Translational developmental studies of stress on brain and behavior: Implications for adolescent mental health and illness?

Abstract

Adolescence is the transition from childhood to adulthood, with onset marked by puberty and the offset by relative independence from parents. Across species, it is a time of incredible change that carries increased risks and rewards. The ability of the individual to respond adequately to the mental, physical and emotional stresses of life during this time is a function of both their early environment and their present state. In this article, we focus on the effects that acute threat and chronic stress have on the brain and behavior in humans and rodents. First, we highlight developmental changes in frontolimbic function as healthy individuals transition into and out of adolescence. Second, we examine genetic factors that may enhance susceptibility to stress in one individual over another using translation from genetic mouse models to human neuroimaging. Third, we examine how the timing and nature of stress varies in its impact on brain and behavior. These findings are discussed in the context of implications for adolescent mental health and illness.

Introduction

Recent reports in the US suggest that stress is taking a toll on the physical and mental health of Americans and their families (APA, 2010), with reports of stress levels exceeding what families consider to be healthy in the majority of homes. It should come as no surprise then that stress related disorders including anxiety and depression affect as many as 10% of our youth, making them the most prevalent of the developmental psychiatric disorders (Newman et al., 1996, Pollack et al., 1996, Kim-Cohen et al., 2003, Kessler et al., 2005, Merikangas et al., 2010) This article highlights recent studies on the impact of stress on the brain and behavior across development and across species that may help to explain the high incidence of anxiety and stress related disorders during adolescence.

Overview

Stress occurs when mental, emotional and or physical demands exceed the regulatory capacity of the organism. Situations that are highly unpredictable or uncontrollable are examples of highly stressful environments (e.g., institutionalization, warfare) (Koolhaas, 2011)). Such environments threaten the survival or coping ability of the organism. When such threats occur, the nervous system responds by releasing stress hormones that help put the organism in an alert state and ready for action. While moderate stress can be adaptive in helping the animal respond to the situational demands, chronic stress can have lasting negative consequences (Shonkoff et al., 2009). Thus, stress can vary along dimensions of frequency, duration (e.g. acute or chronic), and magnitude, each of which has different implications for the stability of the animal. We focus on the adaptive and maladaptive effects of psychological stress across different points in development in this paper.

We highlight three approaches that examine the impact of emotionally charged or environmentally demanding events that can lead to stress. In the context of this review, a potential threat can be a stressor depending on how one perceives that threat. Since threat is invariably associated with negative emotions, how well we can regulate those emotions can influence whether we perceive it as a psychologically stressful event. We begin with a brief review of threat related brain circuitry. We then present findings from recent human imaging and mouse studies that illustrate developmental differences in response to potential threat highlighting changes during the period of adolescence. Second, we describe mouse and human genetic studies that illustrate individual variability in response to threat. We end by providing an example of prolonged early stress in humans; specifically those reared in institutions abroad and illustrate how such challenging environments impact later behavioral and neural responses to potential threat.

Stress effects on the brain

Major circuits involving the amygdala/hippocampal complex together with the prefrontal cortex support behaviors related to threat processing and vigilance (Lupien et al., 2009). Threat results in the release of stress hormones that target regions of the brain and major muscles key for flight or flight. Under non-stressful conditions, these hormones help to support growth and development (De Kloet et al., 1998). However, under conditions of challenge the release of hormones suppress growth and repair in order to support functions necessary for survival. Especially key to the stress response is the release of glucocorticoids that redistribute glucose to the body to help the individual overcome the threat or challenge. Failure to activate the stress response places the organism in a fragile state. Yet, failure to inhibit the stress response can result in disease and lasting adverse effects on growth and development.

