and responsiveness of the care provider to the child’s signals on the positive side, her intrusiveness and over controlling behavior on the negative side. That’s what quality care refers to in this process sense. And here’s one example from us, but other people have seen the same. These are three year olds. They’re actually in family-based childcare; that’s where you go and stay with the lady in her home. These are the top and bottom third of the distribution for how sensitive and responsive the quality of care the care provider is giving. So morning to afternoon, top third of the distribution for sensitive and responsive care, very little increase.

Bottom third, that’s where our real significant stuff is happening. This is AM to PM childcare; this is afternoon childcare relative to home on a non-childcare day. So whichever way we look at it, the pattern is the same. Again now, other relationships also play in a role perhaps in toning the activity of this system. Does it matter? That’s really what I’ve been on the hunt for ever since we first came after this. This is a measure of how temperamentally shy and fearful the child is, through observations in the laboratory and reports of childcare and parent provider. Temperamental fearfulness we strongly suspect has a hereditable component. We’ve got a bit of evidence that it does.

And we’ve always suspected that the children that would be the most influenced by quality of childcare, in fact Belsky has some data like this, or the elevation of cortisol, would be these shy, temperamentally shy and fearful children. So what you’re seeing here is time two measurements of internalizing anxiety and depression symptoms, which is what these kids should be developing in the less supportive kind of context. And the top third and the bottom third of the distribution is just for plotting purposes of temperamental fearfulness. Children whose childcare centers at time one were producing stable across the day or decreasing levels of cortisol, normal diurnal rhythm. No elevation or increasing levels.

And what you can see is it doesn’t matter much for internalizing if you’re not shy, anxious, and fearful. But if you are, it makes a critical difference. And this is not just a stress diathesis model. Temperamentally fearful children, in daycare contexts where they aren’t elevating, appear to be becoming less anxious. They’re protected. They’re learning the skills to play with other kids without getting stressed out. So, I’d better move. Everything I’ve shown you is the normal range of care-giving. There’s nothing maltreating in what I’ve shown you. And what I want to do is shift now to care-giving systems that are broken. And if we go back to our animal models, this is where our animal models provide us with the most information.

Our rodent models, we don’t have models of sexual and physical abuse in the rodent. I [unint.] would never allow you to do that anyway what we do to kids all the time unfortunately. And in non-human primates, that’s where we have our models. Models of peer-rearing, which Steve will probably talk about, or surrogate rearing, where we moved the regulator. These are deprivation models. In humans, children who are severely neglected in their families of origin, yes, but what a messy model because life keeps being difficult for those kids. So I have turned my attention to studying children who start out their lives living in orphanages or other institutions, where the quality of care means that they really don’t have their own emotional regulator or physical regulator person, and then our adopted by families.

So they have this sudden shift into a supportive home, at different points of course, of development. So that’s been a model that I’ve been focusing on; I’ll bring you some results from there. We know that institutional rearing is not good for human development. We know that it produces delays and deficits in all domains of development and in the studies that have tracked healthy babies going into institutions and followed them over several years you simply watch the fallout on development of skills over that period of time. Kids make unbelievable rebounds once they get placed in families. This has also been demonstrated for the last half century. But we know that some of them are at risk for problems in social cognition and social functioning, attention regulation and executive function, big signal of having started out in an orphanage.

Emotion regulation and we’re beginning to work on the neurobiological correlates of different behaviors that are disturbed so we can get a better understanding of what has happened as that brain has tried to put itself together in a depriving context. Now, we got our first hint of what was happening to the HPA axis in a study that I helped design. I actually went over and served as one of their research assistants in ’97 because they didn’t know how to collect cortisol. These are 46 kids living in an orphanage in Iasi, Romania. They are two years of age, two to three years of age. These are family reared kids in Iasi, Romania. This is a normal diurnal rhythm in the HPA axis. Not one of the 46 kids had a normal pattern. We saw a consistent lowering of the first wakeup value and an inability to bring cortisol down over the day to the lowest level.

This kind of atypical, flat pattern we now know from animal research, is at least in adult organisms can be produced by chronic stress. And then it pops back once the stressor is removed. We see somewhat of the same thing happening when kids are adopted. Here, two months into adoption, we still see it still hasn’t gotten to be quite a marked diurnal rhythm relative to US family reared children in Minnesota. I call them Garrison-Keeler kids for those of you who get the joke. And we know that given another six months or so, we’re going to see, from most of the kids but not all of them, a normal diurnal pattern has been reestablished, though we’ll see some other glitches in how the system is operating. This is not just orphanages.

Our friends out at Oregon Social Learning Center, who are also part of this center that we operate, have been looking at preschoolers who are placed in a new foster placement, removed from the home because of maltreatment or neglect. Using community controls of the same socioeconomic class, they’ve identified what is one and a half standard deviation above and below the mean and grouped their kids accordingly. 62% of their kids had AM levels that are within that one and a half, plus minus. 4% have elevated, but then 4% of the controls would be high, too, given those. 34% have this low, flat pattern. They’ve been able to go back into child protection service records to figure out what it is.

It’s not the severity of physical abuse, it’s not the severity of sexual abuse. What is correlated with this is the severity of neglect. As many of us know, neglect is the understudied problem in maltreatment. Most children who are maltreated are neglected. There other forms of maltreatment, but neglect is a very powerful determinant of what happens to their development and is yet understudied. Okay. There’s a lot of variation in what happens once the child is sort of stabilized. We have been on the hunt for what will help us predict which kids will have pretty normally functioning HPA axis once they have a chance to get used to living in a low risk supportive context and where we’ll see some differences still.

And early on, we got curious about the relation between the growth system and the HPA axis. One of the ways that activation of stress can reduce and help you adapt is by reducing expenditure of energy on things that you don’t need for survival today but are future oriented. One of the most energy demanding things that young children do is to grow. And there are very intimate relations between the growth system and the HPA axis. Just to remind you, growth hormone releasing factor is the positive stimulus. Somatostatin is the negative stimulus. It’s sort of the brake on this system. It produces growth hormone, interacts with the liver, produces insulin-like growth factor 1 and interacts with the bones and muscles to produce growth.

Corticotrophin releasing hormone or factor increases somatostatin production, steps on the brake. And elevated cortisol interacts at the point of the liver to reduce its sensitivity to the growth hormone signal reducing IGF-1. Two spots where stress activity of the HPA axis will slow growth. Growth is one thing that we know falls off, along with behavior for kids living in institutions. One estimation is that they lose about one inch of linear growth. They don’t shrink, but relative to other kids, for every two to three months in an institution. And by the time many of them are adopted, particularly in relation to more severe deprivation, they are in the range of what we would call growth stunted. More than one-and-a-half standard deviations below the mean.

Of course this could be because they’re not getting fed well, but we actually know from a good deal of work that if you assess the food that they’re being fed, in many orphanages it’s just fine. And we also know that you can provide stimulation and affection and not change the amount of food that they’re getting and they’ll start to grow. So their growth failure follows a pattern that we have termed, that others have termed psychosocial. Short stature, a decrease in bone growth, weight is proportional to height. And if anything, they put a little more fat around their bellies, which is actually a very classic stress model. And as I said, you increase the attention, you get increased growth.

Once these kids get to families, they grow like weeds, one and a half to two times kids their age. And almost most of them catch up. So when we started dividing our samples by whether the kids were growth stunted, one and a half standard deviations or more, we’ve done three studies. This is the third study where we’ve looked at that. These kids are six to seven years, post-adoption, six to seven years of age. Most of them have been adopted four years to five years. They’ve been in their homes for that long. I’ve divided the post-institutionalized kids into those who are not stunted, those who are stunted, and my non-adopted Garrison-Keeler kids as comparison.

So the non-stunted ones at adoption-note, everyone’s gained pretty good growth by now-just looks like Garrison-Keeler kids. I can’t tell them apart. But these that were stunted at adoption, we see a little lower still, and a little elevated here. We’ve also started taking this in to look at the behaviors. Does it help us differentiate behaviors? Indiscriminate friendliness is a behavior that we see in post-institutionalized children; not all of them but some of them. And we see it in severely neglected kids. This is this bizarre thing where you walk as a stranger into the room and the child might come over and say, “I’ve missed you,” or climb in your lap or play with your earrings or ask you if you are married and whether you’ve got a boyfriend or a girlfriend.

Okay? Indiscriminate friendliness, and it is a component, or we used to think it was a disinhibited form of detachment disorder. That’s not really believed anymore. Cognitive inhibitory control, performing ongoing tests, attention tests, thalamic cortical striatal circuitry. We’ve got good imaging data from BJ Casey on these tests and what we see is kids adopted out of orphanages score more poorly than non-adopted children on both of these. They’re more indiscriminate friendly and have problems with attention. But if we divide them into those who were stunted at adoption and those who were not, the burden is being carried largely by the kids who had this growth failure.

So what we’re on the hunt for is trying to understand the genetics of the prior experience that link these systems of growth, stress, and sort of neurocognitive effects which are all walking into prefrontal cortical areas. And finally, the levels of cortisol here at these late afternoon times predict a level of .05 both indiscriminate friendliness and problems in inhibitory cognitive control. In my last moment, I’m taking you one more place. Do I have time? Okay. And that is to our friend puberty. We’ve been talking about the early period of life as a period of great plasticity, where the experiences that the individual has is coming and being embedded in a sense under the skin into the individual’s regulatory and cognitive functioning.

