Abstract
Depression arises as the outcome or final common pathway of multiple interacting risk factors coming together at a particular time, or over a particular period, in a person's life. Kendler et al. state that ‘Major depression is a multifactorial disorder and understanding its aetiology will require the rigorous integration of genetic, temperamental and environmental risk factors’ [1], p.1139]. This paper proposes that among the environmental risk factors requiring consideration are adverse relational experiences dating from early life.
Vulnerability to depression and overdrive of the hypophysial-pituitary-adrenal axis
Recent adverse life events have long been related to the development of depression [2–4] and were found to be the most powerful predictors of depressive onset in a major prospective study [1]. Clinical experience reveals that individual vulnerability to the development of depression after adverse life events varies greatly. Genetic factors and kindling, or sensitization, of depression pathways via previous episodes of depression and/ or via the effects of stress, are regarded as major factors in determining individual vulnerability to depression [5–7]. Stress causing hypophysial-pituitary-adrenal (HPA) activation during critical periods of early development is currently considered to be of particular importance in sensitizing the depression pathways in the brain [5, 8–12].
There is good evidence from animal models that adults who have been subjected to psychological stress as pups demonstrate exaggerated HPA responses to stress [8–12].
The psychological stresses examined have included maternal separation and non-handling. The non-handled animals demonstrated behaviours consistent with animal models of depression including reduced exploration of new environments [8].
Coplan et al. [13] studied bonnet macaques reared by mothers who were exposed to a condition involving the variable availability of food. These infant monkeys demonstrated persistently elevated cerebrospinal fluid (CSF) concentrations of corticotrophin releasing factor (CRF), which persisted after the mothers had reliable access to food. This animal model could be regarded as one approximating the situation where human babies are exposed to a situation involving the inconsistent or unreliable presence and/or emotional responsiveness of the mother. In the human infant, as will be delineated shortly, this situation is likely to lead to insecurity of attachment, and exaggerated HPA responses to stressful conditions.
Psychological stress in human infants
There is a burgeoning literature on the effects of stress on human infants, which focuses primarily on HPA activation in response to stress and on the effects of stressful experience on brain development. Importantly, there has been a significant increase in understanding the neurobiology of brain development, demonstrating the neuroplasticity of the brain according to experience. The general wiring of the human brain is, of course, genetically determined, but research, particularly over the last 10 years, has shown that environmental experiences impact brain development from birth, and before, if we include the intrauterine environment [6, 10]. Post [14] delineates the ‘sculpting’ of the brain according to experience, and postulates that emotional over-excitation or deprivation in early life can affect the neural substrates of emotional and cognitive development just as over-excitation or deprivation can significantly affect the development of the visual system.
Kandel [15] and Shatz [16, 17] give excellent accounts of the processes involved in neuroplastisticity. Kandel states [15], p.460] that behaviour and social factors can ‘exert actions on the brain by feeding back upon it to modify the expression of genes and thus the function of nerve cells’. He goes on to say that ‘alterations in gene expression induced through learning give rise to changes in patterns of neuronal connections’. These changes in neuronal connection create the alterations in brain structure and functioning that may arise through experience.
The first two to three years of life, during which patterns of attachment to primary caregivers are forming, is considered to be a critical period for the experiencedependent maturation of the brain. The literature suggests that the right brain is dominant in this early period of life [18] and that attachment behaviours are predominantly linked to right hemisphere functions [19–21]. Importantly, the right hemisphere is regarded as playing a predominant role in emotional processing [20–22], and also with neuroendocrine and autonomic activation and the stress response in humans [20].
At the current state of knowledge, it seems that right brain hemisphere functions are predominantly responsible for the perception of emotion, facial expression of affect and non-verbal aspects of language, that is for those aspects of human functioning that enable us to ‘read’ the other person and link us in secure attachments [23, 24]. Fonagy's [25] delineation of reflective function as crucial to the development of secure attachment can be regarded as dependent on the integration of right and left hemisphere functioning. That is to say, reflective function, the ability to conceptualize the mind of another, will be dependent on the social–emotional processing in the right hemisphere, integrated with the linguistic and interpretive capacities of the left hemisphere [23, 24].
The human infant is biologically adapted for participating in human interactions. A basic motive is ‘social fittedness’ [26], and emotional signalling between infant and caretaker provides the basis for communication and regulation in the relationship. The bond involving regulation of closeness for safety and security between children and their caretakers forms the attachment behavioural system. It is now thought that the attachment system does not simply promote proximity to caregivers for protection from predators, as was first thought [27] but is integral to the organization of normal neurophysiological homeostasis [23, 28, 29]. Indeed, Kraemer [29] regards the attachment system as the crucial organizing regulator of normal neurophysiological homeostasis.
