Abstract
Trauma-related conditions
A considerable literature has developed about the subject of borderline personality disorder (BPD) since its formal recognition in 1980. However, its aetiology remains controversial. Saunders and Arnold [1] summarise the various aetiological approaches to BPD, which have been proposed in recent years. They include, for example, links to schizophrenic or affective illnesses [2,04,5] to derailments in the separation-individuation process [6, [7] to actual early losses and separations [8,9,10,11] to severe empathic failures during the first years of life [12] or to more general family based pathologies [13,14,15,16,17,18,19,20]. To these must be added the influential theory of Kernberg [21], who sees the major manifestations of the borderline condition as defences against powerful but unacceptable aggressive drives. Van der Kolk et al. [22] point out that none of these approaches encompasses the observations of Stern [23] in his original description of BPD. Stern noted that, in most cases, there was a history of ‘actual cruelty, neglect, and brutality by the parents’. Stern's observation has gained support from data generated in recent years [24,25,26,27,28,29,30,31,32], which suggest that in a typical cohort of borderline patients, 60–80% will have had a history of abuse of some kind.
The incidence of abuse in BPD is substantially higher than that for the general population [33]. It is hard to avoid the possibility that it has aetiological significance. However, a construction of a theory of the origins of borderline personality disorder based on trauma is complicated by the fact that such a background is not unique to BPD.
Conditions other than BPD, most notably somatisation disorder and various dissociative states, have a developmental history marked by abuse. Such a background is particularly striking in the case of dissociative identity disorder. A number of studies [34,37] suggest that a history of physical and/or sexual abuse is present in nearly 90% of cases. Childhood abuse is also common in the history of those who suffer less severe dissociative states [38]. A background of abuse has been found in those with somatisation disorder [39] and somatising presentations [40, [41] which frequently involve pain [42,47]. An association also exists between somatisation and dissociative states [48, [49]. Predictably, there is also a relationship between somatisation, dissociative states and a background of abuse [50, [51]. In addition, there is an established relationship between borderline and dissociative states [52] consistent with DSM-IV. Finally, there is an association between somatisation disorder and Cluster B personality disorder [53], the various categories of which overlap with BPD [54].
The shared developmental history of BPD, somatisation disorder and dissociative states leads to the possibility that they are related conditions. This possibility is supported by a consideration of the abandoned syndrome of hysteria, studied by Janet [55] and others at Salpetriere during the last century. The stereotypic case of hysteria might be called ‘chronic hysteria’ in order to distinguish it from a conversion disorder immediately following severe stress [56]. Such a patient showed, in varying degrees, somatisation as catalogued by Briquet [57], dissociative states, conversion disorders and personality problems. This syndrome was fractured by DSM-III into its component subsyndromes [58], implying that these conditions arise independent of each other. Although this is sometimes the case, the distinction is often artificial. These diagnostic categories frequently coexist, a fact acknowledged, in part, by the recent DSM revision which now includes dissociative states among the criteria for BPD. Such clinical evidence leads to the suggestion that traumata potentiate the emergence of specific and interrelated symptom clusters which may or may not all be manifest in a particular patient.
The above suggestion is supported by several investigations. For example, Briere and Muntz [59] compared those 15% of a sample of 278 University women who had had a sexual contact with a significantly older person before the age of 15 with those who had not. The abused group reported higher levels of dissociation, somatisation, anxiety and depression than did the remainder. A more extensive study was conducted by van der Kolk and his colleagues on 328 subjects who had experienced trauma either early in life or later [60]. It was found that 21.8% of the sample of 328 subjects suffered posttraumatic stress disorder (PTSD) alone. However, another 47.2% suffered PTSD in conjunction with what van der Kolk has called ‘disorders of extreme stress’ (DES). These disorders involve a number of symptom clusters. The main groupings include: (i) impairment of regulation of affect and arousal; (ii) dissociation and amnesia; (iii) somatisation; (iv) alterations of perceptions of self including guilt, self blame, ineffectiveness, alterations in relations with others, including an inability to trust or maintain relationships; and (iv) alterations in systems of meaning including despair and hopelessness.
A rather similar grouping of symptoms following trauma was described by Bower [61] in his follow-up of holocaust survivors. However, the clinical picture of ‘the concentration camp syndrome’ provides additional information about the effects of severe trauma. Bower found that his subjects had, typically, suffered a permanent personality change. This alteration involved the explosiveness characteristic of the borderline. Other behaviours were frankly sociopathic.
