[Page 102 continued]
Until recently, developmental psychology and neurobiology have been studied in parallel. Whereas the former is concerned with the observation and measurement of behaviour, cognition, and emotion, the latter is concerned with the study of cellular, neurophysiological, and biochemical processes in the brain and the autonomic nervous system. It is now increasingly possible to study simultaneously the neurobiological processes accompanying or underlying observed behaviour, by the use of a variety of neurophysiological measures and brain-imaging techniques [*2] (for the latter, see review by Nelson & Bloom, 1997).
These methods are enabling the apparent mind-brain dichotomy to be bridged. However, it is evident that the relationship between observed behaviour and measured physiological function is complex. Before considering the neurobiological correlates of child abuse and neglect in the following section, in this section a variety of neurobiological equivalents of observed behaviours, emotions, and psychological processes will be discussed, which are of relevance to the field of child abuse and neglect. While some begin from a neurobiological point of view, others commence from a psychological perspective.
There are more than 30 neurotransmitters [*3] in the central nervous system.
Of these, the biogenic amines noradrenaline (norepinephrine) and dopamine (known collectively as the catecholamines) and serotonin (or 5-HT) have been studied in greater detail, since they are particularly involved in the regulation of several human behavioural systems which, in turn, help to regulate the interaction between the organism and its environment.
The biogenic amines are synthesised in discrete nuclei in the brain stem and midbrain.
Projecting axons from these nuclei extend to all higher brain centres, and it is likely that the biogenic amines are involved in regulating the function of higher brain centres.
There are different ways by which neurotransmitter secretions become activated. It is, for example, postulated that peripheral adrenaline, produced as part of the stress response, stimulates receptors in the vagus nerve (part of the autonomic nervous system), which ends in the nucleus of the solitary tract in the brain stem. From there, messages are sent to the noradrenaline-producing locus coeruleus (LeDoux, 1996). The locus coeruleus is primarily activated from within the brain, from nuclei in the medulla (Aston-Jones et al., 1991).
The biogenic amine systems are hypothesised to operate in balance with each other in their regulatory functions (Rogeness, Javors, & Pliszka, 1992). There has been further amplification of Gray's work on behavioural facilitatory and behavioural inhibitory systems, which suggested that these behavioural systems are mediated by neurotransmitters (Gray, 1982, 1987).
Rogeness and McClure (1996) have reviewed their own and others' empirical findings about the functions of these neurotransmitters in regulating behaviour and emotions and about possible early environmental influences on permanent changes in neurotransmitter levels. Dopamine is postulated to mediate the behavioural facilitatory system, activated by rewarding stimuli, or by aversive stimuli when escape or avoidance are possible.
Dopamine is thus involved in approach, escape, and active avoidance as well as in predatory aggression (Quay, 1993). It is suggested (Rogeness et al., 1992) that noradrenaline and serotonin are involved in the mediation of behavioural inhibition in the face of lack of reward, punishment, or uncertainty, These two neurotransmitters are therefore involved in regulating dopamine dependent behaviour.
Attachment is considered here since it is a fundamental aspect of child development and is affected by child abuse and neglect. There are some neurobiological correlates of attachment security, and it has also been shown to influence the infant's response to stress.
Attachment behaviour is defined as proximity-seeking behaviour by a dependent organism (infant or child), when he or she senses discomfort of any sort, including pain, fear, cold, or hunger. The child seeks to get closer to the attachment figure (parent or primary carer) on the assumption that the parent will be able to reduce the discomfort and restore the child's equanimity. It is a biological instinct (Bowlby, 1969).
On the basis of the nature of the mother's or the primary caregiver(s)' responses to the infant's inevitable, normal, and repeated bids for a response to their attachment needs, the child constructs internal working models of self and parent (Bowlby, 1988). These models are beliefs by the child about herself or himself and predictions about how he or she will be treated by others.
There are well-validated measures of the nature of infants' and young children's attachment status, which begins to be formed in the middle of the first year of life. The child's attachment status is assumed to be based on the child's previous attachment experiences and thus reflects the child's internal working models. Attachment security is measured in infancy and early childhood by the Strange Situation Test (Ainsworth, Blehar, Waters, & Wall, 1978), which yields a secure category (B),
two insecure categories -- anxious/avoidant (A) and anxious/insecure (C) -- and a disorganised/disoriented category (D) (Main & Solomon, 1990).