Attachment patterns and infant experience
The principal means of delineating attachment patterns in human infants is via Ainsworth's Strange Situation (SS) [30], an experimental condition involving short periods of separation (and reunion) from the primary caregiver, usually mother. Secure and insecure (subtypes ambivalent, avoidant and disorganized/disorientated) attachment patterns have been identified.
Level of HPA activation is the most commonly used measure of stress response to the SS. Infants with secure attachment do not demonstrate significant HPA activation in response to the SS [31–33]. These infants are able to obtain comfort and reassurance from the caregiver in stressful and/or frightening situations [34]. They develop a model of the caregiver as reliably available in time of need, and, over time, become capable of autonomous regulation of emotional distress [23, 24, 35].
Infants with insecure ambivalent attachment patterns are intensely distressed by the separation in the SS. When the caregiver returns they show resistant and ambivalent behaviour, seeking close contact, yet showing anger, and an inability either to derive comfort from the caregiver, or resume play and/or exploration. Observations of mother–infant interactions in infants with insecure ambivalent attachment patterns, show that the mother is inconsistent and unreliable in her responses, both to the infant's distress and to signals for comfort and reassurance [34, 35].
Infants showing insecure avoidant attachment show little or no distress during mother's absence in the SS and do not seek contact on her return. These infants have been subjected to rejecting, sometimes angry responses to their distress, and have learned that proximity to mother is best maintained by not displaying distress or calling for help and comfort at times of need [34, 35].
Infants with insecure avoidant and insecure ambivalent attachments tend overall to show greater HPA activation in response to the SS than infants with secure attachments [31–33]. Infants with insecure disorganized/ disoriented attachments show the greatest levels of HPA activation in response to the SS [36]. They lack an organized response to separation and reunion, and represent the group of infants most likely to have been subjected to severe stress, including physical abuse, gross neglect, or to have a caregiver with unresolved loss or trauma [36, 37].
It has become clear that secure attachment acts as a buffer against stress in infants exposed to stressful or frightening experiences [23, 24, 33]. Infants with insecure attachment have been subjected to the cumulative effects of repetitive stressful experiences in relation to a primary caregiver not able in any reliable way to provide emotional comfort and reassurance for her distressed infant. The infant in this situation must rely on behavioural strategies or primitive forms of defence [38] to manage distressing situations. When the stress-managing capacities of these mechanisms are exceeded, the infant's primary and intense emotions will not remain regulated within safe limits. It is likely that the cumulative, repeated effects of psychological stress in these infants will result in HPA activation with prolonged elevation of cortisol levels and consequent adverse effects on the neural pathways involved in stress regulation [39].
There has been argument as to whether the development of attachment patterns is dependent on experiences with primary caregivers, or more genetically derived. Siegel [23], p.71] gives a succinct summary of the arguments for the experience-dependent development of attachment patterns, as follows: (i) it has been shown that the child's attachment status is independently correlated with each parent and is related to the parent's Adult Attachment Interview (AAI) findings. (The AAI is a semi-structured interview developed by George, Kaplan and Main [George C, Kaplan N, Main M: unpublished data] to determine the attachment status of adults, which has been extensively used in attachment research); (ii) the child's temperament, strongly genetically determined, does not predict the child's attachment status; (iii) factors in adults which are strongly genetically determined (intelligence, memory, etc.) do not correlate with AAI findings; and (iv) the AAI findings of a pregnant woman and her husband can independently predict the child's attachment status some 18 months later, suggesting that ‘the AAI is not a measure of the parent's response to parenting that particular child’ [23], p.71].
In the first volume of his famous trilogy [27] Bowlby stated that developmental processes could best be understood as the product of a unique genetic endowment interacting with a particular environment. He predicted that it would be found that the attachment relationship directly influences the infant's capacity to cope with stress by impacting the maturation of a ‘control system’ in the infant's brain that comes to regulate attachment functions. His prescience is striking in the light of current knowledge.
Psychological theories of depression
From the time of publication of Mourning and melancholia [40] in 1917 to the present, theorists and clinicians have connected the onset of depression with the experience of loss, real or symbolic, or separation [40–42]. In the light of the neurobiological findings delineated above, we might say that some of the most important losses relate to deprivation of experiences promoting secure attachment from infancy and beyond.
Bibring [43], Seligman [44] and Bowlby [45] all focus on the sense of helplessness and powerlessness associated with the onset of depression. Bowlby, however, brought the sense of helplessness and powerlessness directly into the field of attachment with his statement that the principal issue about which a person feels helpless is his ability to make and maintain affectional relationships [45], p.247].
Aaron Beck [46, 47] proposed that depression is not a primary disorder of mood, but the consequence of how the patient thinks about himself, his world and the future.
Bowlby agreed with Beck with regard to the development of negative cognitive schemas in depression-prone individuals. Yet, unlike Beck, he stressed that these schemas develop as an understandable outcome of the child's experiences in interaction with primary caregivers [45], pp. 249–250].