Bower's findings support earlier observations of a persisting personality change following massive and prolonged trauma [62, [63]. It involves an alteration in the sense of self such that, in an extreme case, the individual might say ‘I am not a person’ [64].
Herman [65], in reviewing this subject, points out that the syndrome of complex symptomatology which follows prolonged and repeated traumata has received no formal recognition and has been the subject of very little systematic inquiry. In contrast, the diagnosis of PTSD has been accepted since 1980. The two syndromes are obviously related. Terr [66] has suggested a categorisation of Type I and Type II trauma, the former syndrome concerning the effects of a single trauma, while the latter is a consequence of prolonged and repeated traumata. In a broad sense, Type I and Type II might be linked in the same way that PTSD is related to borderline phenomena. Such a relationship has been hypothesised by a number of writers in recent years [21, [67,72].
The various pieces of evidence touched upon so far lead to the conclusion that a history of trauma, particularly if repeated and prolonged, might lead to a ‘long-standing severe personality disorganisation’ [73] to which is related a tendency to dissociate and to somatise. This state is typically accompanied by a pervasive dysphoria, akin to depression, but which can be distinguished from depression arising in non-borderline patients [74,76]. Depending upon the relative prominence of the above symptom clusters, such patients might receive the diagnoses of BPD, dissociative identity disorder, somatisation disorder, or dysthymia [77].
To talk of ‘double diagnoses’ in these situations is to miss the point. The data concerning the sequelae to trauma suggests these are not independent diagnoses, but refer to functionally interrelated states.
Van der Kolk et al. [21] note that a substantial proportion of those who no longer meet criteria for PTSD after traumatic experiences ‘continue to suffer from high levels of dissociation, somatisation and affect dysregulation’ (i.e. phenomena associated with BPD).
A pathogenic familial background?
Not all BPD patients have not been forced to endure trauma of the kind included in the categories of sexual, physical, or emotional abuse? Furthermore, some people have suffered abuse equivalent to those with borderline personality disorder yet they have no adult psychiatric diagnosis. It is evident, therefore, that to make a one-to-one relationship between abuse and BPD is to oversimplify the problem, as Paris and Zweig-Frank [78] point out. A vulnerability factor or factors must intervene, in the typical case, between the incidents of abuse and the consequence of BPD.
The most obvious vulnerability factor to consider is genetic. However, there is little evidence that BPD is genetically transmitted [79]. A second possibility is neurological. Recent observations suggest that BPD is associated with ‘soft’ neurophysiological deficits, particularly of a frontal kind. Goyer et al. [80] used positron emission tomography to examine rates of cerebral glucose metabolism in 17 patients with various personality disorders. They found a significant inverse correlation between a life history of impulsive aggressiveness and activity in the frontal region. Van Reekum et al. [81] report that a neuropsychological investigation of nine BPD subjects indicated frontal dysfunction in seven of them. Nevertheless, it is unlikely that any neurophysiological deficit identified in a BPD group will be of a simple localised kind. A systematic dysfunction is more probable. Two recent neuropsychological studies show a complex pattern of deficits [82, [83]. Although such data are sometimes used to argue an ‘organic’ basis to BPD it is equally plausible to suppose that subtle neurophysiological disturbances found in BPD are the results of environmental insults or failures. Increasingly, evidence suggests that certain cognitive and memory deficits might be the consequences of severe trauma [84,88]. One finding suggests that a reduced hippocampal volume is a correlate of these memory deficits [89]. Emotional memories have been found to be enduring and associated with long-term potentiation (LTP) in the amygdala, with the prefrontal cortex ‘controlling expression of emotional memory’ via the amygdala [90].
The vulnerability factor which seems most likely to be operative in the development of BPD is a dysfunctional family environment. A study from Nash et al. suggests such a supposition [91]. One hundred and five abused and non-abused women were examined for patterns of adult psychopathology associated with childhood sexual abuse and to test the extent to which these patterns are independent of other pathogenic properties of the family environment. It was concluded that greater non-specific impairment among abused women may be a consequence, at least in part, of pathogenic family structure rather than sexual abuse per se. Goldman et al. [92] produced complementary data comparing 44 young BPD subjects with 100 non-BPD patients. They found significantly higher rates of psychopathology among the family members of those with BPD. These findings are consistent with the chaos remarked upon by Golomb et al. [93] as typical of the BPD family background.