We have another massive period of plasticity, and that’s the pubertal period. Sex hormones are plasticity hormones. They’re involved in shaping the brain and altering it, opening up systems potentially for the new experiences that are happening in adolescents. Puberty, we know, brings loads of changes in psychological functioning. The social context kids are in, physical growth, and in neuroendocrine activity. In typically developing kids, puberty is associated with an increase in the set point of the HPA axis and an increased HPA reactivity to social stressors. So it’s a period of both risk and opportunity. Plasticity means both.

It is a potential point in which children who have been in adverse conditions and now are living in good conditions might have the opportunity with these sort of plasticity increases to reorganize the context of what they’re living in now. So we’ve been very interested in looking at that. Last little piece of study, we’ve been doing a study where we’ve been looking at puberty. One way to do this is to hold age constant. So we’re looking at 12 and 13 year olds. This is that wonderful period of life where some boys are still three feet tall and the girls are like eight feet tall and well developed, so you get the full range of pubertal development almost in these two years.

We’ve identified children who are early or pre-pubertal, mid or late, same ages. This is typically in developing kids. Cortisol awakening response is a response that is actually on top of the normal diurnal rhythm; that’s a lot of the focus of research today. And you can see we’re getting that increase in set point and maybe a bit of an increase in the reactivity of CAR with pubertal development and typically with developing Garrison-Keeler low risk kids. Here’s what we see for our post-institutionalized kids. Early, pre-early. They have almost no CAR. This very different pattern. I don’t know if it’s just my stunted kids.

I suspect so, but we don’t have enough kids to pull them out that way. This is what we see for the ones who are mid to late puberty. There’s no difference. Now of course this is cross sectional and we’re in the process of doing the longitudinal work to see if they’re actually the same kids we’re reorganizing. This just shows you that if we look at low, medium, or high degrees of deprivation, it’s the high degrees of deprivation that show it to us in the pre-early group, but again, no evidence that this is different once we have puberty hit. So, relationships as regulators. They are important regulators of development. They get embedded under the skin.

They regulate all sorts of systems. I’ve only shown you one; you’re going to be hearing about many, many others in these other talks today. And I want to acknowledge the many people that have been involved in helping us get this work done. Thank you. How much time do we have? So one or two questions. I want to make sure my buddies have plenty of time; they have such cool data. You don’t want to ask me any questions. Yes? The question is do we see difference in the age of onset of puberty. The answer is yes. So one of the things that goes along with that stunting of growth and then the rebound, and we know this from animal models that have been malnutrition models, but it is final common pathways into the growth system, right?

You get an earlier onset puberty, especially for the girls, but that may be because it’s just a lot easier to measure in girls. It’s more obvious. We dealt with that in that study by specifically going out looking for the ones that were pre-early, et cetera. It’s not a whole ton of the kids. But when we see it, it’s about a year advanced, in some cases, maybe two. So there are some families dealing with eight year olds going into puberty whose sort of mental development age is not up there and it is a real difficult issue for them.

MOD: Our next speaker is Dr. Stephen Suomi. Dr. Suomi is chief of the Laboratory of Comparative Ethology at Shriver National Institute of Child Health and Human Development. I know that one. Dr. Suomi earned his bachelor’s degree in psychology from Stanford University and his PhD in psychology from the University of Wisconsin, Madison. Dr. Suomi later joined the psychology faculty at the University of Wisconsin where he obtained the rank of professor before moving to NICHD in 1983. Dr. Suomi’s initial post-doctoral research, and I think it’s important for me to mention this, successfully clarified the adverse effects of early social isolation, previously thought to be permanent in rhesus monkeys.

So he really clarified this for us. This early research led to his election as a fellow in the American Association for the Advancement of Science, and he was recognized in this membership for, and I quote, major contributions to the understanding of social factors that influence the psychological development of non-human primates. Dr. Stephen Suomi’s present research at the NICHD focuses really in three areas: the interaction between genetic and environmental factors and their shaping of individual developmental trajectories. He also focuses on the issue of continuity versus change, and the specific relationship and relative stability of individual differences throughout development.

And I’m very interested in hearing what Steve might say about that today. And he’s also focusing on the degree to which findings from monkeys or from the non-human primates, particularly those studied in captivity, generalize not only to monkeys living in the wild but also to humans living in different cultures. Throughout his career, Dr. Suomi has been the recipient of numerous awards and honors. His most recent include the Award for American Psychological Association, the Distinguished Primatology Award from the American Society of Primatologists, and the Arnold Pfeiffer Award from the International Society of Neuropsychoanalysts.

I am very pleased to introduce Dr. Steve Suomi, who is from the National Institute of Child Health and Human Development.

SS: Well thanks Yvonne for that wonderful introduction. It’s a real pleasure and a certain amount of pride being at the first inaugural meeting of this and getting a chance to talk about some of the research my colleagues and I have been doing, so nicely supported by NICHD over the years. My lab at NICHD, which is actually not on the main campus here in Bethesda but out at the NIH animal center near a little town called Poolesville about 40 miles up the Potomac from downtown DC in the Maryland countryside where we measure lab space in terms of acres instead of square feet, and this space allows us to keep many of our colony of about 350 rhesus monkeys in outdoor naturalistic settings year round.

So when I talk about the Poolesville monkeys, these are the ones that are out in the Maryland countryside. And the lab from its inception has had a major focus on the study of individual differences in personality or temperament among these monkeys and the biological substrates that underlie these individual differences. And in particular, how genetic and environmental factors act, and we think now much more importantly, actually interact to shape individual developmental trajectories. Over the years, we’ve been especially interested in two subgroups of monkeys, both in our colony at Poolesville, but also that we have seen at two field sites that we’ve had access to over the years.

The first subgroup consists of about 20% of the population, both out at Poolesville and at these field sites. And these are monkeys who seem to be unusually fearful or anxious in the face of novel or mildly challenging circumstances of stimuli. So stimuli that most monkeys find really interesting and will readily explore, these guys will try to avoid. And if they’re forced to confront these stimuli they show a pronounced arousal of a variety of biological systems. The system we pay the most attention to is the HPA axis that Megan was referring to, and we see this in terms of prolonged elevations of cortisol, whether we’re measuring it in blood or saliva or more recently in hair.

These monkeys are shy in social situations, they seem to be anxiety prone, and they look very much like a subgroup of human children. Again, at least in the US, 15 to 20% of the population that people like Jerome Kagan at Harvard and my dear friend Nathan Fox at Maryland have been studying for the past quarter century, these are children who have been called behaviorally inhibited by Kagan, and these are children who we now know are at substantial risk for developing childhood anxiety and depressive disorders. And for many of whom this risk carries on through adolescence into adulthood. There’s another subgroup of monkeys, maybe 5 to 10% of the population in Poolesville and also in the wild, who seem to be unusually impulsive.

And this impulsivity gets them into trouble. They do stupid things that other monkeys know better than to try, like get between a high ranking female and their offspring. Or they repeatedly confront a dominant adult male. And when that adult male beats them back, they’re right back in that male’s face. And when that male beats them back again, they’re right back in that male’s face. So not surprisingly, they tend to elicit an awful lot of aggression from other monkeys in their social group. And they themselves are excessively aggressive, often in inappropriate situations, and more importantly, once they get into an aggressive encounter, whereas most monkeys know how to back off before things get out of hand, when these monkeys are involved in an aggressive interaction, the intensity of the aggression tends to escalate up to the point where somebody can actually get hurt.

And these monkeys also have characteristic biological features that differentiate them from others in their social group. The feature we’ve paid the most attention to is apparent deficit in serotonin metabolism that we pick up in terms of chronically low levels of the primary central serotonin metabolite in cerebral spinal fluid 5-hydroxyindole acetic acid or 5HIAA. And just as is the case in some extensive human clinical work in the 1990’s, very aggressive humans and very aggressive monkeys tend to have chronically low levels of this CSF 5HIAA metabolite. These children, these monkeys, represent another subgroups of children that people like Richard Tremblay, my colleague in Montreal, have been studying for a number of years.

These are children who by the time they are two years of age have already identified as being excessively aggressive. And while most two year olds are already aggressive, this is why they call it the terrible twos, in most cases, the aggression goes down and it shows up in fewer situations with less intensity as the kids grow older. But that’s not true of this subgroup of human kids. They still have high levels of aggression. By the time they hit the school system they’re often causing major classroom disruptions, they’re often diagnosed as having various externalizing disorders, and many of them at least in the US, by the time they’re in their teens you can find them either in prison or in the morgue.

So in addition to these obvious parallels to human subgroups, we’ve been interested in these phenomena in our monkeys for a variety of reasons. First, in both cases, for both what we call our uptight monkeys and our impulsive aggressive monkeys, the behavioral and biological characteristics show up very early in life, usually within the first weeks. And in the absence of major environmental change, are remarkably stable from infancy through the childhood years, adolescence, and into adulthood. And we have increasingly solid evidence that many of these characteristics are highly heritable. So for example, among our shy monkeys, the individual differences in cortisol have a high hereditary component for our impulsive aggressive monkeys.