Following the addition of the D category, V. Carlson, Cicchetti, Barnett, and Braunwald (1989b) re-analysed the attachment relationships of a sample of maltreated and non-maltreated 12-month-old infants and found a large over-representation of this disorganised category (82 %) among the maltreated group, in comparison with 19% in the non-maltreated group. More boys than girls were classified as (D).
Main and Hesse (1990) postulated that introjection of fear into the care-giving relationship leads to the (D) category. The frightened or discomforted infant seeks his attachment figure, who is at the same time the source of his discomfort.
Alongside the (D) classification, Crittenden (1988) has added the (A/C) classification, also found more commonly among maltreated children. Most studies of attachment status in maltreated children have included physically abused and/or neglected children, and none have examined the attachment classification of sexually abused infants (Morton & Browne, 1998). This may be explained by the fact that in neglect, emotional abuse, and repeated physical abuse, especially of young children, the abuser and caregiver will, most likely, be the same person, whereas it is far more likely that the sexual abuser will not be the primary or sole caregiver of the child.
The measurement of attachment status in older children using the Strange Situation Test is not applicable. Alternative means, relying on the assessment of internal working models, have been developed using story-stems (McCrone, Egeland, Kalkoske, & Carlson, 1994), a projective task in which children are requested to complete beginnings of a number of stories.
Neurobiological correlates of attachment
Gunnar, Brodersen, Nachmias, Buss, and Rigatuso (1996) postulate that one function of a secure attachment relationship is to buffer or protect the developing brain from the potential deleterious effects of elevated gluco-corticoids on the brain during the protracted postnatal brain development. This is the more so for infants rendered vulnerable due to a fearful or inhibited temperament. An interesting psychobiological attachment theory has been expounded by Kraemer (1992), suggesting a central role for biogenic amines as the mediators of secure or insecure attachment. There is also evidence of a complex relationship between attachment status and infant cardiac reactivity (Izard et al., 1991). Heart-rate variability (a measure of heart-rate pattern) was found to be higher in insecure infants.
Stress is defined as a stimulus or experience that produces a negative emotional reaction or affect, including fear and a sense of loss of control. Potent sources of stress in childhood have now been shown to include severe deprivation and neglect in early life and exposure to violence between parents, as well as the more obvious recognised forms of abuse.
Much of the data on the effects of early deprivation on brain development comes from animal experiments. For example, extensive animal studies have shown that brief and repeated periods of separating a mother from her newly born offspring leads to a stress reaction expressed by increased gluco-corticoid secretion with resultant death of hippocampal cells (e.g. Plotsky & Meaney, 1993). Conversely, increased licking behaviour of their pups by mother rats leads to decreased hippocampal cell loss in old age.
Interestingly, this finding was initially attributed to brief periods of handling of neonatal rats by humans (Meaney, Aitken, Bhatnagar, Van Berkel, & Sapolsky, 1988); it was subsequently found that the mother rat responded to the human smell on her pups by licking them (Liu et al., 1997)! Further work with rats has shown that one day of maternal deprivation was sufficient to decrease brain-derived neurotrophic factor in the hippocampus and bring forward preprogrammed cell death (apoptosis) (Zhang, Xing, Levine, Post, & Smith, 1997). It is important to note that 1 day in the life of a new-born rat is equivalent approximately to 6 months of maternal deprivation in human infants.
The Stress Response
The stress response is a physiological coping response. It involves the hypothalamic-pituitary-adrenal axis, the autonomous [*] nervous system, the neurotransmitter system, and the immune system.
There are individual variations, likely to be enduring, in the response to stress, which are based on differences in temperament (W. Boyce, Barr, & Zeltzer, 1992) as well as on prior experience.
Prior experience can affect responses to stress by sensitisation, by determining the child's attachment security (Nachmias~Guririar, Mangelsdorf, Parritz, & Buss, 1996), and by shaping the child's perception of an experience and its meaning as stressful or not.
Lewis (1992) describes three dimensions along which responsiveness to stress can be measured, namely threshold, dampening, and reactivation.