As Bowlby put it The depressed person is likely to have had the bitter experience of never having attained a secure relationship with his parents, in spite of repeated attempts to do so, including having done his utmost to fulfil their demands and perhaps also their unrealistic expectations they may have had of him. These childhood experiences result in him developing a bias to interpret any loss he may later suffer as yet another of his failures to make or maintain a stable affectional relationship [45], p.247].
Research on parent–child relationships and depression
Many researchers have found that the experience of loss or separation in childhood is linked with an increased risk of depression in adult life [2, 48]. Bifulco and others present compelling arguments to suggest that it is the quality of care the child receives before and after the loss which determines the predisposition to depression, rather than the loss as such [49].
Researchers have also explored the nature of the parent–child relationship for patterns predisposing to depression. Gordon Parker has established that individuals predisposed to neurotic depression in adult life have often experienced childhood adversity in the form of lack of care and warmth, frequently with the additional factor of overprotective or intrusive parenting [50–54]. Using the Parental Bonding Instrument (PBI), a reliable and well-validated instrument addressing the dimensions of caring/rejection and protection, he demonstrated that those who have experienced ‘affectionless control’, characterized by low care and highly intrusive, controlling parenting are particularly predisposed to neurotic depression in adult life.
Extensive research (including longitudinal studies) has been carried out looking at the effect of parental depression on infant development. According to Seifer and Dickstein [55], p.126] ‘the studies converge on the finding that depressed mothers are less contingently responsive, more disengaged and more negative during dyadic interactions with their infants. In turn the infants are less positive and more negative.’ Several studies show that maternal depression results in higher rates of insecure attachment in the child [56–58].
Blatt and others [59–62], hypothesize that people who lack secure internal working models of caregiving relationships (i.e. lack security of attachment) are predisposed to depression in adult life after particular adverse life events. Blatt and Homann [59] argue that insecure ambivalent attachment may lead to depression focused on issues of dependency, loss and abandonment, whereas insecure avoidant attachment may result in depression focused on issues of self-worth and selfcriticism. There is a considerable body of research examining this hypothesis, for which Mazure provides a useful review [4], pp. 302–303]. To summarize the findings to date, there is strong evidence that the combination of high dependent needs and recent adverse social events is an important predictor of depressive onset [4]. The findings are less clear-cut with respect to the trait of self-criticism, largely because of methodological discrepancies in the studies available. We would suggest with Coyne and Whiffen [63] that the trait of dependency with consequent vulnerability to abandonment issues and the trait of self-criticism with consequent vulnerability to perceived failures in achievement, are not mutually exclusive. Having both traits is likely to render the individual particularly vulnerable to depression.
Stability of attachment patterns
There is an important question concerning the stability of attachment patterns from infancy onward. Bowlby [64] and Sroufe [65] argue that these patterns remain stable over time. Fonagy [66] reports studies by Hamilton [67] and Waters et al. [68] showing substantial continuity of secure versus insecure attachment from 18 months to 20 years. Other studies have not shown such substantial continuity [69], and Michael Rutter writes wisely about the likelihood of discontinuities [70]. It is apparent, however, that some insecurely attached infants will continue to be subjected to the environment that first produced the insecurity of attachment, with no positive intervening event(s) or relationship to effect any change in their established pattern of attachment.
Conclusions
There is a group of patients predisposed to depression on the basis of adverse early relational experience. Sensitization of depression pathways in the brain, mediated by sensitization of the HPA axis, underlies this predisposition.
The neurobiology of attachment offers a means of integrating findings concerning the effects of early life experience on brain development, sensitization of the HPA axis under stress and predisposition to depression and other psychiatric disorders. This integration is important for creating a bridge between the effects of interpersonal relational experiences in early life and brain/mind development. It has implications for understanding the development of psychopathology and what may be protective factors.
Current findings in attachment studies suggest that security of attachment can act as a buffer in the face of stressful experience in early life, while insecure attachment correlates with increased HPA activation under stress. According to Fonagy [71] the cornerstone of secure attachment is sensitive caregiving with the caregiver's perception of the child as an intentional being at its root. The question is how best to promote sensitive caregiving in caregivers who have not themselves received it and how best to promote the caregiver's perception of the child as an intentional being, in other words reflective function, in caregivers who do not yet have this capacity?
Clarification of factors and clinical interventions which promote secure attachment and those with protective or preventative effects after adverse early life experiences will be of central importance for the psychiatry of the future. Fonagy [oral presentation, 2001] summarizes the significance of this endeavour well, Even if we do not believe that secure attachment will offer protection, given that we suspect that insecure attachment creates a vulnerability, and given that we cannot realistically guarantee to protect our children from social hardship, it behoves us as a civilized society to try to do everything we can to prevent a situation arising where our children have inadequate psychological resources to deal with the challenges they will inevitably encounter.