The individual differences in 5HIAA concentrations and measurement of serotonin metabolism are also highly inheritable. So genes are clearly involved and underlying at least part of these individual differences. But I’m here to tell you this afternoon that genes are not the entire story, and in addition, early experiences, particularly early experience with attachment objects or caregivers can also alter these developmental trajectories and the influence of these environmental influences or characteristics show up not only at the level of behavioral expression and emotional regulation but also at the level of neural hormonal output, of neural transmission, neural transmitter metabolism at the level of brain structure and function, and we now believe even at the level of gene expression.

Now in the wild, rhesus monkeys, who by the way are a very unusual species of primate relative to non-human primate species. They are not our closest relatives; they only share about 95% of the genes that we have, where by comparison chimpanzees and bonobos share between 98 and 99%. But rhesus monkeys are a lot more like us than are chimpanzees or bonobos or any of the other great apes in one fundamental respect. They are a success story. They are the world’s second most successful species of primates next to us. There are more rhesus monkeys living in the wild and canvassing a wider geographic range, experiencing a wider range of climates than any other species of non-human primates.

Another way in which they are unique is that like us, they are one of the very few primate species that can be genuinely called a weed species. That is, if you take them out of their natural habitat and put them somewhere else, they’re likely to thrive and expand their populations. Whereas this is hardly the case for other non-human primate species, many of which are threatened, endangered, or very close to extinction. They are also unbelievably resilient in the face of unusual and severe stresses, both physical, natural and social. And I could give you plenty of stories; I’ll just give you one. I assume everyone remembers Hurricane Katrina, this major storm that trashed New Orleans and the Mississippi Gulf Coast five years ago this summer.

What you may not know is 30 miles north of New Orleans on the northern shore of Lake Pontchartrain is the Tulane National Primate Research Center where they had, before the hurricane hit, 3,000 rhesus monkeys living in outdoor pens outside. This was their normal living situation at the primate center. When the hurricane came through, it blew away all the pens; just blew everything away. Yet, every single monkey survived. They did not lose one of the 3,000. We think we know what happened. In each of those pens the caregivers had put in a large concrete culvert, the sort of thing you use for drainage under expressways, so the monkeys would have a way to come out of the hot Louisiana sun.

Certainly what happened is when the hurricane hit, the monkeys all headed for those culverts, got down in the culverts, waited for the storm to come through, and then waited for the humans to show up the next day with their food. So these monkeys survived Hurricane Katrina much better than most of the human population as well. In the wild and in such captive environments as we are able to do in Poolesville, these monkeys typically live in large social groups called troops. The troop can range in size from 30-40 individuals to up to several hundred, but whether the troops are large or small, they all have the same basic social structure. Every troop is organized around several female-headed multigenerational families or matrilineal, plus males who come in from the outside.

And this arrangement derives from the fact that females stay in the group in which they are born for their entire life, whereas males stay only until puberty and then they leave, either voluntarily or they are physically kicked out of their troop usually by other females. And then these males usually join all male gangs, hang out in the gangs for anywhere from several months to over a year, and then work their way into a new troop. So many troops have got plenty of adult males and females in it, but the females have all been there since birth and they’re all related to one another, whereas the males have been there only since puberty and initially they’re related to no one until they start having offspring of their own.

Here is a typical rhesus family: mother, newborn infant, that’s grandmother back there, that’s an older sister, the infant, this is a younger sister of the mother. The other thing you need to know-wrong button-about social organization is that in every troop there is multiple dominance hierarchies. So there’s a hierarchy, for example, between these families, so that every member of the highest ranking family, including infants, outranks everybody in the next highest ranking family. There’s a hierarchy within each family that follows a general rule: Younger daughters outrank older daughters, and this lasts for as long as both daughters are alive, even after the mother’s no longer around to reinforce it.

There’s a hierarchy among the males that come into a troop that’s roughly related to tenure, that is the longer the male is in a troop, the more likely that male is to be high ranking. But of course it’s not tenure or length of time that’s important. What’s really important is how good that male is at making friends and sustaining alliances with not only other males but especially with high ranking females who carry the real clout in the troop. And males who are good at this will move up in the hierarchy and males who are not so good don’t do this very well or last long at all and they usually leave early. And then finally, there’s a hierarchy among the infants that are born into the troop every year, and that’s roughly related to the position in the troop of their mother.

So it’s not that some infants are bigger or stronger or smarter or quicker than other infants. Those individual differences undoubtedly exist. It’s that other monkeys associate a particular infant with a particular mother and afford it equivalent social status. So life in a rhesus monkey troop is a pretty complicated business and for an individual to survive, let alone thrive in this complicated social environment, not only must it acquire a complicated social behavior repertoire but it also has to acquire a great deal of knowledge from the other troop members. Who’s related to whom, where do these individuals fit in the various dominance hierarchies, as well as recent social history. Who’s been getting the fights recently?

Who can you count on to back you up if you get into a fight yourself? And monkeys that are able to acquire this information and follow all the rules and [unint.] in these complicated social dominance hierarchies can handle this complex social living very well. And monkeys who are either unwilling or unable to follow these rules don’t survive very long at all either. So in nature and out at Poolesville, at least for our naturally living groups, rhesus monkey infants, who by the way grow up four times faster than we do, which is an enormous advantage for those of us that are interested in looking at long term effects of early experience or lifespan development or citing the transition of characteristics from one generation to the next, this four to one ratio is an enormous advantage because we can see a generation in four or five years rather than having to wait 15 or 20 years for the human equivalent.

So when I give you an age in rhesus monkey months or years, multiply by four and you get a rough human equivalent. At any rate, rhesus monkey infants spend virtually all of their first month of life in intimate social contact with their mother, usually plastered to their mother’s [unint.] surface or no more than arm’s length away from their mother. And during this time, a strong and enduring attachment bond is established between mother and infant. It’s functionally and biologically the same kind of bond that human infants routinely form with their caregivers. And then during this first month of life as this attachment is formed, by the time these monkeys are two months of age, they’re ready to start using this attachment secured base that they’ve established to start exploring their environment.

So just like Megan has described for human children, rhesus monkey infants use their mother as a secure base. Beginning the second month of life they’ll leave the mother for short periods of time and then come running back to the mother for a period of contact. Then they’ll leave their mother for short periods of time, running back to mother for a period of contact. It’s almost as if they’re attached to their mother by a long rubber band. And as recent months go on, the amount of time and distance spent away from the mother increases, so by six months of age they’re only spending about 20% of their waking hours in actual physical contact with their mother, as opposed to 100% of the time they were spending in that first month.

Yet the mother’s presence in the immediate physical environment is absolutely crucial to sustain this exploration because if you lose your secure base or access to your secure base, then any motivation to explore will disappear and you’ll get pretty upset emotionally. So who are these youngsters spending time with when they are away from their mother? Well, they’re spending it with peers, with age mates just like themselves. And in the wild, this is not by accident, because in the wild, rhesus monkeys are seasonal breeders. Although they copulate year round, the actual conceptions only occur within a two to three month period.

So what this means is 80 to 90% of a troop’s infants are born during the two to three month period. And what this guarantees for any infant born during this time is they will have plenty of other youngsters who are about the same age and have about the same physical, cognitive, and social capabilities as themselves with whom to interact. And so that’s who they play with. For throughout the childhood years, throughout the second year of life and the third year of life, these monkeys will spend several hours a day in active social play with peers. And it’s in the context of this social play that virtually every behavior pattern that’s going to be important for normal adult functioning can be developed, practiced, and perfected long before it actually has to become functional.

And this is especially true for reproductive behavior and for the socialization of aggression that comes into these monkeys’ repertory at between four and six months of age. The work by my mentor Harry Harlow way back in the 1960’s and much more recent work from researchers in the Netherlands have demonstrated that monkeys are otherwise socialized but denied the opportunity to play as they are growing up inevitably grow up to have problems with reproduction and real problems controlling aggression. Now not all of our monkeys out at Poolesville grow up this way. Some of them are cross fostered to unrelated females at birth, so we can start separating out genetic and environmental factors.

And others are reared, as Megan suggested, in the absence of access to their biological mother at all. These are monkeys who are reared in a procedure, again, that Harry Harlow developed many years ago called together together or peer only rearing. Peer monkeys are separated from their biological mother at birth well before there’s any chance of any kind of attachment to be developed. They are hand reared in our neonatal nursery where they get all the nutrition and things. They actually weigh more than their mother reared counterparts under these circumstances. And then at a very early age, usually before the end of the first month of their life, they’re put together with other monkeys like themselves, with peers hence the peer rearing, where they then spend the next six months of life, 24 hours a day, 7 days a week, with one another growing up in this particular environment.

The peer groups are usually four or five individuals. Under these circumstances, peer reared monkeys develop hyper attachments to one another, intense attachments that are actually as strong or stronger than you’ll see in a normal mother-infant bond. But these hyper attachments, while intense, are actually dysfunctional. What happens is the infants spend much of their time clinging to one another much as they would early on clinging to their mother. But instead of going out to explore in the second month of age, they continued this high level of clinging. So their opportunities to expose themselves to the environment are much less as those of their mother reared counterparts.