This article focuses on the effects of psychological stressors. Psychological stressors, in contrast to physiological ones (e.g., hypoxia), require higher order processing and interpretation of sensory information, thus making connections with limbic and cognitive circuitry crucial for reacting to this type of stressor. Specifically the amygdala (Davis, 1992) appears to be critical in activating the stress response to cognitive-emotional challenge and threat while hippocampal and prefrontal regions appear critical in the regulation of the stress response. For the purposes of this review we will focus predominantly on frontoamygdala circuitry implicated in threat processing (see Figure 1).

Fig. 1. Frontoamygdala circuit of Fear.

A simplified cartoon of the brain circuitry involved in emotion reactivity and regulation related to threat processing. The amygdala receives multimodal sensory signals that may initiate a fear response, while top-down input from the infralimbic prefrontal cortex to the amygdala can dampen or extinguish fear responses generated there. Abbreviations: BA, basal amygdala; CE, central amygdala; IL, infralimbic prefrontal cortex; ITC, intercalated; LA, lateral amygdala; PL, prefrontal cortex; vmPFC, ventromedial prefrontal cortex. Adapted from Casey et al 2013.

The majority of human and animal stress studies have focused on the effects of stress on the hippocampus, but more attention has been given to the amygdala and prefrontal cortex in recent years (Liston et al., 2009). A large body of research, beyond the scope of this paper, has documented detrimental effects of stress on the hippocampus. In brief, these studies show that repeated threat or chronic stress leads to decreased hippocampal volume, dendritic spine density, and a remodeling of synaptic terminals of this region (Magarinos et al., 1997). If the stressor is short-lived, these effects are reversible (McEwen, 1998). However, if stress occurs over a period of months or years it can result in irreversible apical dendritic atrophy and even cell death (Uno et al., 1989). The amygdala, in contrast to the hippocampus, shows proliferative effects with stress. Specifically, stress and/or the administration of stress hormones leads to enhanced dendritic arborization and increased spine density (Vyas et al., 2002, Vyas et al., 2003, Mitra et al., 2005) that may be less reversible than the hippocampus (Vyas et al., 2004). Human imaging studies show parallel results of stress on these regions, with individuals who have experienced high levels of stress showing smaller hippocampal volume (Bremner et al., 1995, Gurvits et al., 1996, Bremner et al., 1997), larger amygdala volume (Tottenham et al., 2010), and elevated amygdala activity to cues of threat relative to non-stressed individuals (Liberzon et al., 1999, Rauch et al., 2000, Shin et al., 2005).

Repeated stress has been shown to have profound effects on prefrontal functions too (e.g., Arnsten, 1999, Mizoguchi et al., 2000, Bland et al., 2004, Cook and Wellman, 2004, Moghaddam and Jackson, 2004, Maroun, 2006, Radley et al., 2006, Del Arco et al., 2007), by diminishing the ability to flexibly regulate attention, actions and affect (Phelps et al., 2004, Liston et al., 2006b). Whereas the amygdala and hippocampus are involved in learning about cues and contexts that signal threat, the prefrontal cortex has been suggested to be involved in “un-learning” these associations (Morgan and LeDoux, 1999, Nair et al., 2001, Herry and Garcia, 2002, Gottfried and Dolan, 2004, Phelps et al., 2004, Santini et al., 2004, Mickley et al., 2005, Akirav and Maroun, 2006, Kalisch et al., 2006, Corcoran and Quirk, 2007, Milad et al., 2007). Failure to recognize when environmental cues and contexts are no longer threatening (Quirk and Gehlert, 2003) has been suggested to be at the very core of anxiety and stress related disorders that peak in diagnosis around adolescence.

