Skip Navigation

Link to  the National Institutes of Health  
The Science of Drug Abuse and Addiction from the National Institute on Drug Abuse Archives of the National Institute on Drug Abuse web site
Go to the Home page

The Intersection of Stress, Drug Abuse, and Development - Abstracts

Neonatal Stress as a Vulnerability Factor in Drug Abuse

Paul M. Plotsky, Ph.D.

Adverse early experience including abuse, neglect, or severe childhood illness serves as a major risk factor in development and may influence the vulnerability to a variety of physio- and psycho-pathologies. Clinically, this is reflected by a high rate of mood and anxiety disorders in patients reporting early adverse life events and in the rate of comorbidity for substance abuse in these patients. Utilizing a rodent model of neonatal maternal separation that exhibits many behavioral, neuroendocrine, and neurochemical features associated with major mood disorders as well as a vulnerability to substance abuse, we have started to evaluate the acquisition of alcohol and psychostimulant self-administration as well as stress and environmentally cued relapse after withdrawal. Animals exposed to neonatal maternal separation choose alcohol solutions in a two-bottle home cage choice test with no prior training. Under appropriate testing circumstances, these animals also show increased acquisition of cocaine self-administration with a left shift in their dose curve. Relapse is context dependent and can be triggered by either stressors or appropriate cues. This process appears to involve extrahypothalamic corticotropin-releasing factor (CRF) and noradrenergic neurocircuits that are sensitized by neonatal maternal separation.

Stress Models and Glutamatergic Mechanisms

James I. Koenig, Ph.D.

Glutamate is the most abundant excitatory amino acid neurotransmitter in the brain. However, the precise role of this neurotransmitter in the stress response is unclear. This presentation will discuss how glutamate modifies hypothalamic-pituitary-adrenal axis activity and the implications for neuropsychiatric disorders. Moreover, novel stress model systems will be considered.

Circadian Rhythms in the HPA Axis of Rats and Man Mary F. Dallman, Ph.D.

There is a well-known rhythm in basal activity in the HPA axis in which peak circulating glucocorticoid concentrations occur at a time just before or at the onset of the daily activity rhythm. In addition, however, regulatory control of activity in the HPA axis is also affected by the time of day at which it is examined. This presentation will include results of rat and human studies on stress responsiveness and glucocorticoid feedback as a function of time of day.


Dallman MF, Akana SF, Bradbury MJ, Strack AM, Hanson ES, Scribner KA. Regulation of the hypothalamo-pituitary-adrenal axis during stress: Feedback, facilitation and feeding. Sem Neurosci 6:1-9, 1994.

Akana SF, Dallman MF. Chronic cold in adrenalectomized, corticosterone (B)-treated rats: Facilitated corticotropin responses to acute restraint emerge as B increases. Endocrinology 138:3249-58, 1997.

Bhatnagar S, Dallman MF. Neuroanatomical basis for facilitation of hypothalamo-pituitary-adrenal responses to a novel stressor after chronic stress. Neuroscience 84:1025-39, 1998.

Corticotropin-Releasing Factor Engages Brain Biogenic Amine Systems During Stress

Rita J. Valentino, Ph.D.

Neuroanatomical and physiological findings implicate corticotropin-releasing factor (CRF) as a neurotransmitter in two biogenic amine nuclei that have been implicated in stress and stress-related psychiatric disorders, the noradrenergic nucleus, locus coeruleus, and the serotonergic dorsal raphe nucleus. Certain stressors release CRF in these nuclei to alter neuronal activity and consequently neurotransmitter release in targets of these systems. Moreover, neuroanatomical findings suggest that these nuclei may be coregulated by CRF from common afferents. Finally, CRF-biogenic amine interactions are plastic in that chronic stress selectively alters the sensitivity of these systems to CRF and subsequent stressors. This plasticity could play a role in the pathophysiology underlying psychiatric disorders including substance abuse.

Glutamatergic Regulation of the Corticolimbic Stress Response

Bita Moghaddam, Ph.D.

Stress has been known to activate dopamine neurotransmission in cortical and limbic regions. Disruption of this activation is thought to play a role in maladaptive responses to stress. Findings from our laboratory will be presented to indicate that glutamate neurotransmission regulates the dopamine stress response in the prefrontal cortex and amygdala. Implications of these findings are that (1) glutamate projections to these regions contribute to the executive and affective responses to aversive stimuli and (2) in a disease state, manipulations of glutamate-mediated transduction mechanisms may normalize aberrant stress responses.

Postnatal Neurogenesis and Stress

Elizabeth Gould, Ph.D.

The dentate gyrus of the hippocampal formation undergoes substantial neurogenesis during the postnatal period and into adulthood. The late development of this region may render its structure potentially vulnerable to adverse life experiences. Indeed, stressful experiences suppress the production of new granule neurons in the rodent and primate throughout life. Stress-induced inhibition of neurogenesis may alter the structure of the granule cell layer and, potentially, the function of the hippocampus.

What Stress Research Can Tell Us About the Origins of Drug Abuse

Pier V. Piazza, M.D., Ph.D.

Environmental experiences have an important effect on the sensitivity of an individual to drugs of abuse. Studies of drug self-administration in laboratory animals have shown that both physical and psychological stressors facilitate the acquisition of drug self-administration, probably by increasing the reinforcing efficacy of drugs of abuse. The adrenal hormones, glucocorticoids, which increase the sensitivity of mesencephalic dopaminergic neurons to drugs, seem to be one of the biological substrates of the effects of stress on the propensity to develop drug-intake. Stress research has contributed to showing that the biological status of an individual plays an important role in determining a propensity for developing drug self-administration and has highlighted the importance of environmental experiences in inducing a drug-prone phenotype.


Piazza PV, Le Moal M. Pathophysiological basis of vulnerability to drug abuse: Role of an interaction between stress, glucocorticoids and dopaminergic neurons. Ann Rev Pharm Tox 36:359-78, 1996.

Piazza PV, Le Moal M. Glucocorticoids as a biological substrate of reward: Physiological and pathophysiological implications. Brain Res Rev 25:359-72, 1997.

Piazza PV, Le Moal M. The role of stress in drug self-administration. TIPS 19:67-74, 1998.

The Role of Stress in the Reinstatement of Drug-Seeking

Jane Stewart, Ph.D.

Using an animal model of relapse to drug-taking, we have shown that acute exposure stress reinstates drug-seeking in rats trained to self-administer cocaine, heroin, and other drugs of abuse. Specific brain corticotropin-releasing factor (CRF) and noradrenaline systems are implicated in this effect. New evidence about the nature of the processes mediating this effect will be presented. Issues yet to be addressed are the following: Does a history of drug-taking make animals vulnerable to these effects of stress? Does early exposure to stress change the susceptibility of animals to these effects of stress?

Stress and Drug Addiction: Role of Corticotropin-Releasing Factor in the Central Amygdala

Friedbert Weiss, Ph.D.

Recent efforts to understand the role of brain stress systems in drug addiction have focused on the extrahypothalamic corticotropin-releasing factor (CRF) system in the central nucleus of the amygdala (CeA). This system is part of a neural circuitry regulating emotional responses to stress including anxiety. Stress-like symptoms and anxiety are an integral part of acute and protracted drug and alcohol withdrawal syndromes. Evidence is accumulating to suggest that these withdrawal signs involve activation of CRF neuronal mechanisms in the CeA and that this effect may be a common neurobiological element in acute withdrawal from all drugs of abuse. The CRF system in the CeA also shows long-term cyclical disruptions during protracted cocaine and ethanol withdrawal that may lead to enhanced susceptibility to stress. These changes, combined with disturbances in stress-regulatory responses at other levels such as altered responsiveness of central dopamine systems to stress, may be an important factor in long-lasting vulnerability to relapse.

Social Stress, Drug Abuse, and Gene Expression

Klaus A. Miczek, Ph.D.

Social conflict engenders significant immediate and long-term behavioral and neurobiological consequences that are relevant for pain, emotion, and drug abuse. At the pharmacological level, intense, unpredictable, and uncontrollable social stress engenders naltrexone-reversible analgesia, opiate tolerance, and withdrawal. Concurrently, behavioral sensitization to dopaminergic challenges becomes evident, and cocaine self-administration is initiated more rapidly and maintained at higher rates. Dopamine concentrations are immediately increased in nucleus accumbens and prefrontal cortex, but not in striatum. Repeated social confrontations are anticipated by increased dopamine activity in nucleus accumbens and followed by decreased serotonin. Within less than 1 hour of the social confrontation, the immediate early gene c-fos is expressed in periaqueductal grey area, dorsal raphe nucleus, and locus coeruleus. These molecular changes may be critical for the development of tolerance to opioid analgesia, behavioral sensitization to psychomotor stimulants, and increased psychomotor self-administration.

Stress, Craving, and Relapse to Drug Use

Rajita Sinha, Ph.D.

A common clinical observation is that stress leads to relapse to drug use in humans. However the mechanisms by which stress is linked to relapse or continued drug use in humans have not been well investigated. This presentation will cover human laboratory studies examining the association between stress-induced craving and drug cue-induced craving and treatment studies that show a positive association between stress and continued drug use. Finally, new data on the potential link between childhood stressors, recent stressful life events, and laboratory-induced craving states will be presented. Future research directions in the area of stress and relapse to drug use in humans will be outlined.