Attempts have been made to determine whether specific deficiencies can be identified of the family background of those with BPD. Golomb et al. [93] tested the hypothesis concerning failure of parental empathy in the developmental history of those with BPD. They found mothers of BPD subjects treated their children egocentrically, as need-gratifying objects. These observations lead to several predictions. First, since their children were required to meet maternal needs, over involvement and control would be characteristic of BPD parents. Second, since the maternal orientation is towards the parent's needs rather than the child's, the parenting style will be inconsistent. An epidemiologically designed study [94] of a random sample of 776 adolescents supported both predictions. Maternal inconsistency in upbringing of the child predicted BPD but not any other Axis II disorder. However, this effect occurred only in the presence of high maternal over involvement. Neither maternal over involvement in maternal inconsistency alone predicted the emergence of BPD.
Maternal inconsistency is likely to inflict mini-traumata on the child. A failure of responsiveness and/or actual absences will lead, in some cases at least, to the production of separation anxiety and a disruption in the child's development of a ‘secure base’. This prediction was tested by West et al. [95] on a sample of 85 female outpatients. The dimensions and patterns of reciprocal attachment were compared with other scales measuring components of psychopathology and interpersonal relationships. Four of the attachment scales (feared loss, secure base, compulsive care-seeking, and angry withdrawal) identified patients with high scores on a measure of borderline disorder. Of these four scales, feared loss had the predominant effect.
Although the above studies focused on the mother-child dyad, other investigations suggest that failures in paternal care are also important [96, [97]. Further study is needed in order to gain a better understanding of the interplay between a potentially pathogenic family background and the incidents of abuse. Nevertheless, the available evidence makes it reasonable to suppose that the ‘chaotic’ family background of the borderline-to-be, demonstrated in the large majority of this series, creates a vulnerability to the effect of abuse. This hypothetical vulnerability can be understood, at least in part, in terms of the main thesis of this paper which is that the developmental background of those with BPD disrupts the emergence of higher order psychological functions which include the system of self.
A Jacksonian hypothesis
A disturbance of self is a central feature of the diagnostic criteria for BPD. The great neurologist, Hughlings Jackson, believed himself to be the first to use the word ‘self’ in the medical literature (Vol. II, p.96 [98]). Jackson's ideas are helpful in developing a hypothesis concerning the origins of BPD.
Jackson was born in 1835, 21 years before Freud, who was influenced by him. Jackson was a pioneer theorist in the sphere of mental illness. However, since his ideas were not well understood in his own time they were relatively neglected. He is remembered for his neurological contributions, although much of his opus involves an attempt to understand mental illness in terms of evolutionary theory.
Jackson's main hypothesis is well known. He suggested that those functions which have evolved last and which emerge late in human development are more fragile, more easily disrupted by insults to the brain-mind system than those functions which, as it were, are more hard-wired and appear earlier in evolutionary history. This process, the reverse of evolution, he called ‘dissolution’.
A Jacksonian hypothesis for BPD must not only explain disturbances of self but also those deficiencies in regulation of impulse and affect which are also central features of the syndrome. In addition, it should illuminate the origins of somatisation and dissociation since these are related phenomena. In the following discussion, we consider each of these symptom clusters in terms of the dissolution hypotesis: (i) self; (ii) affect-impulse dysregulation; (iii) somatisation; and (iv) dissociation.
Self
The problem of self was prominent in Jackson's theoretical system. His thinking was logical and scientific. He began his argument with fundamentals. Since he conceived the study of mental illness as ‘an experimental investigation of mind’ (Vol. II, p.4 [98]), it was necessary to begin with a definition of mind or self. He saw mind as a function of brain. Nevertheless, the two were not to be considered identical. There was, in his view, a ‘parallel concomitance’ between them (Vol. II, p.85 [98]).
Jackson conceived mind, or self, as double, made up of two poles, subject and object, one pole of awareness, the other of the objects of awareness. The objects of awareness of which self consisted were those of inner life, which he described in terms of imagery (Vol. II, p.93 [98]). Williams James developed this fundamental idea of the duplex self, calling the objects of consciousness ‘the stream of consciousness’ [99].
Self, then, in the theoretical system of Jackson and James was an awareness of the stream of consciousness. Both men emphasised that self was neither a thing nor a substance but a process, a result of what Jackson called a ‘coordination’. He explained this idea in the following way.