Adolescence: A time of stress

By definition, adolescence poses new environmental demands on the organism, as the individual moves from dependence on parents to relative independence. As such, the adolescent must rapidly adapt to new social, sexual, and intellectual challenges (Romeo, 2010, Spear, 2010). In a series of recent experiments we examined changes in the brain and behavior to threat during adolescence. Our work uses two distinct approaches. The first approach involves the use of naturalistic cues (e.g., a frightened face) that over a lifetime become associated with potential threat in the environment. The second behavioral paradigm involves experimentally manipulating a neutral stimulus to take on aversive associations using Pavlovian fear conditioning. This conditioning involves pairing a neutral cue (e.g., tone) repeatedly with an aversive stimulus (e.g., shock), until the neutral cue takes on noxious properties that mimic the aversive stimulus through associative learning. Both approaches have been used to determine how well an individual can suppress a fear response when danger cues and contexts are no longer a source of threat, but differ in behavioral validity when considering developmental and species differences that we will discuss. We present converging evidence for developmental variation in fear regulation using both of these approaches below.

Transitional Studies of Threat

In a series of neuroimaging studies of adolescents, we have examined inflections in behavior as the individual transitions into and out of adolescence (Galvan et al., 2006, Hare et al., 2008, Somerville and Casey, 2010). In the most relevant of these studies, we examined responses to threat cues (fearful faces) in 60 children, adolescents, and adults with functional magnetic resonance imaging (fMRI). This study went beyond examining the magnitude of brain activity that has been shown by several groups to be higher in adolescents than in adults to such cues (Monk et al., 2003b, Ernst et al., 2005, Rich et al., 2006, Williams et al., 2006, Guyer et al., 2008a, Guyer et al., 2008b, Guyer et al., 2009) to show specific changes in adolescents to these threat cues relative to both adults and children. We examined not only transient patterns of frontolimbic activity, but changes in activity as a function of repeated exposure (Hare et al., 2008).

Our adult human results showed that reaction times to threat cues were longer than to neutral ones (Hare et al 2005). Reaction times were positively associated with amygdala activity and negatively associated with the ventromedial prefrontal activity. Adolescents showed an initial exaggerated amygdala response to cues that signal threat (fearful faces) relative to children and adults (see Figure 2, Hare et al., 2008). This initial heightened response in amygdala activity was age-dependent and did not correlate with symptoms of anxiety. Although several groups have shown similar elevated amygdala activity to emotional pictures in adolescents relative to adults (Monk et al., 2003a, Guyer et al., 2008b) few studies have shown a distinct pattern in adolescents from both children and adults. This response was attenuated with repeated presentation of the fearful face (i.e., exposure to empty threat) across experimental trials. The extent to which activation of the amygdala diminished with time was correlated with ratings of everyday anxiety as measured by the Spielberger trait anxiety rating scale (Spielberger et al., 1988). These findings suggest that initial emotional reactivity as indexed by elevated amygdala activity may be typical of or normal for adolescents, but that failure of this response to subside over time with no impending threat is atypical and may be indicative of heightened anxiety during this period. During adolescence, when the amygdala response is heightened relative to that observed in children and adults, more top down prefrontal control may be needed to effectively attenuate the fear response. Failure to dampen this response may lead to symptoms and ultimately diagnosis of anxiety and stress related disorders.

Fig. 2. Amygdala response to empty threat as a function of age and symptoms of anxiety.

(A) Depiction of threat stimulus and location of activation in the amygdala. Middle: Amygdala activity to empty threat (fearful faces) plotted as a function of age. (B) Scatter plot of the correlation between Spielberger trait anxiety scores and habituation (decrease from early to late trials) of amygdala activity for teens and adults (note: anxiety scale was not appropriate for under 13 years) r = −.447, p < 0.001. Adapted from Hare et al., 2008.

Translational Studies of Threat

In a recent parallel study of adolescent humans and mice, we examined sensitivity to threat cues. In contrast to the previous studies, instead of using naturalistic cues (fearful faces) we used neutral cues to control for the amount of history with the cue of threat. Naturalistic threat cues come to be associated with danger over a lifetime. However, our experiences over a lifetime are not equivocal and are limited by our age and opportunity for such experiences. For example, a child may have fewer experiences of dangerous situations or threats than an adult, and an anxious child may have many more experiences of threat than a non-anxious child. These experiences will differentially impact threat-related circuitry. Fear learning paradigms are thus advantageous in that they can assess fear learning equivalently in typically and atypically developing humans. Second, because there is a high degree of neural and behavioral conservation across species, fear learning can be assessed equivalently in humans and mice. The translation of findings from human to mouse provides the added opportunity of delineating mechanisms of change in mice that would be more difficult in developing humans.