Neuroendocrinology of Stress: Developmental Aspects

George P. Chrousos, M.D., Sc.D.

Stress activates the central and peripheral components of the stress system, i.e., the hypothalamic-pituitary-adrenal (HPA) axis and the arousal/sympathetic system. The principal effectors of the stress system are corticotropin-releasing hormone (CRH), arginine vasopressin, the proopiomelanocortin-derived peptides melanocyte-stimulating hormone and endorphin, the glucocorticoid, and the catecholamines norepinephrine and epinephrine. The developing brain undergoes rapid growth and is characterized by high turnover of neuronal connections during prenatal and early extrauterine life. These processes, and hence brain plasticity, slow down during childhood and puberty and plateau in young adulthood. Hormonal actions in early life, and to a much lesser extent later, can be organizational, i.e., can have effects that last for long periods of time, frequently for the entire life of the individual. Hormones of the stress system and sex steroids have such effects, which influence the behavior and certain physiologic functions of individuals for life. Exposure of the developing brain to severe and/or prolonged stress may result in hyperactivity/hyperreactivity of the stress system, with resultant amygdala hyperfunction (fear reaction), decreased activity of the hippocampus (defective glucocorticoid negative feedback, cognition) and the mesocorticolimbic dopaminergic system (dysthymia, novelty seeking, addictive behaviors), hyperactivation of the HPA axis (hypercortisolism), suppression of reproductive, growth, thyroid, and immune functions, and changes in pain perception.

These changes may be accompanied by abnormal childhood, adolescent, and adult behaviors, including excessive fear ("inhibited child syndrome") and addictive behaviors, dysthymia and/or depression, and gradual development of components of the metabolic syndrome X, including visceral obesity and essential hypertension. Prenatal stress exerted during the period of sexual differentiation may be accompanied by impairment of this process, with behavioral and/or somatic sequelae. The vulnerability of individuals to develop varying degrees and/or components of the above lifelong syndrome is defined by as yet unidentified genetic factors, which account for up to 60 percent of the variance. CRH has marked kindling and glucocorticoids have strong consolidating properties; hence both of these hormones are crucial in the development of, and can each alone produce, the above syndrome. CRH and glucocorticoids may act in synergy, as in acoustic startle, while glucocorticoids may suppress or stimulate CRH, as in the hypothalamus and amygdala, respectively. A CRH receptor type 1 antagonist, antalarmin, inhibits both the development and expression of conditioned fear in rats and has anxiolytic properties in monkeys. Profound stressors, such as those from sexual abuse, rape, etc., may elicit the syndrome in older children, adolescents, and adults as well. Most frequently, chronic dysthymia and/or depression may ensue, associated with gastrointestinal complaints and/or the premenstrual tension syndrome. A lesser proportion of individuals may develop the classic posttraumatic stress disorder characterized by hypocortisolism and intrusive and avoidance symptoms; in younger individuals, it may present as dissociative personality disorder.

Maternal Deprivation, Brain Stress Circuits, and Drug Abuse

Delia M. Vázquez, M.D.

Stress experiences in adulthood appear to have a strong influence on susceptibility to drug-taking in various animal models and have also been observed in clinical populations suffering from substance abuse. From the animal data, it is evident that repeated stress not only increases the propensity to drug-taking, but also augments the psychomotor effects of psychostimulants, a phenomenon that is termed behavioral sensitization. For the most part, stress-induced behavioral sensitization is dependent on glucocorticoid hormone action (Piazza and Le Moal, 1996). Central limbic-hypothalamic-pituitary-adrenal (LHPA) axis mechanisms (possibly corticotropin-releasing hormone [CRH]) also appear to be critical. It is apparent that an increased secretion of glucocorticoids or higher sensitivity to the effects of this hormone, either naturally present in certain individuals or induced by stress in others, increases the vulnerability to develop drug intake, possibly through an enhancement of the activity of mesencephalic dopaminergic neurons. In addition, impulsivity and risk-taking behavior, vulnerability traits for drug abuse, have been linked to serotonergic mechanisms. A number of clinical and epidemiological studies have found a strong association between psychosocial stressors early in life (e.g., parental loss, child abuse) and an increased risk for depression, anxiety, impulsive behavior, and substance abuse in adulthood (Kendler et al., 1992). Thus, intrinsic vulnerability to drug intake may be influenced by the exposure to adverse early life experience that may lead to a permanent alteration of the normal physiology of the LHPA and other neurotransmitter systems that mediate drug abuse. Studies using rodents also suggest that early experience may induce long-lasting modifications on individual predisposition to self-administration of psychostimulants. Specifically, prenatal stress is associated with prolonged CORT secretion in response to stress and increases both locomotor reactivity to amphetamine and amphetamine self-administration in adult offspring.

While these are the consequences of prenatal stress, the effect of postnatal life stresses on behavioral sensitization to drugs has received little attention. It may not be coincidence that in humans, early traumatic events and higher life stress scores are associated with early initiation of substance use (Dinwiddie, Reich, and Cloninger, 1992; Duncan, 1977; Gutierres, Molof, and Ungerleider, 1994). This suggests that there may be critical developmental periods that may increase vulnerability to drug abuse. Understanding what is different during development is a critical first step to formulating preclinical studies and prevention and therapeutic strategies. A system that is relevant because of its developmental characteristics and plasticity is the stress-responsive LHPA axis. The brain serotonin (5HT) system is also of relevance because it shares structural components with the LHPA axis, and alterations in both of these systems underlie psychological and behavioral abnormalities. The developing LHPA axis differs from the adult axis in both morphology and function. Most of our understanding of this period is derived from rat studies. In the rodent, there is a delay in the development of stress-related circuits and a change in adrenal morphology and function. Days 3 to 14 of life are characterized by a "silent period" during which the developing animal has a reduced peripheral hormonal response to stress (stress hypo-responsive period [SHRP]) (De Kloet et al., 1988; Sapolsky and Meaney, 1986a; Vázquez and Akil, 1992). The SHRP serves the function of maintaining low glucocorticoid levels at a time of maximal physical growth. Maternal factors appear to exert a strong regulatory role in several physiological processes in the infant rat and in primates (Kuhn, Butler, and Schanberg, 1978; Kuhn, Pauk, and Schanberg, 1990). Maternal care has an inhibitory effect on the infantŐs LHPA system (Suchecki, Rosenfeld, and Levine, 1993), and removing the mother from the home cage results in infant rats that are exquisitely responsive to mild stressors at a time when they should not be responsive (Rosenfeld, Suchecki, and Levine, 1993). Moreover, long-term maternal deprivation during postnatal day 3 to 4 vs. postnatal day 11 to 12 led to opposite hormonal profiles. As juveniles, an increased responsiveness to stress is observed in the animals deprived from day 11 to 12 of life, whereas decreased responsiveness is seen in the 3- to 4-day-old deprived animals. Therefore, as the external source of regulation (i.e., the mother) is removed, the inability of the developing organism to self-regulate has consequences that are specific to the developmental window in which the insult occurs. The result is either an increase or a decrease of the stress responsiveness and expressing or suppressing fearful behavior (Meaney et al., 1993a; Plotsky and Meaney, 1993).

We have started to explore the maternal deprivation rat model with the goal of understanding the presence of critical developmental periods after birth, which lead to alteration in the stress response and sensitization to the effects of psychostimulants. We use a rat model in which pups are deprived of maternal care for 24 hours during particular days within the first 2 weeks of life, when the adrenal response to stress is quiescent but brain response to stress is active. We investigated the regulation of molecules related to the LHPA axis, the extrahypothalamic CRH stress circuit, and the serotonin (5-HT) system. These molecules and brain circuits are believed to play a role in mood, impulsivity, and anxiety, comorbid traits in clinical populations at risk for drug abuse. We concentrated on one developmental age, postnatal (PN) day 13, to explore these systems. CRH is a neurotransmitter that mediates behavior associated with stress and anxiety, and it exerts its action in brain areas within and outside the hypothalamus. We thus studied the effect of maternal deprivation on this system. We focused on areas that are rich in CRH neurons or have reciprocal CRH connections, namely, frontal cortex, amygdaloid complex, ventromedial hypothalamic nucleus (VMH) and the hippocampus. We concentrated on the modulation of CRH itself and its receptors, CRHr1 and CRHr2. The results were complex and varied depending on the anatomical area. We found that maternal deprivation caused a significant increase in CRH, CRHr1, and CRHr2 mRNA in frontal cortex. In the central nucleus of the amygdala, maternal deprivation did not alter CRH mRNA levels. However, consistent effects were detected on the receptors: CRHr1 and r2 gene expression decreased in this nucleus. CRHr2 mRNA also decreased in VMH, where this receptor has been linked to anorexic effects of CRH. Significant increases of CRHr2 were detected over the pyramidal cell region of the hippocampus. Interestingly, chronic injection treatment, as a chronic intermittent stressor, potentiated the up-regulation effect of maternal deprivation in cortex (for CRH mRNA) and over hippocampal CA3-4 region (CRHr2 mRNA). One week of administration of the tricyclic antidepressant desipramine prevented all the maternal deprivation-induced changes in CRHr2 gene expression, regardless of the direction of the change. Desipramine had no effect on CRHr1. This suggests that regulation of CRHr2 gene expression is mediated by different mechanisms than CRHr1. These results also reveal that the modulation of the infantŐs CRH brain system is not unidirectional or at random, but rather site and molecule specific, suggesting neuronal interactions between limbic efferent pathways. Maternal deprivation caused exclusively postsynaptic effects in the serotonin brain system. We observed an increase in the gene and protein expression of the serotonin 2A receptor (5-HT2A) in cortex and in the serotonin 1A receptor (5-HT1A) in the CA1 region of the hippocampus. Maternal deprivation also caused an increase in serotonin 1B receptor (5-HT1B) mRNA in the CA1 region. In contrast, no effects were seen in the mRNA levels of the somatodendritic 5-HT1A receptor and the 5-HT transporter in dorsal raphe. Although desipramine reduced the adrenocortical stress response in the deprived animal, this agent did not alter the maternal deprivation-induced changes in 5-HT1A and 5-HT2A receptors. However, desipramine did prevent the 5-HT1B increase in the deprived animals. It is of interest that chronic injection caused a down-regulation of 5-HT1B receptor levels in the nondeprived animals. We later observed an increase in alcohol consumption in adult animals that received chronic injection treatment as infants. Functionally, the 5-HT1B postsynaptic receptor is associated with impulsive and aggressive behavior (Saudou et al., 1994; Zhuang et al., 1999).