In support of temperamental influences, Lewis showed that in babies there was stability over the first 2 months of life for threshold and dampening levels, particularly in highly reactive babies whose response threshold to stress was low and whose dampening response was low.
Although in Lewis' work, measures of stress responses were behavioural rather than neurobiological, it is likely that there are neurobiological correlates to these observed behaviours, which would suggest that some children will be more vulnerable to the effects of stress.
These children would also be described and perceived by their carers as temperamentally difficult. They are therefore doubly vulnerable, first to their own inherent responses to stress, and second, when they are met with insensitive and punitive care-giving responses, which will be perceived by the infant or child as stressful.
Hypothalamic-pituitary-adrenal (HPA) axis
The (HPA) axis is a physiological pathway connecting the brain to the adrenal cortex, which secretes cortisol [*4].
The hypothalamus secretes corticotropin-releasing hormone (CRH) that in turn stimulates the anterior pituitary gland
to secrete adrenocorticotropic hormone (ACTH). This is released into the blood and, when reaching the adrenal gland, stimulates the cortex to produce and release cortisol into the circulation (Chrousos & Gold, 1992). The rate of secretion of cortisol is regulated by a negative feedback loop to maintain an optimal level. This means that as specific brain centres receive cortisol, they send messages via the HPA axis to reduce the level of cortisol secretion. Conversely, there are situations that require an elevation of cortisol in the body, and when certain signals are perceived by the brain, cortisol secretion increases via messages down the HPA axis. In the normal state, there is a diurnal variation in cortisol levels, with higher levels found in the morning, falling to lower levels in the afternoon. Cortisol levels can be measured noninvasively by assays of saliva and in the urine.
One aspect of the body's coping response to acute stress is an elevation of serum cortisol, a stress response that commences in early infancy. It has, for instance, been found that a more reactive HPA axis is associated with greater professional competence in air traffic controllers (Rose, Jenkins, Hurst, Livingston, & Hall, 1982) and with greater emotional and social competence in boys with haemophilia (Mattson, Gross, & Hall, 1971 ).
Serum cortisol acts in a number of different ways and on most tissues and organs. Its actions include suppressing the immune response, increasing the level of circulating glucose, and dampening of fear responses to the stressor, as well as adverse effects on the hippocampus (see later).
However, what is particularly interesting and important is the notion, proposed by Munck, Guyre, and Holbrook (1984), that the stress-induced increase in glucocorticoid (cortisol) levels protects not against the original stressor but, instead, against the body's normal and immediate responses to stress, with the aim of avoiding the over- reaction of these responses, to the detriment of the body's homeostasis.
Behavioural responses to stress
Cortisol level is only one measure of a stress response, behaviour being a further one. Interestingly, there is a differential response to the same stressor from the two measures.
Gunnar , Brodersen, and Krueger (1996) showed a diverging stress response in infancy. Whereas there was a parallel decrease of both cortisol and behavioural responses to stress between 2 and 6 months, crying increased again by 15 months, while the cortisol response to stress had fallen to a lower level than in early infancy.
Another example of this differential response is the finding that pacifying a young infant by, for instance, use of a dummy, which reduces crying, belies the fact that elevation of cortisol continues (Gunnar, 1992). Coping may take the form of a reduction in felt anxiety or distress, alongside (and possibly facilitated by, or at the expense of) elevated cortisol.
The sympathetic nervous system and catecholamines
Both the sympathetic nervous system and the catecholamines are activated by stress. Fear messages from the amygdala and hippocampus arriving in the medulla of the brain stem (LeDoux, 1996), and CRH from the hypothalamus, released by stress, stimulate the locus coeruleus and thus noradrenaline secretion in the brain, demonstrating the direct links between the adrenergic and the cortisol responses to stress (Southwick et al., 1993).
Messages are also relayed by sympathetic nerves from the brain stem to the medulla of the adrenal gland, which secretes adrenaline and noradrenaline. The effects of these hormones include raising heart rate and blood pressure, sweating, and activation of the fight or flight response.
As mentioned earlier, adrenaline probably also activates noradrenaline secretion in the brain via the vagus nerve. There are thus several connections between the peripheral sympathetic adrenaline and central (brain) noradrenaline actions in response to stress (Krystql, Southwick, & Charney, 1995), which explain the emotional as well as the physical experiences associated with stress.