In our experiments we used a Pavlovian fear conditioning paradigm to more directly examine how responses to threat change during the period of adolescence (Lau et al., 2008, Lissek et al., 2009, Pine, 2009, Waters et al., 2009, Britton et al., 2011, Lau et al., 2011, Pattwell et al., 2012). Specifically we wanted to examine the ability of the adolescent to regulate fear once the threat of fear was removed (i.e., extinction learning). We tested over 80 individuals between the ages of 5 and 28 using skin conductance response (SCR) to measure physiological responses of arousal during both fear conditioning and extinction (Fere, 1888, Cacioppo et al., 2007). Because of the developmental nature of this study we used an aversive sound for the human subjects rather than shock as our aversive stimulus and paired it repeatedly with a neutral stimulus (yellow or blue square). Our results indicated no effect of age on fear acquisition, but a significant effect on fear extinction (Pattwell et al., 2012). Adolescents showed attenuated fear extinction relative to both children and adults (Figure 3). This effect remained when co-varying for both gender and trait anxiety in the humans.

Fig. 3. Developmental variation in fear extinction learning.

(B) Extinction learning is attenuated during adolescence in the human as measured by less change in galvanic skin response with repeated presentation of the conditioned stimulus alone during extinction trials. (C) This finding is paralleled in the mouse as measured by less change in freezing behavior. Reproduced with permission from (Pattwell et al 2012).

In a parallel study with mice postnatal days (P)23, 29 and 70, we used freezing behavior to measure the fear response, electric shock as the unconditioned aversive stimulus, and a tone as the conditioned stimulus. We observed a similar developmental pattern. The adolescent (P29) mice, like human subjects, showed diminished fear extinction learning compared to the preadolescent and adult mice. These findings are consistent with rodent studies that show adolescent rats require twice as many extinction trials as adults, or prolonged duration of the conditioned stimulus to achieve reductions in conditioned fear behavior comparable to those seen in adult rats (McCallum et al., 2010, Lai et al., 2012). Thus, adolescence is a time when threat cues appear to be highly salient and more resistent to exinction than at any other time in development.

The mouse model provides the opportunity to assess the mechanism underlying the conserved age differences in fear extinction learning across species. As such, we used immunohistochemical and electrophysiological methods to assess neurobiological changes in frontolimbic circuitry in the mice across developmental stages. We focused on the infralimbic cortex because of its role in extinction learning (Santini et al., 2004) and because of the behavioral findings indicating diminished extinction learning in adolescents. We measured activity-induced expression of the immediate early gene c-Fos in the infralimbic cortex. Consistent with previous studies, the density of c-Fos-labeled cells in the infralimbic cortex of adult mice was significantly higher than non-extinguished, fear-conditioned controls. In contrast, there was no change in density of c-Fos labeling in the adolescent (29-day-old) mice. These data suggest that the neural circuit engaged by fear extinction learning in adults is not active during adolescence, providing a likely neural substrate for the inefficiency of cortical control of fear responses during adolescence.

To further delineate changes in frontoamygdala circuitry with age, we performed electrophysiological recordings in ventromedial prefrontal cortex (vmPFC) brain slices of mice after both fear acquisition and fear extinction. Previously it has been shown that fear conditioning involves a decrease in intrinsic excitability of infralimbic cortex whereas fear extinction reversed this decrease in excitability (Santini et al., 2008). Electrophysiological recordings at infralimbic and prelimbic cortex synapses across age showed a fear-conditioning-induced potentiation of prelimbic synapses present in adult mice that was absent in adolescent mice. Extinction-induced enhancement of infralimbic cortex synaptic plasticity in adult mice was lacking in adolescent mice (Pattwell et al., 2012)

Together, these studies suggest blunted regulation of amygdaladependent fear responses during fear extinction in adolescents. These findings may help provide novel insights into the heightened prevalence and treatment of anxiety disorders during adolescence, as the main form of cognitive behavioral therapy relies on principles of extinction learning.