At this point, we speculate that these changes in stress brain circuits and serotonin receptor system may alter the developmental program and increase the vulnerability to maladaptive stress behaviors later in life. Are there developmental windows that render the developing organism more susceptible to maladaptive stress responses and behaviors conducive to drug abuse later in life? We investigated animals deprived during PN day 3 to 4 vs. PN day 11 to 12 since deprivation at these times led to opposite hormonal profiles at 40 days of age. At this age, animals that were maternally deprived on day 11 of life showed significantly increased ACTH levels after novel environment exposure compared with PN 3 (p< .005; n = 8 per group). The plasma CORT levels of both the PN 3 and PN 11 deprived animals were significantly lower than the nondeprived unhandled controls (p< .05). PN 3 animals had lower CORT levels than the PN 11 animals. We also found that animals deprived on PN 11 were 25 to 30 percent more active in a novel environment than those animals deprived on PN 3 (p< .05; n = 12 per group). However, PN 11 deprived animals were not different from nondeprived unhandled controls. Starting at 50 days of age, the individual vulnerability to sensitization to the psychomotor effects of cocaine was tested. On the first day of drug treatment, we found that deprivation during PN 3 to 4 was protective for male and female animals, which did not increase their locomotor activity to a low cocaine dose (7.5 mg/kg). However, significant motor activity was observed in female animals deprived from PN 11 to 12 (p< .005; n = 6 per group). After 7 days of chronic cocaine treatment, followed by 10 days of withdrawal, animals were challenged with an acute dose of cocaine. Both groups of deprived animals showed lower sensitization when compared with the unhandled group. Therefore, the groups of animals showing low CORT response to novelty also exhibit lower sensitization to cocaine.

It thus appears that there is a protective window of time very early in the life of the infant rodent. It also appears that effects by early life adverse events on predisposition to drug use may depend on the timing and chronicity of the event. The gender of the animal is also an important variable. Is there a similar biological phenomena in the developing human? We have begun to study children who were victims of childhood neglect. Based on epidemiological studies, these children are at risk for psychopathology and substance abuse (Kendler et al., 1992). A complete endocrinological evaluation is performed in our institution as part of an evaluation for growth failure. We have found that compared with normal short-stature children, children with a history of neglect fail to effectively "shut down" the ACTH and cortisol response to insulin hypoglycemia. Failure to effectively inhibit ACTH secretion is also observed when a constant cortisol infusion is administered. These data suggest that in children with a significant history of neglect there is a failure to inhibit the CRH drive once initiated. It would be important to follow these high-risk children for whom we have biological markers to determine whether these predict future development of substance abuse.


De Kloet ER, Rosenfeld P, Van Eekelen AM, Sutanto W, Levine S. Stress, glucocorticoids and development. Progress in Brain Research 73:101-20, 1988.

Dinwiddie SH, Reich T, Cloninger CR. Prediction of intravenous drug use. Compr Psychiatry 33:173-9, 1992.

Duncan DF. Life stress as a precursor to adolescent drug dependence. Int J Addict 12:1047-56, 1977.

Gutierres SE, Molof M, Ungerleider S. Relationship of "risk" factors to teen substance use: A comparison of abstainers, infrequent users, and frequent users. Int J Addict 29:1559-79, 1994.

Kendler KS, Neale MD, Kessler RC, Heath AC, Eaves LJ. Childhood parental loss and adult psychopathology in women: A twin study perspective. Arch Gen Psychiatry 49:109-16, 1992.

Kuhn CM, Butler SR, Schanberg SM. Selective depression of serum growth hormone after maternal deprivation in rat pups. Science 201:1034-36, 1978.

Kuhn CM, Pauk J, Schanberg SM. Endocrine responses to mother-infant separation in developing rats. Psychobiology 23:395-410, 1990.

Meaney MJ, Bhatnagar S, Larocque S, McCormick C, Shanks N, Sharma S, Smythe J, Viau V, Plotsky PM. Individual differences in the hypothalamic-pituitary-adrenal stress response and the hypothalamic CRF system. Ann NY Acad Sci 69:70-85, 1993a.

Piazza PV, Le Moal M. Pathophysiological basis of vulnerability to drug abuse role of an interaction between stress, glucocorticoids, and dopaminergic neurons. Ann Rev Pharmacol Toxicol 36:359-78, 1996.

Plotsky PM, Meaney MJ. Early, postnatal experience alters hypothalamic corticotropin-releasing factor (CRF) mRNA, median eminence CRF content and stress-induced release in adult rats. Molecular Brain Research 18:195-200, 1993.

Rosenfeld P, Suchecki D, Levine S. Multifactorial regulation of the hypothalamic-pituitary-adrenal axis during development. Neuroscience and Biobehavioral Rev 16:553-68, 1992.

Sapolsky RM, Meaney MJ. Maturation of the adrenocortical stress response: Neuroendocrine control mechanisms and the stress hyporesponsive period. Brain Research Reviews 11:65-76, 1986a.

Saudou F, Aït Amara D, Dierich A, Lemeur M, Ramboz S, Segu L, Buhot MC, Hen R. Enhanced aggressive behavior in mice lacking 5-HT1B receptor. Science 265:1875-78, 1994.

Suchecki D, Rosenfeld P, Levine S. Maternal regulation of the hypothalamic-pituitary-adrenal axis in the infant rat: The roles of feeding and stroking. Dev Brain Res 75:185-92, 1993.

Vázquez DM, Akil H. Development of pituitary pro-opiomelanocortin gene and peptide expression: Characterization and effect of repeated intermittent maternal isolation. Neuroendocrinology 56:320-30, 1992.

Zhuang X, Gross C, Santarelli L, Compan V, Trillat AC, Hen R. Altered emotional states in knockout mice lacking 5-HT1A or 5-HT1B receptors. Neuropsychopharmacology 21:52S-60S, 1999.

Maternal Opiate Exposure: Long-Term CNS Consequences in the Offspring

Ilona Vathy, Ph.D.

The problem of drug abuse among pregnant women is of major concern. Many abused drugs, including opiates, can cross the placenta to affect the development of the central nervous system (CNS). The influence of a specific drug on the developing brain is dependent on the presence of specific cellular recognition sites for the drug, the maturational state of the receptive sites, and the maturational state of neural circuitry. The observation that the highest concentrations of opioid receptors are present in several brain areas including the limbic system, thalamus, striatum, hypothalamus, midbrain, and spinal cord suggests that physiological mechanisms other than analgesia and pain perception could be affected by opiate exposure. Thus, during development, opiates to which an organism is exposed become part of its chemical environment and may act to perturb homeostatic balance in those systems that are developing at the time of drug exposure. We have evidence that prenatal opiate exposure alters several CNS systems, including opioids, catecholamines, excitatory amino acids, and GABA.

It is also important to note that the rate and timing of brain development as well as the hormonal background are different in the two sexes (2). Thus, these differences could subsequently produce differential responses of developing males and females to prenatal drug exposure. There is also evidence that catecholamine neurons differ in developing male and female rats beginning as early as gestation days 14-16 (7, 20). Dopamine (DA) and norepinephrine (NE) neurons may develop morphological and functional sex differences in the absence of sex steroids (3, 4, 5, 6, 19, 20). Our work consistently demonstrates that the alterations in the opioid, catecholamine, excitatory amino acid, and GABA systems are sex-dependent. In all of our studies reviewed below, animals are injected with morphine (10 mg/kg) or saline twice daily during gestation days 11-18 (for details see 26, 28). Brain neurochemistry and behaviors are examined in adulthood unless stated differently.