And when you think about it, it’s not surprising, because a peer is not going to be as good a mother, even a relatively poor mother, at being able to serve as a secure base for exploration or being able to console the infant when it gets upset because usually the peer is as upset as the infant itself. And when these monkeys get to the age where they should be developing patterns of play, they do indeed develop patterns of play, but the play is always primitive, never reaching the level of complexity or intensity that you see in their mother reared counterparts. And when you think about it again, it shouldn’t be a surprise because these peer reared monkeys have to serve as attachment objects and playmates at the same time and they end up not doing such a great job of either.

So these monkeys, although they do play a little bit, it’s much less sophisticated than what you would normally see and it turns out that even though they’re with potential playmates 24 hours a day, these monkeys are actually play deprived as they are growing up. Now when these monkeys reach six or seven months of age, which is a normal age for weaning among mother reared monkeys in the wild, we take the peer reared monkeys and their mother reared counterparts and we put them all into the same large social group where they have to grow up together until puberty and when they go into various other studies. So the differential early experience I’m going to be talking about is only the first six or seven months of life where these monkeys are either mother reared or peer reared.

After that, they’re all living in the same physical and social situations. And when we do that, we find that peer rearing has its consequences. One of the things that happens is that peer reared monkeys grow up to be more fearful than you would expect from their mother had they been reared by their biological mothers. And behaviorally, they look very much like the subgroup of 20% of the natural population that seems to be unusually fearful and anxious under mildly challenging circumstances. And like those fearful monkeys in the wild, these peer reared monkeys also show elevated levels of cortisol under comparable conditions of challenge. Again, whether we’re measuring it in blood or saliva or more recently in hair.

And another thing happens for these peer reared monkeys. As they grow up, not only do they become more fearful than you would expect, they also become more aggressive. And they start developing these patterns of escalating aggression that is so characteristic of that five to 10% of the natural population I mentioned at the outset. So they look like these 5% impulsive naturally occurring impulsive aggressive monkeys. And like those naturally occurring aggressive monkeys, they also show deficits in CSF metabolism. So here are 5HIAA levels in the first years of life and later on. In the first year, these levels are pretty high for everybody. They drop for everyone later on. But if anything, the rearing condition difference has gotten bigger and this is not the result of genetic differences, this is the result of differential early experiences.

One of my colleagues in the NIAAA, the alcohol institute, got access to some of my monkeys when they were adolescents and young adults. They put them through the monkey version of the happy hour where for over a period of weeks and months, for an hour a day every day, these monkeys had unlimited access to 9% NutraSweet flavored alcoholic beverage of nonalcoholic NutraSweet flavored beverage or clean tap water so they are not fluid deprived. And they consume these fluids in familiar situations. And in these circumstances, some monkeys consume more alcohol than do others, and as a group, the peer reared monkeys consistently consume more alcohol than their mother reared counterparts. More recently, we’ve been able to do some neural imaging studies of these monkeys in collaboration with people from NIDA as well as from NIMH.

Here are data from a PET scan study published three or four years ago. This is looking at PET scans of two year old mother or peer reared monkeys, again, only for the first six months of life. For the remaining 18 months, they’re living in the same situation. And what you see on the left, their serotonin transporter binding and cerebral blood flow on the right. Comparisons are between a composite front view of mother reared monkeys versus peer reared monkeys. Top view of mother reared monkeys versus peer reared monkeys, side view of mother reared monkeys versus peer reared monkeys, and similarly for cerebral blood flow: mother versus peer, mother versus peer, and mother versus peer. And simple inspection shows that the peer reared monkeys’ brains are not lighting up as much as their mother reared monkey counterparts.

When you look at the actual data, throughout the entire brain you see these differences as a function of early experience. So whether you’re looking at the thalamus, the striatum, the temporal or the frontal lobes, frontal cortexes of the brain, peer reared monkeys have consistently significantly less serotonin binding potential than their mother reared counterparts in these various brain areas. And I could show you a graph of cerebral blood flow and it would almost be superimposed on these particular findings as well. We just had accepted for publication in collaboration with people from NIDA and NIAAA a paper looking at PET scans of serotonin 1A receptors. We find the same type of individual differences in brain function.

We also recently published a functional structural MRI study looking at mother reared versus peer reared monkeys when they’re three years of age. Again, differential rearing only for the first six months of life, and again there are significant differences in the way their brains are wired as well. So peer rearing has the effect of influencing not only behavioral expression and emotional regulation, more fear, more aggression, more hormonal activity, higher levels of cortisol, neurotransmitter metabolism, lower levels of CSF 5HIAA, and even brain structure and function. And very recently, not yet published currently under review, we’ve been collaborating with colleagues both from McGill University Moshe Szyf and Michael Meaney’s group and also in collaboration with Allison Bennett, our former post-doc now at Wake Forest, looking at genome-wide methylation patterns.

So Moshe, Professor Methylation as we now call him, now has a high put through system where he can look at every gene in the genome in various tissues and see whether its gene is excessively methylated or under methylated. And we have done this looking at tissue samples from eight year old monkeys, prefrontal cortex on the one hand and lymphocytes, white blood cells, on the other. And the results have been astonishing. Here’s chromosome by chromosome. Each one of those lines that you see, bars, is a gene. And if it’s above the line it’s significantly more methylated in mother reared, and if it’s below the line it’s significantly more methylated in peer reared monkeys. The top line is prefrontal cortex, the second line for each chromosome is lymphocytes.

So chromosome one, prefrontal cortex lymphocytes, chromosome two, prefrontal cortex lymphocytes. Here’s another comparison, here’s the X chromosome prefrontal cortex lymphocytes. And the first thing that jumps out is there are a hell of a lot of genes being differentially methylated as a function of what happened in that first six months of life. Fully one fifth of the entire rhesus monkey genome is differentially methylated as a function of what happened in that first six months of life. Over 4,400 genes, both in prefrontal cortex and in lymphocytes. And furthermore, while there is a great deal of tissue specificity, there is also a great deal of overlap. So here for example on chromosome one, our regions in both prefrontal cortex and in lymphocytes, where exactly the same genes are being exactly methylated in exactly the same direction in prefrontal, in brain, and in blood.

So all these genes are being over methylated for mother reared in the brain, in the blood. And similarly over here, they are overly methylated in peer rear both in the brain and the blood. 30% of those 4,400 genes follow that pattern. There’s another 25% in which the pattern is exactly the opposite. So if it’s overly methylated in mother reared in the brain, it will be overly methylated in peer reared in the blood samples. And the reason this is important is while it’s not obviously 100% overlap, but there is at least a sufficient where you can get some information on some of these genes on what’s going on in the brain by looking at what’s going on in the blood. And why this is important is given present technology, you can only get one brain sample in a lifetime, but you can get blood every week. And that’s what we’re now doing.

We’re engaged in a prospective longitudinal study where we’re looking at these differentially reared monkeys starting at birth and watching these patterns emerge as these monkeys are growing up. And finally, these patterns are not random. They follow certain pathways. So here, for example, is the HPA axis. These are all the genes that are differentially methylated as a function of early experience. And this axis, we’ve heard about the HPA axis before, and for some of these genes, like all the vasopressin genes, the oxytocin related genes, you see the effects only in brain, not in blood. Others, you see it largely in blood and some of the [unint.] functions, but in some cases you get a complete overlap. Again, what this means is this is not a random distribution. These things are following pathways.

And given that we now have the rhesus monkey genome to compare with the human genome, we know what these genes are doing right now. These are methylation patterns. We are also now very recently, in a collaboration with Steve Cole at UCLA, looking at actual patterns of gene expression and mRNA levels. These now are data, this is a heat plot from Steve Cole’s analysis of leukocytes, another blood sample from four month old monkeys who are either mother reared or peer reared. So they’re still in their differential early rearing. Consider these are early points, these are microarray data. Consider each of these the equivalent of a gene. If it’s red, it’s over expressed. If it’s green, it’s under expressed. And each of these things across here is an individual subject.

So mother reared, mother reared, mother reared, mother reared, mother reared, peer reared, peer reared, peer reared. And there’s almost no overlap. If it’s red in mother reared it’s green in peer reared; if it’s green in peer reared it’s red in mother reared, with one exception and that’s MR55. MR55 mother reared looks more like a peer reared monkey, except MR55 had an abusive, neglectful mother. And we now know from another set of research from friends and colleagues at [unint.] that offspring of abusive mothers look very much like peer reared monkeys. They’re more fearful and more aggressive; they have higher levels of cortisol and lower 5HIAA’s than monkeys reared by non-abusive mothers. Saw even on a level of an individual subject, you can start to see this. Now we talked about gene by environment interplay.

For the last almost decade now we’ve been taking a candidate gene approach with the help of Peter Lech who used to be at the NIMH, now back in his native [ph.]. The strategy is taking genes that have been identified largely by biological psychiatrists as being important in emotional regulation. Genes for which at the human level there is a polymorphism or polymorphisms, and for which rhesus monkeys have the same gene and the same or equivalent polymorphism. And the gene that we’ve spent the most time with over the years has been the serotonin transporter gene. This is a gene that has acquired a great deal of interest over the last 15 years. It is a gene thought to regulate the reuptake of serotonin presomatically, it’s the same area of mechanism that drugs like Prozac seem to have their effect.

In both humans and rhesus monkeys, there’s a polymorphism due to link variation in the number of base pairs in the promoter region, such as there’s a long version of this gene and a short version of this gene. In both humans and rhesus monkeys these differences are functional; that is the short version of the gene is associated with less efficient transcription of DNA to RNA, the very first step in the protein cascade that leads to gene expression. So the question is does it make a difference whether you have the long or short version? When Peter Lesch first published these studies, he reported that in humans, individuals carrying the short allele were disproportionately likely to be A. incarcerated for aggression, B. hospitalized for severe depression, or C. suicide victims.