Genetic Factors

The previous work is consistent with different developmental trajectories of distinct limbic brain regions being involved in adaptive fear responses. However, within any developmental stage there is marked individual variability. An important source of variability is that of genetic variation. The main avenues for understanding gene function in anxiety and stress related disorders have been human genetic association studies on one end and genetically engineered mouse models on the other. We have used brain imaging to link structural and functional abnormalities seen in knockout/transgenic mouse models to abnormal patterns of brain activity seen in humans to bridge these (Casey et al., 2010). In an effort to implement a translational approach to human genetic variability, we focused on a common polymorphism in the human gene for brain-derived neurotrophic factor (BDNF). BDNF is a growth factor that plays a central role in neuronal survival, growth, and synaptic plasticity---all core aspects of associative learning in the central nervous system and adaptive fear learning in particular. Human populations contain a common single nucleotide polymorphism (SNP) that causes a valine-to-methionine substitution at codon 66 (Val66Met). This polymorphism leads to decreased trafficking of BDNF into the regulated secretory pathway, which in turn leads to impaired activity-dependent release of BDNF. The BDNF gene is highly conserved from mouse to human, and wild-type mice naturally express the ancestral valine form of the BDNF peptide. To study the effects of the human Val66Met polymorphism in mice, we created a knock-in mouse with a BDNF protein identical to the wild type except it contains a methionine in codon 66 (BDNFmet). Hippocampal neurons obtained from these BDNFmet mice have impaired activity-dependent BDNF secretion and show reductions in dendritic arborization. These mice also exhibit hippocampaldependent learning deficits similar to the findings in humans with the variant human BDNF, validating this mouse as a model of the human Val66Met polymorphism.

We examined the impact of the variant BDNF on fear regulation using similar fear conditioning and extinction paradigms in mice and humans as those described above. In adult humans and mice, we observed less extinction in Met allele carriers than in Val allele carriers, as shown in Figure 4 (Soliman et al., 2010). Moreover, human functional neuroimaging data provided neurobiological validation of the cross-species translation. Specifically, we showed alterations in frontoamygdala circuitry, as a function of BDNF genotype. During extinction, Met allele carriers showed less vmPFC activity (Figure 4c) but greater amygdala activity (Figure 4d) than non-carriers. These findings suggest that cortical regions essential for extinction in animals and humans are less responsive in Met allele carriers. Moreover, amygdala recruitment which should show diminished activity during extinction was elevated in Met allele carriers, suggesting less dampening by vmPFC and more fear response as generated by amygdala output to neuromodulatory systems, the hypothalamus, periaqueductal gray, and vagus (LeDoux, 2000, Phelps et al., 2004).

Fig. 4. Genetic variation in fear extinction learning and limbic activity.

(A) Extinction learning is attenuated in mice with the BNDF Met (M) allele relative to non-Met allele (V) carriers as measured by less change in freezing behavior with repeated presentation of the conditioned stimulus alone during extinction trials. (B) This finding is paralleled in the human as measured by less change in galvanic skin response. (C) Brain activity as indexed by percent change in magnetic resonance (MR) signal during extinction in the ventromedial prefrontal cortex (vmPFC) by genotype (x, y, z = −4, 24, 3), with Met allele carriers having significantly less activity than Val/Val homozygotes (VM < VV is blue), image threshold P < 0.05, corrected. (D) Genotypic differences in left amygdala activity during extinction (x, y, z = −25, 2, −20) in 70 humans, with Met allele carriers having significantly greater activity than Val/Val homozygotes (VM > VV is orange), image threshold P < 0.05, corrected. *P < 0.05. **MM were included in the analysis with VM, but plotted separately to see the dose response. All results are presented as mean +/− SEM. VV, Val/Val; VM, Val/Met; MM, Met/Met. Adapted from Soliman et al 2010.