Prenatal Morphine Exposure Affects Adult Brain NE and Opioid Systems

Exposure to morphine during mid to late gestation induces long-term, sex-specific neurochemical alterations in young adult male and female rats. In male rats, prenatal morphine exposure increases NE content and turnover in the hypothalamus (30, 32) and increases tyrosine hydroxylase-immunoreactivity (TH-IR) in the locus coeruleus (LC) and in the paraventricular nucleus of the hypothalamus (PVN; 29). In female rats, these indices are all decreased (30, 32). It is noteworthy that these long-lasting, sex-specific changes in NE neurotransmission are localized to brain regions known to participate in the regulation of stress responses. NE neurons in the LC mediate the neural responses to stress, and corticotropin-releasing hormone neurons in the PVN are the central neuroendocrine regulators of the hypothalamic-pituitary-adrenocortical (HPA) axis.

The same prenatal morphine exposure also affects the endogenous opioid systems. In male rats, prenatal morphine exposure decreases pro-enkephalin (pENK) mRNA levels in the medial preoptic area (mPOA) relative to controls. In all other brain regions, pENK mRNA levels are comparable in the control and drug-exposed males. As reported by others, estrogen (E) administration to ovariohysterectomized (OVX) control females increases pENK mRNA in the ventromedial nucleus of the hypothalamus (VMH). Prenatal morphine exposure blocks the E-induced increase of pENK mRNA in the VMH. No changes are found in any other brain regions of morphine-exposed females. (10, 11).

The sex-specific effects of prenatal morphine exposure on the binding characteristics of mu, delta, and kappa receptors are also examined in brain membrane homogenates (21). Prenatal morphine has no effects on either the density or affinity of mu, delta, or kappa opioid receptors in any brain region of exposed males. In contrast, prenatal morphine exposure alters the binding of all three major opioid receptor subtypes in females. The alterations in opioid receptor binding in females are receptor subtype and site specific. Mu binding: In OVX females, prenatal exposure to morphine reduced the Bmax of mu opioid receptors in the hypothalamus. E increases the density of mu receptors in this brain region of drug-exposed, but not control, females. The binding capacity of mu opioid receptors is not altered in the POA, ventral tegmental area (VTA), cortex, striatum, or cerebellum of females. Delta binding: Prenatal morphine exposure increases the density of delta receptors in the cortex and striatum and decreases it in the VTA of OVX female rats. E increases the density of delta receptors in the cortex and the striatum of control but not morphine-exposed females. Additionally, our most recent receptor autoradiography study demonstrates that prenatal morphine exposure differentially alters gonadal hormone regulation of delta opioid receptor density in the cortex of gonadectomized male and female rats (25). Kappa binding: E administration increased kappa opioid receptors in all six brain regions of control females. E increases the density of kappa opioid receptors in the hypothalamus, POA, and cerebellum, but not in the cortex or the VTA of morphine-exposed females.

Prenatal Morphine Exposure Affects Adult Behaviors/Functions

Gonadal hormone-dependent sexual behaviors. There is evidence for interactions between endogenous opioid peptides and catecholamine neurotransmitters. Behaviors that are stimulated by catecholamines, including sexual behaviors, are often inhibited by endogenous and exogenous opiates (14, 15, 18, 27, 34). These inhibitory actions of opiates are thought to result in part from tonically active, endogenous µ opioid inputs to NE terminals (12, 15). Prenatal morphine inhibits lordosis (reproductive) behavior without altering estrous cyclicity in young adult female rats (28). In contrast, morphine-exposed male rats display somewhat enhanced copulatory activities, including more frequent mounting and intromitting behaviors and significantly shorter post-ejaculatory intromission latencies (PEI; 26, 28). More frequent mounting and shorter PEIs may reflect increased sexual motivation.

Gonadal hormone-independent spinal sexual reflexes. Because in utero exposure to morphine differentially alters adult sexual behaviors, which are dependent on circulating gonadal hormones in male and female rats, we also investigated the urethrogenital (UG) reflex in adult offspring (31). The UG reflex is a hormone-independent, spinal sexual reflex in both male and female rats. This reflex is tonically inhibited by supraspinal neurons but can be consistently evoked by urethral stimulation in spinally transected animals. Prenatal morphine exposure does not remove the supraspinal inhibition of the UG reflex in spinally intact male or female rats. In spinally transected animals, the UG reflex is qualitatively similar in all groups. However, morphine-exposed males require a stronger stimulation to elicit the UG reflex, but once elicited, the frequency of UG reflex is significantly higher compared to control males. In contrast, in female rats, prenatal morphine exposure has no effects on the UG reflex. Thus, the UG reflex appears to be differentially affected by prenatal exposure to morphine in males and females.

Learning and memory. Clinical studies report negative effects of prenatal drug exposure on cognitive and neurobehavioral development during the neonatal period, childhood, and adolescence (8, 9). Preclinical studies show hippocampal involvement in learning and memory in animals (24). We have evidence that prenatal morphine exposure decreases long-term potentiation in hippocampal slices in male rats (37). Recently, we used the eight-arm radial maze to evaluate spatial working memory, a hippocampal-dependent task, in prenatally morphine- and saline-exposed male and female rats (24). All female rats were OVX, and male rats were gonadally intact. After training, animals were tested once per week on 3 consecutive weeks. Both male and female morphine-exposed animals needed more time to complete each trial than controls, suggesting that working memory is impaired in these animals. Interestingly, morphine exposure did not affect reference memory in male rats but hindered it in OVX female rats. In OVX females, replacement of both E and progesterone (P) were able to restore the impairment of reference memory produced by prenatal morphine exposure. Thus, prenatal morphine exposure differentially alters spatial working and reference memory in male and female rats.

Prenatal Morphine Exposure Affects Seizure Susceptibility

Excitatory amino acids. Because prenatal morphine exposure alters the endogenous opioid and NE systems, and both NE and opioids play a significant role in seizure susceptibility and generation, we asked whether prenatal morphine exposure affects seizures induced by several convulsant agents in male and female rats (22, 23). Several doses of N-methyl-D-aspartate (NMDA) were administered to prenatally saline- and morphine-exposed adult male and female rats. Another group of control and morphine-exposed adult male and female rats were injected with kainate (KA). In the NMDA group latency to onset of stereotypy and seizures and in the KA group the latency to onset of wet dog shakes and seizures were recorded. In male rats, the lowest dose of NMDA induced seizures in the control males but not in the morphine-exposed males. In contrast, morphine-exposed males demonstrated higher sensitivity to 200 mg/kg of NMDA than controls. A systemic injection of KA in morphine-exposed males elicited a shorter latency to onset of stereotypy but had no effects on clonic seizures. In female rats, prenatal morphine exposure increased the latency to onset of stereotypy only in OVX, E-injected females after NMDA administration. Prenatal morphine exposure did not alter NMDA-induced seizures in either OVX or OVX, E-injected females. An E injection increased the latency to onset of NMDA-induced seizures in both saline- and morphine-exposed females. Morphine exposure increased the latency to onset of stereotypy in both OVX and OVX, E-injected females after KA administration. However, it increased the latency to onset of KA-induced seizures only in OVX females. An E injection decreased the latency to onset of KA-induced seizures in both saline- and morphine-exposed females. Thus, prenatal morphine exposure differentially alters the sensitivity to excitatory amino acid-induced seizures in male and female rats.

Epileptiform activity in vitro was evoked in entorhinal cortex (EC)-hippocampal slices by removing magnesium from the perfusion solution, which is known to unblock the NMDA subtype of glutamate receptors (35, 36). EC-hippocampal slices of morphine-exposed male rats have shorter latency to onset of initial seizure activity and recurrent discharges under low magnesium conditions. Seizure activity is not affected by prenatal morphine in females. Thus, prenatal morphine exposure differentially alters low magnesium-induced seizure-like events in male and female rats.

GABA is the major inhibitory neurotransmitter in the CNS, and chronic morphine exposure in adult animals alters the GABAergic system (13). Alterations in the GABAergic system are associated with changes in seizure susceptibility (16). To investigate whether prenatal morphine exposure alters the GABAergic system, we used bicuculline, an antagonist at GABAA receptors, to monitor changes in seizure susceptibility of prepubertal and peripubertal (17) male and female rats. These ages were chosen because changes in seizure susceptibility are often observed around the time of puberty (1). Morphine-exposed males showed increased susceptibility at prepubertal ages and decreased susceptibility at peripubertal ages. There were no differences in seizure susceptibility in morphine-exposed females at either age.

Effects on Drug Abuse Liability

To evaluate the effects of prenatal morphine exposure on the propensity for drug abuse, electrical brain-stimulation reward methods were used in a recent pilot study. One week after a monopolar stainless steel electrode was implanted in the medial forebrain bundle, adult prenatally saline- and morphine-exposed male rats were tested in an operant chamber modified for delivery of brain stimulation. Once the reward thresholds were stable, animals were injected with 5 mg/kg cocaine. The maximum response rate and M50 were analyzed. Our preliminary data indicate that prenatal morphine exposure significantly enhances the rewarding effects of an acute cocaine injection. Therefore, our preliminary data suggest that prenatal morphine exposure alters the way that animals respond to illicit drugs in adulthood, which might increase the propensity of progeny to abuse drugs.