But others have not been able to replicate this finding. There has been a lot of controversy, what I consider to be a landmark study by Casby and Moffitt. Brook might have something to say about that later, where they reported that in a sample of individuals of a prospective longitudinal study, individuals in Dunedin, New Zealand, individuals who carry the short allele in their late twenties had higher levels of depression and more depressive symptoms than individuals who either one carried the long allele, but only if they had a high level of concurrent stress, or from my standpoint had a history of childhood maltreatment. Absent either of those environmental characteristics, the individuals carrying the short allele had no more risk of developing depression than individuals carrying the long allele.

And this also has been controversial with meta-analysis flying back and forth. I think the data, when you take a solid look at it, are overwhelmingly in favor of the results published by Casby and Moffitt and their group in Science back in 2003. Well, what about monkeys? Well, with Peter Lesch’s help we were able to characterize all our monkeys out at Poolesville with respect to long and short allele. And some monkeys had the short allele and some had long allele and some were mother reared and some were peer reared, and the question is does it make a difference? And unlike the human literature, which may be considered muddy, the monkey literature is pretty clear.

Let’s look at serotonin metabolism. Here, 5HIAA concentrations, if you have the short allele and you’re peer reared, you show a real deficit in serotonin metabolism. If you show a long allele and your peer reared, if anything slightly higher than normal levels. But if you had a good mother, it doesn’t make a damn bit of difference which allele you had. You had perfectly normal serotonin metabolism. Here is the gene by environment interaction, where I would argue that the motive of the interaction is a buffering effect of mothering. Good mothering is preventing these individuals who carry this gene from developing that kind of deficit. You look at aggression, you get the same picture.

Short allele, peer reared, high levels of aggression. Same short allele, confident mother, low levels, as low or lower than individuals carrying the long version independent of rearing condition. It’s even more impressive when you look at alcohol consumption. Here, if you are peer reared and you carry the short allele, you drink to excess. But if you have that same short allele and you had a good mother, you actually drink less than normal. What might be a genetic risk factor for individuals with adverse early experience may actually be a genetic protective factor for individuals with good early experience. Gene by environment interaction buffering by good mothering. And it starts right at the beginning.

We run all of our monkeys through the monkey version of the Brazelton Neonatal Assessment Scale, a human measurement of neonatal reflexes and temperament developed by T. Berry Brazelton and his students back in the late 70’s. T. Berry just turned 90, by the way, and he’s going strong as ever. The monkey version of the Brazelton translates almost exactly item by item and cluster by cluster from the human version, because in both cases you’re dealing with the same neonatal reflexes, they’re born with the same basic emotions, and at least at that age, there’s no language involved. And in one of the major clusters of the monkey Brazelton, we found over the years visual and audio orienting where you take a toy and move it across the infant’s visual field and see how well the infant follows that toy, monkeys who do very poorly on that during the first month of life almost always turn out to have problems with aggression and low 5HIAA later on.

So when you break our orientation cluster by genes and environment, notice here our peer reared monkeys, there’s a deficit if they have the short allele. But once again, same short allele, perfectly good mother, perfectly normal orienting, perfectly normal development. And it’s not just for this candidate gene. We’ve now looked at about 10 other candidate genes, here’s another familiar one, MAOA, also seen by Casby and Moffitt in their Dunedin study to be associated in this case a gene by environment interaction where individuals carrying the less efficient version of this gene and history of child neglect or abuse showed higher levels of aggression but only with that combination.

We basically replicated that finding in our monkeys. We have a structurally different but functionally equivalent view, and then other candidate genes: dopamine transporters, DRD2, DRD4, neuropeptide Y, neo opioid gene, NOS1 gene, BDNF gene, very recently [unint.] receptor gene. In every case, we see the same pattern. That is, if you take the less efficient version of this gene and you pair it with adverse early experience, you get a deficit. But if you take that same less efficient gene or allele and pair it with good parenting, you actually get an advantage. And here just as an example, alcohol consumption data from Christina Barr’s study of mother reared versus peer reared, the gene in question, this is a snip in the CRH, this is your so-called risk allele, and you can see when it’s paired with poor early experience you have very high levels of alcohol consumption, but with a good mother, it’s low or lower than that.

So almost complete crossover. What does this tell us? It tells us, among other things, that good mothering is pretty damn important. You can go a long way from protecting individuals who carry genes that might otherwise put them at risk to they can have these genes and show perfectly normal development. And why this is even more important is that in rhesus monkeys as in rodent studies and increasingly the body of human data, parenting patterns tend to be passed on to the next generation. Surprising we know our finding by largely non-genetic means of modes of transmission, but nevertheless, daughters tend to rear kids the way they were reared themselves.

So here’s a mechanism by which you can have a so-called risk allele stay in a population generation after generation with a parent null allele effect. Because if you have a mother who has offspring with these risk alleles but she’s a good mother, then not only will her offspring develop normally but that offspring will be a good mother to her own offspring who may continue to carry these risk alleles, et cetera. On the other hand, if something breaks the cycle, something happens to break the compromise the ability of a mother to take care of her kids; maybe there’s been a death in the family, maybe there’s increased competition from other members of the troop, maybe another troop has moved into the area or maybe there’s a food shortage, maybe there’s increased predation, maybe a disaster like Hurricane Katrina.

Any of these things that might affect the mother’s ability to take care of her kids will put those of her kids who carry these risk alleles at risk to develop the kind of patterns that I’m describing. So we’ve gotten very interested now in what is it about good mothering, what is it about these early mother-infant interactions that seem to be so important? I want to show you a couple of videos that I think illustrate this nicely. This first video is of a low ranking mother who is distracting her infant in the face of a fight going on. She doesn’t want the infant to get involved in the fight because she’s low ranking. She doesn’t have the clout to rescue that infant if it goes out and explores and so she waits until the problem is solved.

Let me play that again because we didn’t set it up right. So here’s low ranking mother, infant’s looking away at the fight. Mother grabs the infant, grabs a flower, distracts the infant with a flower-you can see the monkeys running-the fight’s over. Look at that face to face interaction. Rhesus monkeys aren’t supposed to do that. That’s supposed to be uniquely human; only humans do that, not monkeys. But you just saw it. Okay, coast is clear, off goes the infant, on goes the mother. That’s a secure attachment relationship. Here’s a 19 day old infant. Now not only is there face to face interaction but there’s actual exchange of lip smacks.

And again, 50 years of studies of rhesus monkey mother-infant interactions, nobody ever reported this. But when you actually look for it, it’s quite obvious. It’s there a lot, it just disappears early. So whether you’re talking about mutual gaze, lip smacking by the mother, lip smacking by the infant or back and forth, blue is mother-infant interaction, red is interaction of infant with everybody else. It’s only with the mother. It grows and then at three weeks it disappears. What’s happening at three weeks, that’s the age where infants are starting to leave their mother. Here’s a three day old infant. That’s not its mother, that’s Pierre Ferrari from Giacomo Rizzolatti’s lab in Parma, Italy.

Rizzolatti’s the guy who first characterized [unint.] neurons 15 years ago. These are neurons that fire when you perform an action, and they also fire when you see someone else perform an action. Rizzolatti’s group wanted to see how early this might show up, so he sent Pierre to our lab. Pierre, after 30 seconds, is going to start opening his mouth at the infant. And watch the infant. Here’s another three day old infant. Now the lab tech is going to lip smack; he’s going to stick his tongue out. Look at the infant. So this is neonatal imitation, first reported by Metzoff and colleagues in the field 30 years ago. So you’re seeing it in monkeys. So we’ve spent a lot of the last couple of years making faces at our monkeys. We now know quite a bit more about this.

First like in the human case it disappears early. In humans it’s gone after a month; in monkeys it’s gone after a week. Secondly, not everybody does it. About 60% of our monkeys in the nursery show this pattern, about 40% do not. The question is does this make any difference whether you imitate in that first week of life? When we look at the one measure on the Brazelton of sensory motor integration, coordination reach and grasps that involve both sensory and motor integration, that’s the one task on which imitators outperform non-imitators. More interestingly, when these monkeys get to the age when they should play, those who imitated during that first week of life play twice as much as those who did not.

And perhaps most intriguingly, these monkeys, when they get older to two years of age, those who imitated show higher levels of self-directed behavior and those who failed to imitate develop self-directed stereotypic autistic-like stereopathies. None of the monkeys who imitated in that first week of life develop these stereopathies. We think we might have stumbled on a very early screening for autistic behavior. We’ve done work looking now at EEG’s and we’ve got distinctive patterns. This work will continue for some time; these are some of the individuals who contributed to this work. And these are those of the many who contributed to the larger body and I thank for you for your time.

MOD: Thank you Steve, thank you for some fascinating research. We won’t take any questions at this time; we’ll wait and get Steve’s questions as we assemble our panel. Our final speaker is Dr. Jeanne Brooks-Gunn. Dr. Brooks-Gunn has a Masters of Education from Harvard University and a PhD in human learning and development from the University of Pennsylvania. Dr. Brooks-Gunn is the Virginia and Leonard Marx professor of child development educators at the Teacher’s College and the College of Physicians and Surgeons at Columbia University. She co-directs the National Center for Children and Families at Teacher’s College and the Institute for Child and Family Policy at Columbia University. Dr. Brooks-Gunn’s specialty is policy oriented research focusing on family and community influences upon the development of children and youth.