These genetic findings provide an example of bridging human behavioral and imaging genetics with a molecular mouse model to suggest a role for BDNF in anxiety and stress. Individuals with the BDNF Met allele may be more vulnerable to developing symptoms of anxiety as teens, in that they show higher and prolonged patterns of amygdala activity and less vmPFC activity in response to threat. During a period when evaluating social cues from peers is essential in forming and maintaining healthy peer relationships, the failure to suppress heightened emotional responses to empty threat (e.g., failure of a peer to notice or smile at a teenager, without any negative intent) could lead to over interpretation and ruminations of self-doubt (Guyer et al., 2008a, Monk et al., 2008). The genetic data provide an example of how an imbalance in amygdalavmPFC coupling during typical development could predispose the adolescent to anxiety and, when exacerbated by an individual factor such as the BDNF Met66 allele, lead to clinical levels of anxiety.

Effects of Early Adversity

The variability observed in both our developmental studies of fear regulation may in part be due to genetic variation, but clearly individual experiences impact behavior. A number of studies have shown the significance of environmental factors such as early adversity and stress on the brain and on behavior (Liston et al., 2006a, Liston et al., 2009, Tottenham et al., 2009, 2010, Tottenham et al., in press). Individuals who experience adversity or multiple traumas during development, may be especially vulnerable for developing symptoms of anxiety or depression as teens or adults. Non-human animal studies have shown that early rearing conditions can have long-term consequences on emotional behavior and the effects of early experience can be more significant than later experiences (Sabatini et al., 2007). Many of these behavioral outcomes are associated with changes in limbic circuitry. Within this circuitry, the region of the amygdala is particularly sensitive to early life rearing conditions (Plotsky et al., 2005, Sabatini et al., 2007, Kikusui and Mori, 2009) and its growth and hyperactivity under such stress, have been shown to mediate the expression of hyperemotionality as measured by increased anxiety-like behaviors in animals (Vyas and Chattarji, 2004).

We recently examined the effects of suboptimal early rearing conditions on human development by examining threat related behavior in children and adolescents reared in an orphanage before being adopted to the U.S. These children exhibit elevated emotional reactivity (Colvert et al., 2008) more anxiety (Casey et al., 2009, Zeanah et al., 2009), internalizing problems (Juffer and van Ijzendoorn, 2005) and difficulty regulating behavior in emotionally arousing contexts (Tottenham et al., 2009) - a profile that can persist for many years. Our question was to what extent these children could suppress threat responses with repeated presentation of an empty threat such as a fearful face. We used the same paradigm as that described earlier by Hare et al. (2008) and collected both structural and functional MRI data. Behavioral and imaging data were collected from nearly 60 children (28 adopted and 27 non-adopted) with an average age of 10 years (Tottenham et al., 2010).

Our structural MRI results revealed larger amygdala volumes for children adopted from institutions abroad, specifically for those adopted at ages older than 15 months (see Figure #5). No differences were observed in the hippocampus, however. Given that several years have passed between the offset of early institutionalization and when children visit the lab for an MRI, it is possible that hippocampal volumes have recovered from any stress effects, while their amygdalae have not. These findings would parallel animal studies showing that changes in the hippocampus and amygdala following termination of stress often result in recovery of the hippocampus but not the amygdala (Vyas et al., 2004).

Fig. 5. Structural changes in limbic structures with early adversity.

(A) Anatomical segmentation of the amygdala. (B) Children institutionalized for more than 15 months had larger amygdala volumes than those institutionalized for less than 15 months, or control children. No differences in hippocampal volume were observed between groups. Adapted from Tottenham et al., 2010.