Our neurochemical and behavioral results suggest that prenatal morphine exposure has widespread, long-term, sex-specific effects on the CNS of exposed progeny. Some of our data suggest that neural systems mediating stress responses are affected by prenatal morphine exposure. Therefore, it is of utmost importance to fully understand the long-term consequences of prenatal drug exposure for neurobehavioral development. This may permit us to develop methods to prevent potentially life-threatening effects such as increased propensity for drug abuse in future generations.


  1. Backstrom T, Rosciszewka D. Effects of hormones on seizure expression. In: JJ Engel and T Pedley (eds.), Epilepsy: A Comprehensive Textbook. Lippincott-Raven Publisher, Philadelphia, pp. 2003-11, 1997.

  2. Beyer C, Feder HH. Sex steroids and afferent input: Their role in brain sexual differentiation. Annual Reviews in Physiology 49:349-64, 1987.

  3. Engele J, Pilgrim C. Nigrostriatal dopaminergic neurons in culture are sexually dimorphic. Neuroscience 22:S235, 1987.

  4. Engele J, Pilgrim C. Early sexual differentiation of diencephalic dopaminergic neurons of the rat in vitro. Cell Tissue Research 255:411-17, 1989.

  5. Engele J, Pilgrim C, Kirsch M, Reisert I. Different developmental potentials of diencephalic and mesencephalic dopaminergic neurons in vitro. Brain Research 483:98-109, 1989a.

  6. Engele J, Pilgrim C, Reisert I. Sexual differentiation of mesencephalic neurons in vitro: Effects of sex and gonadal hormones. International Journal of Developmental Neuroscience 7:603-11, 1989b.

  7. Eusterschulte B, Reisert I, Pilgrim C. Sex differences in the development of the Wolffian duct. A parameter for early gonadal testosterone secretion. Verhandlungen der Anatomischen Gesellschaft (Jena) 86:76, 1991.

  8. Eyler FD, Behnke M. Early development of infants exposed to drugs prenatally. Clin Perinatol 26(1):107-50, 1999.

  9. Hans S. Maternal opioid drug use and child development. In: I Zagon and T Slotkin (eds.), Maternal Substance Abuse and the Developing Nervous System. Academic Press, London, pp. 177-213, 1992.

  10. Hnatczuk OC, Vathy I. Expression of proenkephalin mRNA in adult male and female rats treated prenatally with morphine. Abstr Soc Neurosci 23(2):2148, 1997.

  11. Hnatczuk OC, Vathy I. Expression of proenkephalin and prodynorphin mRNA in the dentate gyrus of adult rats treated prenatally with morphine. Abstr Soc Neurosci 24(2):1489, 1998.

  12. Kalra SP, Kalra PS. Opioid-adrenergic connection in regulation of luteinizing hormone secretion in rat. Neuroendocrinology 38:418-26, 1984.

  13. Lopez et al., 1990.

  14. Meyerson B, Malmnas C. Brain monoamines and sexual behaviour. In: J Hutchinson (ed.), Biological Determinants of Sexual Behavior. John Wiley and Sons, New York, pp. 165-91, 1978.

  15. Mitchell V, Beauvillian JC, Poulain P, and Mazzuca M. Catecholamine and male sexual behaviour. European Journal of Pharmacology 42:191-92, 1977.

  16. Morimoto K. Seizure-triggering mechanisms in the kindling model of epilepsy: Collapse of GABA-mediated inhibition and activation of NMDA receptors. Neurosci Biobehav Rev 13:253-60, 1989.

  17. Ojeda S, Urbanski H. Puberty in the rat. In: E Knobil and J Neill (eds.), The Physiology of Reproduction, Vol. 2. Raven Press, New York, pp. 363-409, 1994.

  18. Pfaus JG, Gorzalka BB. Opioids and sexual behavior. Neuroscience and Biobehavioral Reviews 11:1-34, 1987.

  19. Reisert I, Engele J, Pilgrim C. Early sexual differentiation of diencephalic dopaminergic neurons of the rat in vitro. Cell Tissue Research 255:411-17, 1989.

  20. Reisert I, Pilgrim C. Sexual differentiation of monoaminergic neurons - genetic or epigenetic. Trends in Neurosciences 14(10):468-73, 1991.

  21. Rimanóczy A, Vathy I. Prenatal exposure to morphine alters brain µ opioid receptor characteristics in rats. Brain Research 690:245-48, 1995.

  22. Slamberová R, Veli_ek L, Vathy I. Prenatal morphine exposure alters N-methyl-D-aspartate and kainate-induced seizures in adult male rats. Pharmacol Biochem Behav, in press.

  23. Slamberová R, Vathy I. Estrogen differentially alters NMDA- and kainate-induced seizures in prenatally morphine- and saline-exposed adult female rats. Pharmacol Biochem Behav, in review.

  24. Slamberová R, Schindler CJ, Pometlová M, Urkuti C, Purow-Sokol JA, Vathy I. Prenatal morphine exposure differentially alters learning and memory in male and female rats. Horm Behav, in review.

  25. Slamberová R, Rimanóczy A, Vathy I. Prenatal morphine exposure alters gonadal hormone regulation of delta opioid receptors in rats. Abstr Soc Neurosci 25(2):883, 1999.

  26. Vathy I. Effects of prenatal morphine and cocaine on postnatal behaviors and brain neurotransmitters. NIDA Research Monograph Series 158:88-114, 1995.

  27. Vathy I, Etgen AM. Hormonal activation of female sexual behavior is accompanied by hypothalamic norepinephrine release. Journal of Neuroendocrinology 1:383-88, 1989.

  28. Vathy I, Etgen AM, Barfield RJ. Effects of prenatal exposure to morphine on the development of sexual behavior in rats. Pharmacology, Biochemistry and Behavior 22:227-32, 1985.

  29. Vathy I, Huang-Jun He, Iodice M, Hnatczuck OC, Rimanóczy A. Prenatal morphine exposure differentially alters TH-immunoreactivity in the stress-sensitive brain circuitry of adult male and female rats. Brain Res Bull, in press.

  30. Vathy I, Kátay L. Effects of prenatal morphine on adult sexual behavior and brain catecholamines in rats. Developmental Brain Research 68:125-31, 1992.

  31. Vathy I, Marson L. Effects of prenatal morphine and cocaine exposure on spinal sexual reflexes in male and female rats. Physiology and Behavior, in press, 1998.

  32. Vathy I, Rimanóczy A, Eaton RC, Kátay L. Alterations in regional brain catecholamine turnover following prenatal exposure to morphine. Brain Research 662:209-15, 1994.

  33. Vathy I, Rimanóczy A, Eaton RC, Kátay L. Sex dimorphic alterations in postnatal brain catecholamines after gestational morphine. Brain Research Bulletin 36(2):185-93, 1995.

  34. Vathy I, van der Plas J, Vincent PA, Etgen AM. Intracranial dialysis and microinfusion studies suggest that morphine may act in the ventromedial hypothalamus to inhibit female sexual behavior. Hormones and Behavior 25:354-66, 1991.

  35. Velisek L, Stanton PK, Moshé SL, Vathy I. Prenatal morphine exposure enhances seizure susceptibility but suppresses long-term potentiation in the limbic system of adult male rats. Brain Res, in review.

  36. Velisek L, Vel’skova J, Moshé SL, Vathy I. Prenatal morphine exposure alters ovarian steroid regulation of seizure susceptibility. Brain Res 796:247-56, 1998.

  37. Vathy I, Velisek L. Expression of synaptic potentiation in adult male rats exposed prenatally to morphine. Abstr Soc Neurosci 25(2):596, 1999.

The Impact of Prenatal Stress, Fetal Alcohol Exposure, or Both on Development: Perspectives From a Primate Model

Mary L. Schneider, Ph.D.

The increased prevalence of substance abuse and violence in American families has resulted in a growing concern about the effects of fetal alcohol exposure and prenatal stress on child outcome. It is well established that consumption of large quantities of alcohol during pregnancy causes harm to the fetus (Clarren and Smith, 1978; Jones, Smith, Ulleland, and Streissguth, 1973; Streissguth, Clarren, and Jones, 1985); however, there are limited data on the effects of moderate-level alcohol consumption on child outcome (Jacobson and Jacobson, 1994). Furthermore, whether the effects from alcohol exposure in utero are exacerbated by maternal psychological stress has not been established. These studies are difficult to conduct in humans because of the confounding of variables of substance abuse, prenatal care, poverty, and postnatal environments. Rodent studies have provided identification of a number of relevant mechanisms that underlie the developmental effects associated with fetal alcohol exposure or prenatal stress, under carefully controlled laboratory settings (Sulik and Charness, 1997; Weinstock, 1997). Primate studies can add to our understanding by providing the opportunity not only to systematically manipulate the timing, intensity, and duration of exposure to alcohol or stress but also to study the offspring developmentally using tests and measurements that are similar to those used with humans.