Whereas the catechol amines epinephrine (adrenaline) and norepinephrine (noradrenaline), whose production marks the sympathetic nervous system's response to stress, are secreted within seconds of the body sensing a stressor, glucocorticoids are secreted during the following minutes and take several hours to exert their effect (M. Howe, 1998).
In pre-clinical studies, stress has been shown to enhance the release and metabolism of dopamine in the prefrontal cortex (Charney , Deutch, Krystal, Southwick, & Davis, 1993), one of whose functions is to produce coping responses to stress. Raised levels of noradrenaline and dopamine are positively associated with dysfunction of the prefrontal cortex (Arnsten, 1999), whose functions also include the planning and organising of actions using "working memory" and the inhibiting of inappropriate responses and to attention to distractions ("executive functions"). This dysfunction leads to symptoms clinically recognised as Attention Deficit Hyperactivity Disorder (ADHD).
Long-term effects of the stress response
Very young animals and human infants show reactions to stress that, in some circumstances, may become enduring and, in others, can be modulated by maternal behaviour. In monkeys, infant-mother separation activates the HPA axis, leading to elevation of ACTH and plasma cortisol as well as to increased activity of the noradrenergic sympathetic nervous system, which includes elevation of heart rate. In some monkeys, this reaction may become sustained and this physiological arousal is accompanied by passive and withdrawn behaviour (Suomi, 1991).
McEwen (1999) describes several more and less adaptive responses of the stress hormone axis to chronic stress. When faced with repeated or chronic stress, suppression of the stress response leads to a restoration of cortisol levels to within normal limits (Yehuda, OilIer, Southwick, Lowy, & Mason, 1991) by down-regulating the HPA axis response. However, since cortisol also exerts an effect on the amygdala, which is concerned with actively responding to fear-inducing stimuli in times of acute threat, there may be a cost to the reduced cortisol levels, namely a dysfunctional, or less than optimal, response to frightening experiences, and feelings of passive fear (Hart, Gunnar, & Cicchetti, 1995). These are not uncommonly met with in children who have suffered long-term abuse (Shields, Cicchetti, & Ryan, 1994) and may be reflected the (D) category of infant attachment.
Stress, Elevated Cortisol, the Hippocampus, and Memory
While being a necessary physiological response to acute stress, elevated cortisol may also be harmful. Direct evidence for the harmful effects on brain development of a reactive HPA axis, and consequent elevated cortisol levels, particularly in early life, first came from animal , experiments.
An indication of a similar mechanism in ii humans comes from Gunnar and Nelson (1994) who showed that in 12-month-old infants, there was a negative correlation between EEG event-related potentials
(ERPs), which reflect hippocampal activity in "laying down a memory", and salivary cortisol levels. In other words, it appeared that higher cortisol levels interfered with activity of the hippocampus.
In his paper entitled Why stress is bad for your brain, Sapolsky (1996) has succinctly summarised recent evidence indicating a significant correlation between sustained stress, excess cortisol, and damage to the hippocampus in humans.
The hippocampus, part of the temporal lobe of the brain, is integrally concerned with memory (Squire, 1992). At the time of recollection, the hippocampus is believed to integrate the different aspects of a memory, as well as to locate the memory in time, place, and context (Bremner & Narayan, 1998).
Whereas the left hippocampus is believed to playa more important role in verbal memory, the right side is more involved with visual memory (Bremner et al., 1995). The hippocampus has a high concentration of receptors for glucocorticoids. Primate hippocampal neurons are adversely affected by sustained high levels of cortisol, which promotes early degeneration of these neurons (Sapolsky, Uno, Rebert, & Finch, 1990).
Exposure to high levels of cortisol causes atrophy of hippocampal dendrites, which is reversible when exposure is brief. Prolonged high levels of cortisol lead to hippocampal cell death, probably due to increased neuronal vulnerability to glutamate toxicity. Long-term elevated, but not toxic, cortisol levels render hippocampal neurons susceptible to the effects of commonly encountered threats to the brain, namely hypoxia, epileptic seizures, hypoglycaemia, physical trauma, and toxic chemicals.