She designs and evaluates interventions aimed at enhancing the wellbeing of children, particularly children living in poverty, and associated conditions. She also conducts research on transitional periods focusing on school and family transitions as well as biological transitions in childhood, adolescence, and adulthood. Her numerous books on these topics include neighborhood poverty, contexts and consequences for children and girls at puberty, biological and psychosocial perspectives. Dr. Brooks-Gunn’s honors are many. They include election into the Institute of Medicine of the National Academies and the Society for Research on Child Development. She has a special award for her distinguished service as it relates to contributions to public policy for children.

She has been recognized by the American Research Association, the American Academy of Political and Social Science, the American Psychological Society, the American Psychological Association, and the Society for Research on Adolescents. It gives me great pleasure to bring to you Dr. Jeanne Brooks-Gunn.

JB: Thanks Yvonne, thanks Roz. We’re going to continue our discussion today looking at relationships at a larger level, and I think this is important because we want to understand parent relationships as regulators in the range of families that exist today in the United States and in other Western countries. So what I’m going to do is talk about one study today. This is called fragile families; it was funded by NICHD. Part of the impetus for this study grew out of a decade long research network that was part of NICHD looking at child and family wellbeing, which I was privileged enough to be part of. So why did we start this study? We started this study because births to unmarried parents in the United States has soared.

Notice this is from 2005. The most recent are 2009 and 49% of all babies born in America are born to unwed mothers. We will probably reach 50% sometime in the next 10 to 15 years. This is true across different racial groups, even though there are some differences as a function of race. For example, for white mothers, we’re up to 30%. Because of this fact, our group of collaborators, and this includes demographers, developmental psychologists-me basically-sociologists and economists, and now geneticists. What we’re interested in doing in this big study is understand what are the capabilities and the circumstances of unmarried parents? What pulls and pushes parents together in the rearing of their children?

How do these relationships change over time? And how do basically stability of families and the structure of families affect what happens to children? I think this is really important in terms of the two talks we’ve already heard because we have talked in a way about family structure in Steve’s work and we’ve talked about stability in terms of the work that Megan does. And we’re interested in how these two factors come together to influence children. The study I’m talking about is the longitudinal birth cohort study. We have 5,000 children and their parents. We took advantage of the magic moment that people have actually talked about, or Megan has, in terms of intervention to talk to mothers right after the baby was born, and you know how fast mothers leave the hospital.

This meant we had to be in the hospitals within 24 hours to get their permission to join our study and then to ask them whether the father would join. Now everybody said, I must say, when we started this study, well you’re going to have almost no fathers. Everybody forgot that whether you’re a married or unmarried father, you tend to go see the baby at the hospital. Consequently by our design, we oversampled unwed couples, but we got 75 to 85% of the fathers to participate. This is basically the only study in America where we have a decent, i.e. representative and large sample of fathers to really understand how fathers and mothers are contributing to child development and what’s happening with their relationship, particularly important because since we know unwed couples are more likely to break up.

Our kids at this very point in time are nine years of age. We’re hoping to see the kids once again at 15 to 16. Just to give you a few examples of this particular sample, if we look at our married and unmarried families, remember we sampled about a quarter were married and three quarters were unmarried; clearly the unmarried are at more risk. They’re more likely to be poor-I’m not showing that-they’re more likely to have low education, they’re more likely to be teenage parents, they’re more likely to not have had two parents growing up. You can see here that I’ve got the married on one side and the unmarried on the other. And in terms of blended families, which is also very interesting in terms of stability of relationships, if you look at the percentage of families that had other children-these would obviously be children of second births or more-60 to 70% of the unmarried moms and dads have other children.

And you’ll see later a lot of those are with other partners. So we’re talking about a situation in America. That’s why I wanted to paint a big picture. 40% of our kids are born to unwed couples. A huge proportion of these have children from other relationships, and the vast majority are poor and have low education. One other point to make is the level of incarceration of our fathers. This is by the time that the children are age three. By the time the children are age three, a third of the unwed couples have a father that was incarcerated. And remember, this is a representative sample. I’m not sampling, you know, some one city. This is a representative sample of fathers. This has a lot to do with stability and it may have a lot to do with regulation in terms of what’s going on in the home.

Of my unwed families, about 50% at the birth of the child were living with the mother and the child. A third were visiting, 8% were friends, they’d already broken up, little or no contact, 9%. These were cases where we think the pregnancy may have been because they were in an abusive relationship. Clearly the fathers we don’t have much information on those 9%, but that is a minority of folks. Am I going the wrong way? Oh, I am going the wrong way. What I wanted to show here is these are unmarried dads at birth, and the one point to take from this is there are very high hopes of these fathers in terms of interacting with their babies and their mothers.

Please note that 85% of the dads have their name on the birth certificate and 88 to 90% visited the mothers in the hospital. So these are fathers who really do care about the babies and want to be involved. When we asked the moms and dads at the birth of the baby what their attitudes were about marriage, please note that for the unmarried mothers and fathers, 60 to 80% said that they would marry the mother or father. Now this high hope we might call an unrealistic hope, but at the time the baby’s born, the majority of these families believe that they would get married. And here’s what really happened. We’re looking now at data through age five. Remember I’ve seen these families repeatedly over time.

We just look at the ending of the relationship with the biological dad by age five. 20% of our marriages had split up. About 40% of our cohabs, our visiting relationships that was 80%, and those who had no relationship, virtually none of them were seeing the father anymore. You can see similar rates of beginning new partnerships. Again, think about the stability for the child and what stability may mean for regulation, and please look at the percentages for child with a new partner also increases. We also are interested in transitions, and I’m just showing you the percentage of transitions unwed families, not the married in this case.

Notice that the fathers, a third had five or more relationship transitions since that baby was born. That’s huge. That’s basically one a year. For the moms, that was 12%, but for the three to four, it was a third. So we’re talking of these unwed couples, a nationally representative sample, basically maybe a quarter don’t have multiple transitions. The rest do. We look at things like maternal stress and harsh parenting and I wish I’d put depression up here but I didn’t. Stress and parenting, you have heard about already from Megan, are things that are related to doing a poor job of helping babies regulate. So in our study we look at maternal stress, we look at harsh parenting, we also look at literacy activities because we’re interested in cognitive behavior, and we also look at maternal depression, which I’m going to come back to in a minute.

But what this shows is that mothers, and this is controlling for a host of factors, the more transitions, the more stressed mothers are, the more harsh they are in their parenting. This is true for dating relationships as well as fathers moving in and out. It’s more strong for recent transitions. That’s not really surprising. We also find that father involvement drops off, let’s say for the 50%, 60%, 70% depending on what category that aren’t living with the children at age five. How often are they actually seeing these children? And what you see here is that if the dad has a new partner or if the mother has a new partner, the interaction with that child drops precipitously. That’s the negative 0.54 and the negative 0.76. Very, very high.

And notice that a lot of people have multiple partners. Also we looked at multiple children, same relationship for the dad. If the dad has other children, he is pulling away from this dyad or triad, depending on what is in the home, relating both to instability but it’s also related to having the presence of a second parent in that child’s life. Now I’m going to show you just three quick slides and then I want to talk about genetics for just a moment to follow up on what Steve was talking about. This is a vocabulary test at age five, and what I’ve done is divided my sample into groups looking at what happened to them from birth to age five.

So I have married, stable, and unstable. The cohabs could get married, they could stay stable, or they could break up. Single mothers could stay stable, meaning no guy moved into the house. Single unstable means guys have moved into the house. The stars really are showing whether this vocabulary test score is lower compared to the married stable for the different groups. And what you see is that married stable and cohab to marry have significantly higher vocabulary test scores at age five than all the other groups. So instability and single parenthood are both relating in this case to lower vocabulary scores. For those of you who are demographers, we controlled for about 25 other control variables here.

Aggression, higher scores mean more aggression. Again, think about some of Steve’s work. Here, what you find is the singles and cohab unstables are the three groups that are showing higher aggression at age five and also at age three. And then finally, I want to just put up obesity since NIH was also interested in obesity. And here what you see is that the cohab and single, much more likely to be obese. Please note these adjusted means, again, adjusting from 20 other variables. Over 22% of the cohabiting unstable kids at age five are obese. We have data that we’ve published on the whole sample at age three, we have a very high rate of obesity at age three in this sample. But this is to show that both family structure and instability are relating to what happens to children. We’re continuing this work and other work like it but what I wanted to do, again, because we want to merge here, we want to talk about animal and human models, is talk about early work we’re doing on genetics. We’re looking at three kinds of environments, we’re looking at the Great Recession, maternal education and family income, and then family instability. And we’re looking at two different systems: the serotonin and dopamine systems at the moment. Steve talked about both of those.

I gave you examples from the first two today. I particularly want to show the Great Recession because we have actually just received a grant from NICHD to look at effects of the great recession on our nine year olds, mainly because we were in the field when the Great Recession. We didn’t plan this, but we were. So clearly as we all know, unfortunately there was a decline of GDP. The unemployment rate has gone up, actually that hasn’t come down even though supposedly we’re out of the Great Recession, and there really was an economic shock which could be seen as an environmental stressor.