This study examined rhesus monkeys derived from one of four groups of females: (1) alcohol-consuming, 0.6 g/kg/day, daily throughout gestation, (2) exposed to chronic unpredictable psychological stress (three random noise bursts) from gestation day 90 to 145, (3) alcohol-consuming (as above) and prenatally stressed (as above), and (4) sucrose-consuming, equivolemic and equicaloric to the alcohol solution. Compared with controls, infants from alcohol-consuming females were impaired on early attention, motor maturity, and state regulation, even though the infants were normal in birth weight, gestation length, and facial dimensions. Moreover, all alcohol-consuming females produced viable offspring; however, alcohol accompanied by stress during gestation resulted in 23 percent fetal losses and more severe motor impairments (Schneider, Roughton, and Lubach, 1997). At 6 months of age, when the monkeys were separated from their mothers and put into peer groups, the monkeys from alcohol and alcohol + stress conditions showed enhanced hormonal responses to a stressful event. In peer groups, infants from the alcohol condition showed locomotor hyperactivity. Adolescent offspring were tested in the Wisconsin General Test Apparatus (WGTA) on the number of trials to reach 90 percent criterion for Delayed Nonmatching to Sample (DNMS). A main effect for alcohol but not for stress was found in the number of trials to reach criterion on acquisition of the DNMS task. Once the task was learned, there were no significant differences across groups for memory tasks. Prenatal exposure to both alcohol and stress, but neither alone, was associated with hyperactivity and high levels of stereotypies during testing. Finally, pilot data using positron emission tomography methods for characterizing the dopamine system of these animals in vivo suggest increased dopamine D2 receptor availability in the striata in two alcohol + stressed monkeys compared with two matched controls; this finding will be further studied in larger population samples.

Neuroimaging Methods for Examining the Effects of Early Stress on Brain Structure and Function

Kathleen M. Thomas, Ph.D.

The advent of noninvasive methods of magnetic resonance imaging (MRI) has opened the door for developmental scientists to study the structure and function of the developing brain in vivo. In the past, developmental examinations of brain structure were limited to post mortem studies or invasive imaging methods requiring exposure to radiation, sedation, or other risks. Studies of brain function during childhood relied solely on neuropsychological and behavioral measures, identified predominantly by adult and animal lesion work. These techniques limit our ability to capture normal developmental changes in brain structure and function. In recent years, several research groups have begun to use structural and functional MRI techniques to examine brain development and cognitive function in both healthy children and pediatric patient populations. One of the most established uses of MRI with children has been in tracking the continued structural development of various brain regions with age. Structural MRI provides information at the systems level and does not address cellular level changes. However, careful quantitative measurements can be used to address global anatomic differences between patient and control groups. When combined with behavioral measures, quantitative morphometric measurements can provide additional information regarding the relationship between structure and function. For example, MRI measurements of basal ganglia volume have been combined with behavioral ratings of psychiatric symptom severity to demonstrate a positive correlation between obsessive compulsive symptomatology and size of the caudate nucleus and globus pallidus. An exciting area in pediatric imaging is the use of functional imaging paradigms with young children. Like structural imaging, functional MRI provides a systems level perspective of the brain regions activated during specific behavioral tasks and the network of regions that appear to be involved. Unlike lesion methods that address the effects of dysfunction in a localized region, functional MRI allows examination of the entire network of activation and potential compensatory activity in the face of dysfunction. Examinations of the correlations between behavior and activation provide researchers with an additional means of interpreting the functional significance of various regions of activity.

Our own work has examined the development of various cognitive skills such as attention and memory across normal development and in various patient populations. Of particular interest for this conference are our recent studies combining behavioral, structural MRI, and functional MRI measures of attention and memory in children with intraventricular hemorrhage (IVH). Germinal matrix hemorrhage, and subsequent IVH, is a relatively common complication of premature birth, occurring when immature blood vessels burst under stress and bleed into the lateral ventricles of the brain. This type of injury can be mild, with no obvious tissue damage, or quite severe, affecting the basal ganglia and surrounding white matter, as in some cases of cerebral palsy. We conducted structural whole brain volume MRI scans of 6- to 9-year-old children with IVH, including mild, moderate, and severe cases, but excluding children with motor dysfunction. Compared with full-term controls, children with IVH demonstrated 7 to 30 percent decrements in caudate volume despite controlling for total brain volume. No group differences were observed in volume of the putamen or globus pallidus. This finding is significant in that localized structural differences can be observed even 6 to 9 years after the original insult. In addition, the majority of these children do not demonstrate obvious developmental deficits. Rather, subtle problems have emerged in school learning and other cognitive domains. Functional MRI data were collected during performance of a go-nogo inhibition paradigm. Children were asked to press a response key for every letter except the letter "x." Responses were required on 75 percent of trials, resulting in a prepotent tendency to respond. Children with IVH had significantly more difficulty inhibiting a response than did control children, producing more false alarms. Functional MRI data demonstrated that approximately one-half of the IVH sample activated anterior cingulate and orbitofrontal cortices, regions identified in our studies of go-nogo performance in adults and healthy children. However, even those children who did activate these regions did not demonstrate the expected correlations between activity and behavioral performance. These data and those of other pediatric imaging studies suggest that both structural and functional MRI techniques can be used effectively with children and can be informative for researchers interested in the developmental outcome of children exposed to suboptimal environmental conditions. However, like all new methods, the limitations of MRI must be acknowledged, and the most effective combination of structural, functional, and behavioral techniques must be applied to develop strong paradigms for addressing brain development and function.

Neurobiological Correlates of Childhood Maltreatment

Martin H. Teicher, M.D., Ph.D., Carl M. Anderson, Ph.D., Susan L. Andersen, Ph.D., Ann Polcari, R.N., M.S.C.S., Ph.D.(c)

The focus of this presentation will be on neurobiological and neuropsychiatric abnormalities in individuals with a history of childhood abuse. These abnormalities will be compared with the reported effects of early stress on brain development of rodents and primates. To a great degree, the neuropsychiatric abnormalities observed in humans who have survived intense early stress dovetail with results from controlled studies in mammals and nonhuman primates. These observations underscore the importance and relevance of animal research to our efforts to delineate the genesis of mental disorders and to devise therapeutic or preventive strategies.

Research that we have conducted delineates a constellation of abnormalities that have been associated with childhood abuse. There are three major components to the constellation. One factor we have labeled "limbic irritability." It is manifested by markedly increased prevalence of symptoms suggestive of temporal lobe epilepsy (TLE) and by an increased incidence of clinically significant EEG abnormalities. The second component is deficient development and differentiation of the left hemisphere. This is manifest throughout the cerebral cortex and hippocampus. The third component is deficient left-right hemispheric integration, which is indicated by markedly lateralized shifts in hemispheric activity and by attenuated development of the middle portions of the corpus callosum, which is the primary pathway connecting the two hemispheres.

The initial sequelae of abuse may be either a constellation of internalizing symptoms, such as depression, anxiety, suicidal ideation, and posttraumatic stress, or an externalizing cluster that includes aggressive-impulsivity, delinquency, and substance abuse. Studies have shown that abused children may show some degree of neurological dysfunction even in the absence of apparent or reported head injury. Green et al. (1981) found that soft neurological signs and nonspecific EEG abnormalities were more common in physically abused children without known head trauma. Davies (1979) reported that in a sample of 22 patients involved as the child or younger member of an incestuous relationship, 77 percent had EEG abnormalities and 36 percent had clinical seizures. However, neither Green et al. (1981) nor Davies (1979) postulated that these abnormalities were a consequence of the abuse. Rather, they perceived the neurobiological abnormality as a preexisting vulnerability that increased risk of being victimized.

We have initiated a program of research endeavoring to test the hypothesis that early childhood maltreatment acting as a severe stressor produces a cascade of physiological and neurohumoral responses that lead to enduring deleterious effects on brain development. This is a difficult hypothesis to test in humans, as abuse is not necessarily a random act. If we observe an association between a history of abuse and presence of an abnormality, it may be the case that the abuse caused the abnormality. It can also be argued that the abnormality occurred first and increased the likelihood of abuse or that the abnormality runs in families and leads to an increased frequency of abusive behavior in family members or relatives. One approach to this problem is to conduct analogous studies of severe early stress in animals that are randomly assigned to treatments. The observation of parallel outcomes bolsters the hypothesis that this is a cause and effect relationship.

Association Between Early Abuse and Ratings of Limbic Irritability in Adulthood

To explore the potential relationship between early abuse and limbic system dysfunction, we devised the Limbic System Checklist-33 (LSCL-33; Teicher et al., 1993) to ascertain the frequency with which patients experience symptoms often encountered as ictal TLE phenomena (e.g., Spiers et al., 1985). These items consisted of paroxysmal somatic disturbances, brief hallucinatory events, visual phenomena, automatism, and dissociative experiences. Psychometric evaluation indicated that LSCL-33 scores had high test-retest reliability (r = .92; n = 16). Scores were low in normal controls (< 10) and elevated in patients with documented TLE (> 23).