Most striking of the imaging findings were those from the functional imaging experiment. Specifically, children reared in orphanges showed elevated amygdala activity to emotional distractors, relative to children reared with their biological families (see Figure 6). The enhanced amygdala activity in the adopted children may suggest that they were less able to suppress irrelevant emotional information relative to the comparison group when performing the task. Examination of prefrontal regions involved in modulating the amygdala (Phelps et al., 2004, Quirk and Beer, 2006) showed atypical activity. Unlike the comparison children, the adopted children showed little to no change in prefrontal regions. In healthy populations, the amygdala and ventromedial prefrontal cortex showed inverse patterns of activity during performance of such tasks (Phelps et al., 2004, Hare et al., 2008), which might be mediated by the integrity of the white matter tracts between them (Kim and Whalen, 2009). Populations with anxiety and stress related disorders show less inverse coupling between theses two regions (Shin et al., 2006, Marsh et al., 2008). Therefore the findings are consistent with less top down prefrontal control of amygdala related fear responses in the children raised in the orphanage, which is supported by reports of reduced white matter between these regions in an independent sample of previously institutionalized children (Govindan et al., 2009).

Fig. 6. Frontoamygdala activity to distracting emotional stimuli with early adversity.

(A) Previously institutionalized (PI) children exhibited greater amygdala activity in response to emotional distractors than their typically reared counterparts, suggesting an inability to suppress emotionally laden irrelevant information. (B) Post-hoc t-tests of activity vs. baseline. Adapted from Tottenham et al., 2011.

Conclusion

In this review, we described three sets of experiments showing differential responses to acute threat and chronic stress across development. In the first, set of studies we illustrated developmental differences in the regulation of the fear response to potential threat in both mice and humans. Specifically, we showed that adolescents have diminished ability to suppress fear responses when the threat is no longer present. Second, we described parallel human and mouse genetic experiments that showed striking individual variability in response to threat and the underlying neural circuitry as a function of genetic factors. Finally, we provided evidence for environmental factors such as the early life stress of institutional rearing on the fear response and underlying neural circuitry. Each of these approaches alone provides limited information on developmental, genetic and environmental factors that influence the impact of stress on behavior and later outcomes. Taken together, the findings indicate that increased risk for anxiety and stress related disorders in adolescence may be associated with different developmental trajectories of subcortical emotional systems relative to cortical control regions involved in suppressing emotional responses. This differential development can lead to an imbalance in control by subcortical regions over prefrontal ones leading to heightened emotional reactivity. Although elevated emotional reactivity appears to be a typical part of development during the period of adolescence, failure to suppress that emotional reactivity over time seems to be associated more with individual differences in, or symptoms of, anxiety. Both environmental (e.g., early institutional experience) and genetic (BDNF Val66Met polymorphism) factors can exacerbate the imbalance between limbic and cortical regions resulting is dysregulation of limbic circuitry and sustained rather than transient emotional responses to cues of threat.

Our findings suggest that it is sustainment of the emotional response that leads to anxiogenic feelings and possible risk for anxiety disorders. Important future directions will be to consider how genetic, environmental and developmental factors inter-relate in sufficiently large human samples or in mouse models to directly test these effects from a developmental perspective. Such genetic studies will need to entertain dynamic models that capture the effects of changing environmental and developmental demands on the organism.

Highlights.

• The effects that stress has on brain and behavior in humans and rodents

• Developmental changes in frontolimbic function during the transition into and out of adolescence

• Genetic factors that may enhance susceptibility to stress in one individual over another

• We examine how the timing and nature of stress varies in its impact on brain and behavior

Abbreviations

BDNF

brain-derived neurotrophic factor

BDNFmet

methionine in codon 66 of the BDNF protein

fMRI

functional magnetic resonance imaging

MRI

magnetic resonance imaging

Met

methionine

P

postnatal day

PI

previously institutionalized

SCR

skin conductance response

Val66met

valine-to-methionine substitution at codon 66

Val

valine

vmPFC

ventromedial prefrontal cortex