Even though those of us in this room are employed, certainly we’re worried about what this Great Recession would mean not only for the country but for our own families. So if you want, there was an economic shock that influenced everybody, and this is how we analyzed our data. It turns out because we’re in 20 cities and it takes six months to see all the kids and the fathers and the mothers and the children in each city, the interviewing gets rolled out. So we’re able to separate out the interview period, because we can do it month by month, into three periods. We call it early, middle, that’s when the Great Recession really hit, and late interview period.

And we can tie to this these periods for each city, for each kid, for each month, the unemployment rate in the city as well as some lagged things. So that’s what we did. We were looking at harsh parenting at age nine; this is spanking, slapping, screaming, yelling, harsh parenting. And we are looking at one of the dopamine receptors and we’re looking at an allele the T or the C. The T allele has been related to impulsivity in other studies and I just want to show you here is the rate of harsh parenting in our study, and the point is-I did it month by month here-to show you that we got a big upswing, if you will, in that middle period, correct?

And then it went down in terms of just the harsh parenting levels in our sample, again controlling for all sorts of things that we control for if we’re social scientists. What’s particularly interesting is look at my middle period. The moms, these are the moms’ genetic profiles, who had the TT or the TC pattern were the ones that showed the huge increase, if you will, at the time when the Great Recession was at its peak, at least in terms of economic uncertainty. You can’t do that for unemployment rates because they haven’t dropped. Whereas the CC allele moms seem to be protected. And just to show you by the three time points, the TT, the TC, and the CC for that middle group show a very nicely ordered pattern.

And what I particularly like about this example is the fact that you can take a shock that’s exogenous, not part of the family, the Great Recession happened to all of us, and show that there is an interaction between that shock and the profile that you have vis-à-vis genetics in terms of how your parenting looks. We’ve gone back to make sure this isn’t something weird about the people who are seen in the middle period. We’ve gone back and looked at the data between age five and age three. There was no Great Recession but people were seen at different points and it did not show this pattern. So we think it really is a neat, neat example that most people when they do this kind of work don’t have that exogenous factor coming into play that they can look at.

Second example, also harsh parenting as I said is one of those parenting behaviors that we think from Megan’s talk is related to how well kids self-regulate. Another one is depression. Mothers who are depressed are very withdrawn and look more like the neglect, not as much like the neglect, but something like the neglectful mothers that Megan talked about. So what we have done in this paper is to review what Colter Mitchell asked at the moment is look at depression levels when the babies were a year of age and we looked at serotonin transporter genes, in this case two genes which I don’t have up here, to see how those might interact with different environmental stressors.

We started with these data. There is, as Steve said, Casby and Moffitt, those guys in 2003 published a very influential article in Science. What they showed in a sample of 20 year olds is that stressful life events at 20 to 21 and serotonin transporter interacted such that the preponderance of folks that showed high depression were those who had a high number of stressful life events at 21. I’m putting this up to show that in this research and almost all of the research done on this with the exception of the animal research and two other child studies, people have only looked at the lower left hand quadrant.

They haven’t really looked to see if certain genes, these reactive genes in Steve’s terms, might actually be protective in good environments, which would be putting up the other. So all they’ve done is looked at half the distribution of what Steve was showing us with the rhesus monkeys. So what we did is we took our moms and we have looked at--we are calling these reactive alleles rather than risky because they could be protective, and we also are looking at more genes than people have looked at. You know when you look at one gene it’s kind of the same problem as we have if we’re development psychologists and we’re looking at mother-child interaction. If I look at two or three minutes of a mother-child interaction, I’m not going to get a good measure of how sensitive that mother is.

If genes do act in concert, looking at multiple genes simultaneously might make a difference. So we looked at two genes, STin 2, and the serotonin transporter that Steve was talking about. As he said, short short is the negative or the we will call it-we should never use negative anymore, we should use reactive-reactive and in our sample that was 10%. About 40% had the long short. For STin 2, having the longer allele, the 12 12, seems to be related possibly to less efficient serotonin processing. So we put those two together and came up with a score from zero to four in terms of the number of susceptible or reactive alleles that we might have.

Then we looked at the association of number of reactive alleles by years of education. We also did family income, but I’m just going to show years of education, on the probability of showing postpartum depression. About 17% of our sample did show postpartum depression using a CIDI measure. What’s interesting is look at the half of the graph in terms of the left, and what you see is what you might expect, which is more of these reactive alleles. When you have low education, the more likely you are to have postpartum depression. However, look at what happens to the lines.

By the time you get up to 14 to 16 years of education, actually having the reactive alleles is related to a lower level of postpartum depression. This is again an example from human behavior of what Steve was talking about with his animals. We have very few examples of this in the human literature so we’re very excited about these findings. Here I’ve just put on zero or one versus three or four with the confidence intervals, for those of you who like to see confidence intervals, to show that the confidence intervals are not overlapping for these two distributions. From my perspective, these data are just the beginning.

What we’re doing here is a little different than what Steve is because we’re looking at maternal genetic profiles to start with and how that’s influencing things like depression, which affect kids negatively, and possibly harsh parenting. Obviously we’re moving now to looking at the children’s genes as well, in concert with the mother’s genetic profiles, but also individually. So that’s one piece of the research that is going forward, particularly looking at instability. The second piece that we are doing is some of the methylation work that Steve talked about. Steve and Moshe Szyf have been helping us with this and we’ve been trying to get methylation patterns out of saliva.

We have 3, 4,000 families and we’re doing saliva not blood mainly because we’re in the field, and the first runs that I actually saw last night suggest that we are going to be able to get some methylation patterns out of our saliva and we’re hoping to collect similar data when the children are age 15. So during that period of puberty that Megan talked about, we can see if we’re having differential methylation patterns across some of these children. And I just wanted to end saying that given our interest in single mothers and low education mothers, I just want to make the point, and this is related to Yvonne’s concern about policy, in these analyses that Sara McClanahan did, she divided-these are national data, not our data-she divided groups into the top quartile of education, the middle half, and the bottom quartile.

And what I want you to look at is how close, in terms of single mothers, the three quartiles were in the 1960’s and the fact that by 2000 we have a fanned pattern. What this is really suggesting is that we’re seeing diverging destinies for our high educated than from our low or middle educated. Here’s another example, there was an article in the New York Times two days ago talking about this that divorce rates have plummeted for people with college educations, but they have not for those without college educations. And you’re getting this fan-like pattern. And then finally, income. You’re seeing exactly the same thing with the top quartile of educated mothers pulling away dramatically by 2000 in income.

The reason I mention this is given my interest in income in single parenthood, I think that the disparities that we see between groups in this country are just continuing to grow, so there is great concern. And I’d like to thank you and thank NICHD.

MOD: We’re a few minutes over the time that we’re allowed, but I still feel that there is a need to have some discussion and perhaps a few burning questions are still there. So Steve, you’re not able to do a lot of walking around, so maybe we can make it up there. Okay, well we’re going to bring the panel or the speakers to the front and give you a chance to maybe ask some questions of them and they may even have questions of each other as we think through the next steps. Given that we are a little pressured for time, I wanted to take one minute to thank Roz King and the members of the OppNet both steering and coordinating committees who have been involved in trying to put together this symposium, but most importantly who have been working on the OppNet initiative since November of 2009.

And thank again our speakers and thank you all for coming. I understand that many of you came from different parts of the NIH campus. It was a cold day to be able to do that, so we thank you all very much for coming. And now I’d like to turn it over to you, if you have any questions of the panelists or the panel has just a few points that it would like to make.

MS: Thank you for your talk, I really enjoyed them. I was wondering if you’ve looked at relationships between early mother and infant attachment and the onset of later disorders like PTSD or more like adolescent adult onset anxiety or depressive disorders and what you found.

JB: Well in our study, we’ll be able to do that if we see the kids at 15, and we will.

MG: The question you’re asking though is a very active long term interest in the field of child development and our colleagues who study attachment as their primary focus and the security of attachment, I don’t know how many attachment aficionados we have out here, but there are several forms of organized patterns of attachment, secure one being labeled B, but there is also a pattern D that is disorganized and disregulated form. That’s the form that is associated with maltreatment, though you can see it in kids where it isn’t really maltreatment that’s happened, and that’s the one that in infancy is the most highly predictive of an increased risk in psychopathology just in and of itself.

The other forms may interact with child temperament to provide greater or lesser support during that period. This is disorganized disorder pattern that seems to be associated with increased risk of real psychopathology.

JB: Just to add to that, in the fragile families we actually did collect some attachment data at age three that somebody in Australia is analyzing for us but being very slow on analyzing it. But we are very interested in the disorganized group from our big sample of 5,000, because we’ll have a bigger sample than anyone else to look at some of those questions. And also I actually think the other kinds of insecure attachment may interact with some of this family instability that we’re seeing in our sample to put kids at particularly high risk for problems in adolescence. I mean time will tell, but that would be my prediction right now.

SS: In the primate work, we spent a lot of time looking at early experiences and their relatively short term effects. The period of adolescence which monkeys go through just like humans is a period of dramatic not only biological and hormonal change but social change as well. The world’s very, very different, for the males in particular, and an exceedingly stressful time. And of course one of the issues is how much of these early insults do they show up later on and as we get genetic tools, are we going to see changes in gene expression differentially laid out as a function of what happened in those initial years.