We studied 253 adults presenting for outpatient psychiatric assessment (mean age, 34 yr; 58 percent female). The Life Experience Questionnaire (LEQ; Bryer et al., 1987) was used to ascertain abuse history. Subjects who reported no abuse (n = 109) had mean LSCL-33 scores of 13.6 (±11.3). Total LSCL-33 scores were 38 percent greater in patients with physical abuse, but not sexual abuse (p < .01; n = 77) and were 49 percent greater in patients with sexual abuse, but not physical abuse (p < .02; n = 26). Patients who acknowledged both physical and sexual abuse (n = 41) had scores 113 percent greater than patients denying abuse (p < .0001). LSCL-33 scores from both males and females were affected by abuse in the same manner.

Association Between Early Abuse and EEG Abnormalities in Childhood

A chart review was conducted to blindly examine the association between different types of abuse and quantifiable abnormalities on imaging, EEG, and neuropsychological testing (Ito et al., 1993). Medical records were reviewed on 115 consecutive admissions to a child and adolescent psychiatric hospital. Four groups were established based on abuse ratings. Subjects in the nonabused group had no evidence of abuse in any of the four categories (n = 27). Patients in the psychological abuse group had experienced psychological abuse or neglect but had not been physically or sexually abused (n = 22). Patients in the overall physical/sexual abuse group had experienced probable or definite physical or sexual abuse (n = 55). Patients in the severe physical/sexual abuse subgroup had a documented history of severe physical or sexual abuse (n = 38).

There were no differences between abused and nonabused patients in the prevalence of abnormal neurological exams. Abnormal imaging studies were found in 15 percent of nonabused patients and in 26 percent of abused patients. Abnormal EEG studies were found in 26.9 percent of the nonabused patients but in 54.4 percent of the patients with a history of early maltreatment (p = .021). Abnormal EEG studies were observed in 42.9 percent of the patients with psychological abuse or neglect, 59.6 percent of the total sample with physical/sexual abuse (p = .014), and 71.9 percent of the subsample with serious physical/sexual abuse (p = .0013). Maltreated and adequately treated patients differed most clearly in the prevalence of left-sided frontotemporal abnormalities (p = .036). They did not differ in the prevalence of either right-sided abnormalities (p > .8) or bilateral abnormalities (p > .5). Neuropsychological test results were reviewed for evidence of right-left hemispheric asymmetries (i.e., substantially better visual-spatial ability than verbal performance). Overall, in the adequately treated group, left hemisphere deficits were 2.25-fold more prevalent than right hemisphere deficits. In the total maltreatment group, left-sided deficits were 6.67-fold more prevalent than right, and left hemisphere deficits were 8-fold more prevalent than right-sided deficits in patients with a history of neglect or psychological abuse. Thus, maltreatment appears to be associated with an increased prevalence of left-sided EEG abnormalities and left-sided neuropsychological deficits.

EEG Coherence Measures of Brain Development

EEG coherence is a parameter that indicates the degree of synchrony ("shared activity" or interconnectivity) between two EEG leads across a portion of the bandwidth. Coherence is affected by the adequacy and degree of myelinization of long association fibers and by the degree and complexity of local axodendritic connections in the underlying cortex (Thatcher et al., 1987, 1992). Highly developed local cortical connections modify the EEG signal under the lead and decrease coherence between leads. In most instances, abnormally elevated levels of coherence are an indication of inadequate cortical development or maturation. Further, coherence decreases as EEG leads are moved farther apart. The rate of decay of coherence over distance is a direct index of the complexity and differentiation of local cortical connections (Thatcher et al., 1986). Hence, combined assessment of EEG coherence and coherence decay provides an objective measure of cortical maturation and differentiation.

Fifteen child or adolescent inpatients (10.7±2.5 yr, 7M:8F, 10 medicated) with a history of intense physical or sexual abuse, confirmed by DSS, were recruited (Ito et al., 1998). The controls were 15 healthy volunteers. All subjects were between 6 and 15 years of age, right handed with no history of neurological disorders or abnormal intelligence. Diagnostic data were derived from discharge diagnosis and structured clinical interviews.

To reduce the number of possible statistical comparisons, we derived composite measures of average left and average right hemisphere alpha EEG coherence based on all possible lead pairings within each hemisphere. Abused children had greater average left hemisphere coherence than normal children (p = .007), but a comparable degree of right hemisphere coherence (p > .7). In controls, the laterality index was -3.21 percent. In contrast, abused children had an average laterality index of +7.21 percent (p < .02), indicating significantly greater left vs. right coherence. There was no specific relationship between degree of asymmetry and diagnosis. The asymmetry was apparent whether their primary diagnosis was depression, PTSD, or conduct disorder. Abused subjects differed from controls in degree of asymmetry in central (p = .02), temporal (p < .03), and parietal regions (p < .055). Alpha EEG coherence decayed markedly over distance, and the rate of decay was well fit (r > .9) by a power function. In normal controls, coherence decayed at a more rapid rate in the left vs. right hemisphere. In contrast, patients with a history of severe abuse had a lower rate of decay in their left hemisphere than in their right hemisphere (p < .05). Overall, normal controls had a 16.3 percent greater rate of coherence decay in their left vs. right hemisphere, while abused subjects had a 6.8 percent lower rate of coherence decay in their left vs. right hemisphere (p < .04). Hence, these findings suggest that childhood abuse is associated with deficient maturation of the left hemisphere - both in relationship to normal controls and in relationship to their own right hemisphere.

Right-Left Evoked Response Asymmetry During Recall of Unpleasant Early Memories in Psychologically Traumatized Subjects

Schiffer, Teicher, and Papanicolau (1995) used probe auditory evoked potential (AEP) attenuation as a measure of hemispheric activity to study the effects of early trauma on cerebral laterality. We sought to assess whether early traumatic experience affected the degree of right-sided activation during recall of painful memories. To evaluate this hypothesis, hemispheric activity was measured in adult subjects under two conditions: (1) during recall of a neutral memory and (2) during recall of an unpleasant affectively laden early memory. Subjects were exposed to repeated auditory clicks, during measurement of the amplitude of the AEP. Theoretically, if one hemisphere is more actively involved in a competing mental activity than the other hemisphere, then AEPs recorded over the more distracted hemisphere should be weaker (Papanicolau and Johnstone, 1984; Papanicolau et al., 1983a,b).

Twelve unmedicated right-handed adults who had a history of emotional abuse were recruited and compared with 12 similar nonabused controls. AEPs were recorded while subjects recalled a neutral memory and responses to an abbreviated POMS scale were ascertained. A short empathic psychiatric interview was conducted to evoke recall of a disturbing childhood memory. The AEPs were then repeated. The unpleasant memory always followed the neutral memory task because of concern that unpleasant memories would interfere with the neutral task. Averaged AEP recordings were blindly read by an experienced researcher to obtain N1 and P2 peaks. Of the 20 subjects with artifact-free AEPs, 10 subjects (32.9 yr; M5/F5) had experienced significant childhood trauma and 10 had not (33.0 yr; M4/F6). There was no significant difference between the POMS scores of the two groups during the neutral memories. Both groups reported higher mean POMS scores following the unpleasant memory, but the scores were slightly greater in patients with a history of emotional abuse.

The trauma group displayed significant left dominant asymmetry during the neutral memory (asymmetry index = -15.9 percent; p < .02), and right dominance during the unpleasant memory (asymmetry index = +12.2 percent; p < .10). Overall, trauma subjects had a highly significant asymmetry shift between conditions (p = .007). In contrast, the control group had no significant asymmetry or shift during either state. ANCOVA was used to control for differences in degree of emotional response (all combinations). Statistical differences between groups in degree of asymmetry shift were not a consequence of differences in magnitude of expressed emotional response.

Adult patients with a history of significant early stress had more lateralized shifts in cortical activity as a consequence of their affective state or nature of recalled memory. While they were thinking about a neutral memory, their left auditory cortex was more preoccupied and less available to process the sounds. In contrast, the right auditory cortex was more preoccupied and otherwise engaged when they were thinking about a disturbing childhood memory. In healthy controls, both hemispheres were equally involved in processing and responding to neutral and disturbing memories. These findings are consonant with observations of Rauch et al. (1996), who found, using PET scans, that individuals with PTSD (including several with childhood trauma) showed specific activation of their limbic system in the right but not left hemisphere during exposure to traumatic reminders.

Association between childhood abuse and regional anatomy of the corpus callosum in children. As our research suggested that early abuse is associated with diminished left hemisphere cortical development and decreased right-left cortical integration, we sought to ascertain whether there were abnormalities in the regional anatomy of the corpus callosum in children with a history of abuse or neglect. Medical records were reviewed on 115 consecutive pediatric patients admitted to McLean Hospital. Patients were eliminated for possible preexisting neurological abnormalities, including loss of consciousness and perinatal complications. MRI records were available on 51 subjects from this group. Mean age was 12.9 ± 2.9 years, and 52 percent were male. Records were blindly reviewed by two independent raters using all clinical information and DSS investigative reports to ascertain whether the children had a history of physical abuse, sexual abuse, psychological abuse (witnessing domestic violence, verbal abuse), or neglect.