High-dose cortisol medication has been shown to affect memory adversely, both in adults and in children. Bender, Lerner, and Poland (1991) showed that the verbal memory of asthmatic children on high doses of prednisolone was poorer than on low doses. Healthy adults have shown a decline in explicit memory with sustained increases in dexamethasone (a glucocorticoid) (New- comer, Craft, Hershey, Askins, & Bardgett, 1994).
There are other brain regions with glucocorticoid receptors, including the cingulate gyrus, amygdala, and frontal brain regions. It is possible that some of these regions (although not the amygdala) may be adversely affected by excess cortisol levels early in life (Gunnar, 1998).
Buffers to the Stress Response and the Influence of Attachment Status
Gunnar (1998) suggests that in view of the potentially damaging effects of elevated cortisol levels on the brain early in development, mechanisms have evolved to lower reactivity of the HPA axis to stress. From animal (rodent) work, we learn that hypo-responsiveness to stress in the very early days of postnatal life is achieved by maintaining close contact between mother and pup (Suchecki, Rosenfeld, & Levine, 1993).
Gunnar, Brodersen, and Krueger (1996) have shown that in 72 infants, the cortisol response to the same procedure (well-baby examination with inoculation) decreased from a high to a lower response between the age of 2 and 4 months, and further between 6 and 15 months. This change possibly reflected the child's capacity to recall the previous experience, a repetition of which may therefore not be as aversive. The salivary cortisol levels in 9-month-old infants, left with an unfamiliar babysitter for 30 minutes, did not rise if the caregiver was friendly, playful, and sensitive, in contrast to a cold and distant babysitter (Gunnar, Larson, Hertsgaard, & Brodersen, 1992).
It is postulated (Gunnar, 1998) that the equivalent protector, or buffer, of the HPA axis in human infants is the security of attachment with the primary caregiver . Thus, for example, Nachmias et al. (1996) have shown that 18-month-old children who had a secure attachment to their mother, who was present, showed no elevation of cortisol when responding fearfully to the approach of a stranger (a clown). This finding held whether the children were rated as constitutionally inhibited in new social situations, or not.
By contrast, constitutionally inhibited and insecurely attached children showed a significant elevation in salivary cortisol when approached by the clown. It was also shown that for these latter infants, maternal intrusiveness and insensitive encouragement of the infant towards the clown contributed to the elevated cortisol response.
Infants who showed a disorganised/disoriented attachment response were found to have higher cortisol levels during the Strange Situation Test (Hertsgaard, Gunnar, Erickson, & Nachmias, 1995). This test is a mild stressor compared to most experiences of child abuse and neglect and these results reflect the extent of these children's vulnerability to stress. The disorganised attachment pattern is associated with abuse and neglect (V. Carlson, Cicchetti, Barnett, & Braunwald, 1989a).
Fox, Calkins, and Bell (1994) have shown particular EEG asymmetries between the electrical activity of the right and left frontal lobes in 4-month-old infants who had shown early difficulties in coping with the arousal induced by new stimuli. In some of these children, by 24 months, there had been a behavioural adaptation but the original EEG asymmetry endured.
Whether these enduring physiological changes are indicators of continuing vulnerability in these children, despite the behavioural adaptation which they showed, is not known. Davidson (1994) has suggested that a particular form positive emotion, associated with the motivation towards attaining a goal (e.g. enthusiasm), is manifested in the brain as activity in the left prefrontal cortex.
In adults, enduring changes have been found in the pattern of EEG asymmetry between the left and right prefrontal lobes in patients who had suffered from depression but had made a symptomatic recovery (Henriques & Davidson, 1990). Similar patterns of asymmetries have been found in 3-year-old children who were temperamentally extremely inhibited and shy (Davidson, 1994J:
Psychological Therapies and Neurobiological Changes
Alongside evidence of specific brain functioning and structure, which are the physiological concomitants of observed behaviour, there is evidence that behavioural therapies, as well as drug therapy (e.g. for obsessive compulsive disorder) bring about change on PET scanning (Baxter et al., 1992).
This has important implications for intervention, underscoring the contribution of non-physical, psychological treatments to neurobiological change.