The other thing we’re particularly interested in is how reversible are these. We have designed some environments for our monkeys so that by the time they are late juveniles or even young adults, many of the behavioral and biological abnormalities or things that were quite different in our peer reared monkeys we can actually reverse. And the question is how deep is that reversal? Does it affect only behavior? Is it also affecting these neural transmitters? Are some of these genes turning back on again? And this is going to be an area of intense study over the next couple of years.

MS: Thank you. FS: Hi, thank you very much for some wonderful presentations. I was wondering, I’d realized the focus was on maternal-infant bonding, but where that bond is disrupted in any way, can you touch on whether there’s a slightly adequate at least substitute such as a grandparent, because we see that so often in some of the fragile families in the country. Can a grandparent substitute at all?

JB: There’s actually not a lot of research on that question, although people draw from work on the kibbutz for example. Some people have done work looking at attachments to fathers, and other people have just looked at levels of sensitivities of mothers and fathers. It’s a little different than attachment but it’s related. And my read of the literature is mixed. I’d say the most recent stuff that’s come out has suggested there is substitutability, but this is in sensitivity of fathers if a mother is not sensitive. So that’s my read of the current literature. The old kibbutz literature might suggest that as well.

Grandmother, I’ve done tons of research on grandmothers in a former life and clearly grandmothers can be very pivotal-this is with teen mothers where the grandmothers are very likely to be involved-and in some cases when they are being very sensitive when the teen mother is being very withdrawn, teen mothers in general tend not to interact a lot with their babies, that seems to have a protective effect. What’s even more interesting is grandmother, you get the same thing that Steve gets in the intergenerational transmission of parenting. So if you have a grandma who is very harsh, she will have a teen mom who is very harsh.

We didn’t get as much transmission in that study of the sensitivity, but we did with the harshness.

MG: I was going to just say that from a basic science standpoint, the issue really is a consistent responsive caregiver. It wouldn’t matter if Charlie the dog was doing it.

JB: Stable, but stable.

MG: Stable response. You like stable. In fact, in the old animal literature there was wonderful work that was done by Bill Mason showing that infant monkeys taken away from their mothers but given to a dog to rear did pretty well.

SS: The dog’s rearing the monkeys, not the other way around.

MG: Yeah, the dog reared the monkey, but the dog was responsive and attentive and the monkeys looked infinitely better than when they were raised on a surrogate hobby horse. So the critical ingredient here is sensitivity, responsiveness, stability of the relationship, continuity of the relationship, and stimulation.

SS: There’s also substantial literature going back to Brown and Rudder, individuals who grow up in the first environments who come out looking really good and the most consistent predictor is they had a mentor or somebody or some people or some individuals who took them under their wing and provided them with that support, stability, and those sorts of things. Again, social situations that are increasingly predictable, you have a lot more room to maneuver and a lot more room to get out of a hole.

JB: Which is why I’m so worried now about these transitions that I’m seeing in my families because they’re so much higher than even those of us who started the study thought. So what happens to the kids long term, even though they’re always with the mother. You’ve got all this chaos in and out going on and you know, we just don’t on that particular fact, we just don’t have the data before the study to really address that issue, but we are now.

BE: Thanks Yvonne. I’m Bill Elwood, the OppNet facilitator, so on behalf of all of us in NIH and OppNet and all the people in the public who will see your presentations on the internet because we will post this on our public website, thank you very much for setting a high standard for any of the symposia that follow you. My question is you folks have done so much to advance the field in your discrete areas as well as building links among the basic science, applied behavioral and social sciences, as well as the basic BSSR. With the three of you together and the experience you have on your own with other colleagues and you as colleagues who have worked together these few years out of your post-docs, where do you see research going from here? Where would you like to see it go?

SS: I’d like to see more interdisciplinary collaboration. I’d like to see the presumed bridges and barriers between so-called behavioral and so-called biological and so-called genetic research start to vanish. I think increasingly, we now know that genes are not as deterministic as they were once thought to be. Their characteristics, at least in their expression, I think, is going to turn out to be just as plastic as just about anything else. And so to understand genetic actions, you’re going to have to know in what environment they are occurring. To understand why certain social things are you have to appreciate the differential risk that the different alleles Brooke was talking about are as well, and what this means is that you have to have more communication among individuals within that.

I think the three of us are all beneficiaries of in some cases the same and in other cases quite different integrative research networks that have really been worth every penny that was put into it. Interdisciplinary work is not easy to do. You have to have the right set of players. As Megan said, you have to find people who are willing to play in the same sandbox, but when it happens it’s just a wonderful, exciting thing to see.

MOD: I took copious notes as you all were giving your presentations because we get back as a steering committee, and again talking about next steps and initiatives and activities surrounding many of the areas you’ve addressed today, we’ll be able to use some of these comments that I’ve created. But I think one of the things that was very startling to me and of great interest, as I listened to each of you and your presentations, the title of this series is you know, to put together a group of talks that involve human-animal models. And then this first seminar was on maternal-infant bonding.

Well actually, when I think about it, it’s not just maternal-infant bonding that we’re talking about here, and I think the question that we just got from Ms. Caser in the audience addresses that. It’s really who provides that initial critical care to the infant during that first stage of life, and one of the questions I had was related to something that I gleaned from each of your talks, and it really is related to the timing in which the care begins, but most importantly the timing at which the care stops or is interrupted or becomes harsh, in the term that you’ve used.

And as you think about your data, looking at adoption and kids who are adopted, I’m really interested in the age of adoption and particularly as it relates to the measurement of the stress hormones, particularly here cortisol. We showed and we saw the data as it relates to individuals, children who are five, six, seven years old. How would that profile look if we were actually to measure the cortisol levels in the child much younger and perhaps at a time in which they may have been adopted at one year of age or three years of age? Do we still see that elevation?

MG: The children had been adopted at varying points between about seven months of age and about two to three years of age. We were seeing them multiple years afterwards. But the whole question of the ration of early deprivation, how plastic or what systems at which time, if you move them out of conditions, is a huge one. And unfortunately, we cannot randomly assign from a scientific standpoint, unfortunately, to when kids are going to be adopted. However, I believe NICHD is supporting one of the only studies that has ever been able to do that, and that’s the Bucharest Early Intervention Project. Aren’t you supporting…? Is NIMH doing that? Dang, I was trying to help NICHD.

But this was an amazing study where they actually worked in Bucharest foster care, which didn’t exist there. They had baseline measures and then they randomly assigned kids to care as usual, which at that point was to stay in the institution, though that changed, or to go into foster care. And because they were the ones doing the random assignment, the kids were between seven months and 32 months at the time that they went into foster care. So they’ve been working very hard to see if they have some sensitive periods in that data for different systems on when you get into a more improved care setup. And the kids are now, I think they’ve finished their eight year data collection and I think they’re actually supporting the 12 or 11 year collection.

So that’s really the only study we have to look at, and unfortunately they weren’t able to randomly assign from one month on. Because we think from the nonrandom assignment studies that there is a remarkable window in that first six months of life where if you can shift the care, you really don’t see in almost any system you look at the effects of early adversity. The system is so terribly plastic. And then it becomes different systems becoming increasingly less plastic over time. Though, you know, where there is life, there is hope, and one of the things humans are immensely good at is figuring out workaround systems to be able to achieve the same goals but maybe with different neural systems.

MOD: And this is a more practical question. How do you measure shyness in children as well as in animals? What’s the criteria for measuring shyness? Let’s do animals first.

SS: You put four of them in a playroom and see what they do. And some of them go into the corner and watch, and some of them interact actively and go into the corner and try to drag the one in the corner out. The one doing the dragging is certainly not shy. And the other thing as you look for adaptation, because almost all of the shy individuals, once they’re used to a situation or with familiar individuals, they come out of their shell, which is something nice to see. But you can track these changes over time and they’re really quite obvious. MOD: And will a mother or a parent early in life be able to identify when she or they think that their child is shy?

SS: For the monkeys, the good mothers, that’s what they’re really good at doing.

MG: And of course, human parents are extremely sensitive to the issue of how their child reacts when they take them to see Santa Claus or you know, they have to enter into a new childcare setting or there’s a new thing that happens. In the United States, we highly value children who are exuberant and outgoing and not shy, and there’s a varying range of strategies and sensitivities for what you do to the child that is cautious and inhibited, and some of those children have parents that of course think it’s inappropriate behavior and will stand on their hand to try to get them to approach strange and novel things. Early on, it shows that that behavior drives the HPA system like crazy.

SS: It’s stressful.

MG: It’s stressful; your parent would like you to go and approach the thing that is scaring you, and that the parents who are sensitive and responsive and are there for the child, and so pretty quickly the child is then the one-because they would like to explore the new thing-that figures out how to use their parent to get them over there and do what they need to do. So a lot of variation in that is associated with attachment security. Insecure parents are more intrusive and overwhelming, and so it’s precisely in those contexts where you’ve got the child being frightened of what’s out there and the parent evaluates that it’s not really frightening.

MOD: Let’s give our panel a round of applause. (END OF TRANSCRIPTION)

Back to Top

This page last reviewed: January 28, 2011