An automated algorithm was used to divide the corpus callosum into seven regions as defined by Witelson et al. (1989). In order to control for differences in corpus callosum size due to age or gender, regional volume was corrected for total brain volume. MRI measures were performed by independent researchers blind to all clinical variables. In addition, MRI images from 97 carefully screened healthy normal control children were obtained from Dr. J. Giedd at the Child Psychiatry Branch of the National Institute of Mental Health (NIMH). Mean age was 11.5 ± 3.5 years, and 63 percent were male. Data were analyzed separately for males and females, as previous research has shown that early experience exerts sexually dimorphic effects on the development of the corpus callosum (Denenberg et al., 1983).

Relative regional corpus callosum size was assessed in hospitalized boys with history of abuse or neglect, hospitalized boys without any history of abuse or neglect but with serious psychiatric illness (contrast group), and normal healthy boys studied at the NIMH. Overall, there were prominent group differences in the rostral body and anterior and posterior midbody of the corpus callosum (all p < .006). Within these regions, there were no significant differences between the contrast group and the health controls. However, in the abused group these regions (and also the isthmus) were between 23 percent and 31 percent smaller than in the healthy normal controls (all p < .004). Hence, abuse or neglect, but not psychiatric illness, was associated with a marked and significant reduction in the size of the middle portion of the corpus callosum.

Stepwise regression analyses were performed to ascertain the factors most directly associated with the changes in relative regional corpus callosum size. Results from this analysis indicate that neglect was associated with a marked (24 - 42 percent) reduction in the relative size of all regions of the corpus callosum. In contrast, physical abuse was associated with a significant reduction only in the anterior midbody, and sexual abuse was associated with a reduction in the splenium. Hence, this analysis suggests that neglect may produce particularly severe impact on the relative regional size of the corpus callosum of boys.

Surprisingly, stepwise multiple regression of the female population revealed a different pattern or association. While neglect was consistently associated with diminished corpus callosum size in boys, sexual abuse was a more powerful factor in girls. Indeed, a history of sexual abuse was associated with an 18 percent to 30 percent reduction in size of the rostral body, anterior and posterior midbody, and isthmus. This again points out that the middle portions of the corpus callosum may be most vulnerable to the effects of early experience but that girls and boys may differ in their windows of vulnerability. Neglect in girls was associated with an increase in the size of region 6 and a decrease in the size of region 7. This most likely reflects an alteration in the shape of the corpus callosum with a bulging of the isthmus at the expense of the splenium.

Sánchez et al. (1998) found that differential rearing experience affected the development of the corpus callosum and the cognitive function of male rhesus monkeys. Briefly, infant monkeys raised individually in a nursery from 2 to 12 months were compared to age-matched infants raised in a seminaturalistic social environment. Although overall brain volumes did not differ, the corpus callosum was significantly decreased in the nursery group. Rearing differences were not found in the hippocampus, cerebellum, or anterior commissure. Cognitive difficulties emerged in delayed non-matching to sample and object reversal learning, and degree of impairment correlated with alterations in corpus callosum size.

Berrebi et al. (1988) showed that male and female rats differed in the effects of early handling (a form of brief beneficial stimulation) on the development of their corpus callosum. In adulthood (110 days), handled male rats had significantly greater width of their corpus callosum than non-handled male controls. The opposite results were observed in females. Juraska and Kopcik (1988) found that rearing in a complex environment (post weaning) enhanced the size of the corpus callosum of both male and female rats, though the magnitude of the effect varied by gender. Moreover, ultrastructural studies showed that the effects of early experience were mediated through different mechanisms. Rearing in the complex environment produced a significant increase in the number of myelinated axons in the corpus callosum of female but not male rats. In contrast, rearing in the complex environment markedly increased the diameter of myelinated axons in the corpus callosum of male rats but had no effect on myelinated axon diameter in females. These findings lend credence to the hypothesis that early experience can affect the development of the human corpus callosum and that males and females may differ to a significant degree in the manner in which the corpus callosum is affected.

Association Between Limbic Dysfunction and fMRI Measures of Cerebellar Vermal Blood Flow

Childhood maltreatment has been associated with dissociation, increased prevalence of abnormal EEGs (Ito, 1998), and development of symptoms suggestive of temporal lobe epilepsy or limbic irritability (Teicher et al., 1993). The cerebellum, like the cortex, corpus callosum, and hippocampus, has a protracted postnatal ontogeny and is markedly affected by early exposure to corticosteroids (Lauder, 1983). The cerebellar vermis appears to play a role in the control of epilepsy or limbic activation (Heath, 1976; Strain et al., 1979; Cooper et al., 1974, 1985; Riklan et al., 1976). Research by Mason and Harlow (Mason, 1975) has shown that vestibular stimulation during early life (which largely acts on the cerebellum) markedly attenuates the adverse effects of rearing without maternal contact. fMRI was used to assess the relationship between symptoms of limbic irritability and blood volume in the cerebellar vermis of young adults with a history of sexual abuse or intense verbal abuse.

T2 relaxometry was used as a novel fMRI procedure to derive steady-state blood flow measures (Teicher et al., in press). Although conventional blood oxygenation level dependent (BOLD) fMRI is a valuable technique for observing dynamic brain activity changes between baseline and active conditions, thus far it has failed to provide insight into possible resting or steady-state differences in regional perfusion between groups (Ogawa, 1998). T2 relaxometry, like BOLD, hinges on the paramagnetic properties of deoxyhemoglobin (Pauling, 1936). However, the mismatch between blood flow and oxygen extraction that occurs as an acute reaction to enhanced neuronal activity in BOLD does not persist under steady-state conditions. Instead, regional blood flow is regulated to appropriately match perfusion with ongoing metabolic demand (Kety, 1960), and deoxyhemoglobin concentration becomes constant between regions in the steady state. Therefore, regions with greater continuous activity would be perfused at a greater rate, and these regions would receive, over time, a greater volume of blood and a greater number of deoxyhemoglobin molecules per volume of tissue (van der Kolk, 1991). Thus, there should be an augmentation in the paramagnetic properties of the region dependent on the amount (but not the concentration) of deoxyhemoglobin that would be detectable as a diminished T2 relaxation time.

Thirty-two young adults (9M/32F, 18-22 yr) participated, including 15 (3M/12F) with a history of sexual (n = 12) or verbal (n = 3) childhood trauma exclusive of physical trauma. Each subject underwent echo-planar fMRI to assess basal blood perfusion in the cerebellar vermis, cerebellar hemispheres, anterior temporal lobe, and entire left and right cerebral hemispheres. This was accomplished by collecting a series of 32 "TE stepped" echo-planar images (EPI) in 10 axial slices under resting conditions. A regional decay curve was generated from median pixel intensity within the ROI at each value of TE examined to calculate true T2 relaxation time (T2-RT) with high precision and reproducibility.

There was a strong association between T2-RT and measures of limbic irritability in both groups. The correlation was -0.807 (p < .003) in abused subjects and -0.677 (p < .004) in controls. Slopes were parallel, but the regression line was much greater in the abused subjects (p < .0001). Average T2-RT in abused patients was 100.1 ms vs. 91.8 in controls, even after data were controlled for differences in LSCL-33 scores. Elevated T2-RT measures are associated with decreased blood volume and neuronal activity. These findings suggest that early abuse is associated with a functional deficit in the development of the cerebellar vermis. Blood volume in the vermis increased with greater LSCL-33 scores, possibly representing an effort by the vermis to modulate and contain the irritability. Although this is not a brain region we normally think of as playing an important role in psychiatric symptomatology, there is an enormous convergence of new data suggesting that abnormalities in the cerebellar vermis may be involved in a wide array of psychiatric disorders including bipolar and unipolar depression (Fischler et al., 1996; Lauterbach, 1996; Loeber et al., 1999; Beauregard et al., 1998), schizophrenia (Loeber et al., 1999; Jacobsen et al., 1997), autism (Courchesne et al., 1991), and ADHD (Berquin et al., 1998; Mostofsky et al., 1998). The cerebellar vermis exerts strong modulatory effects on the locus coeruleus, ventral tegmental area, and substantia nigra, which are cell body regions for projection of the primary dopamine and norepinephrine pathways (Reis and Golanov, 1997; Snider and Maiti, 1976; Snider et al., 1976).

In children with ADHD, we found a strong dose-dependent effect of methylphenidate on T2-RT in the cerebellar vermis (Anderson et al., 199x). Progressively higher doses of methylphenidate increase T2-RT. Presumably this occurs through direct effects of methylphenidate on dopamine transporters in the basal ganglia, and the net enhancement in dopamine neurotransmission produced by the drug reduced the need for vermal activation of the substantia nigra and ventral tegmental area. Dose-dependent increases in Fos-like immunoreactivity with acute administration of d-amphetamine and cocaine have been observed in rat cerebellar vermis (Klitenick, 1995), suggesting that abuse-related abnormalities in vermal development may play some role in the proclivity of individuals with early abuse to develop drug abuse.

Archive Home | Accessibility | Privacy | FOIA (NIH) | Current NIDA Home Page
National Institutes of Health logo_Department of Health and Human Services Logo The National Institute on Drug Abuse (NIDA) is part of the National Institutes of Health (NIH) , a component of the U.S. Department of Health and Human Services. Questions? See our Contact Information. . The U.S. government's official web portal