Jackson built up his model of the central nervous system (CNS) from its simplest functional unit. For Jackson, the basic unit was reflexive, the smallest element of sensorimotor function. He conceived each of these units as a representing system. The brain, in his view, evolves and develops through an increasingly complex coordination of these units. As the organism evolves to a higher stage of function, it is not as if something new were being tacked on. Rather, there is a re-representation, and at a higher stage still, there is a re-re-representation, so that the most recently evolved part of the brain, the neocortex, is ‘universally representing’ (Vol. II, p.82 [98]). The whole nervous system, he considered, is a ‘sensori-motor mechanism, a co-ordinating system from top to bottom’ (Vol. II, p.41 [98]).
So where is self in this coordinating system? Jackson rejected the idea that any special new form of neural function had been built into the human brain. The appearance of self, he believed, is the manifestation of a more complex coordination than previously. What is new is a more complete system of unification of the whole organism. Nevertheless, the emergence of self is dependent upon the evolution of anatomically new structures, most importantly, he suggested, the prefrontal cortex (Vol. II, p.399 [98]). This structure, he considered, is necessary to the cerebral process which gives rise to self.
Jackson's model closely corresponds to a recent model of self by the noted neuroscientist, Antonio Damasio [100]. Like Jackson, Damasio considers that the evolution of prefrontal activity is necessary to the emergence of self.
According to the Jacksonian hypothesis, functions which involve the prefrontal area, being the last to evolve, should be lost first as a consequence of insults to the brain-mind system. Accordingly, the development of ‘self’ will be relatively deficient.
The stream of consciousness will be stunted, or perhaps, absent, leaving a painful emptiness which is a cardinal feature of BPD. Much of the impulsive and maladaptive behaviour of BPD can be understood as a desperate attempt to fill this emptiness. The subject is as if addicted to stimuli, including certain of the complex stimuli provided by the social environment.
Affect regulation
Affect regulation, recent studies suggest [101], is dependent upon prefrontal activity. This idea has been extensively developed by Schore [102], who emphasises the significance of the right orbito-frontal cortex. The evidence he reviews supports the possibility that a cascade of descending inhibitory tracts, emerging from frontal and prefrontal areas, provide the capacity for higher order modulation of affective expression. Deficient development of the function of these inhibitory systems will lead to a relative dysregulation of emotional experience and expression. Affect dysregulation as a consequence of an insult to CNS function is predicted by Jacksonian theory, which postulates not only coordinating, elaborating and selecting activities for late evolving aspects of brain function, but also a certain degree of inhibitory control over those which emerged earlier. Van der Kolk, citing the evidence of his own study [60], noted that ‘the lack of loss of self regulation is possibly the most far-reaching effect of psychological trauma in both children and adults’ [103, p. 187]. The younger the age at which the trauma occurs, and the longer its duration, the more likely people are to have long-term problems with the regulation of anger, anxiety, and sexual impulses.
Somatisation
A preoccupation with, and exaggerated awareness, somatic sensation is a feature of the conversations of many people with BPD. This often takes the form of extended catalogues of complaints about symptoms and bodily distress [104]. The explanations of this phenomenon are various and of a psychosocial kind. We suggest another basis to this behaviour, which can be seen as complementary to these explanations. The somatisation found in BPD may be a consequence, at least in part, of a disturbance of attentional focus [105]. This disturbance is conceived, once again, as a disruption of the activity of a notional cascade of neural loops emanating from the prefrontal region. Areas of the prefrontal region, which differ from those involved in the regulation of affect are concerned with attention [101].
Selective inattention, which must involve descending inhibitory mechanisms, is a crucial aspect of the development of effective attentional processes. Without the capacity to ‘screen out’ or ‘turn off’ redundant stimuli, the individual will be unable adequately to focus on meaningful stimuli. A profound deficiency of this function, as indicated by the habituation phenomenon, has been demonstrated in ‘hysteria’, the diagnostic ancestor of somatisation disorder [106]. Somewhat similar findings are typical of BPD [107]. A review by van der Kolk shows that they are also typical of those with PTSD [108].
Those with somatisation disorder also show increased amplitude of event-related potentials in response to auditory stimuli when compared with controls [109,111]. These data are consistent with a perception of increased stimulus intensity. In addition, these patients have an impaired capacity to distinguish finely between meaningful and non-meaningful stimuli when these attentional processes are indicated by amplitudes of event-related potentials [112]. McFarlane, Weber and Clark [113] report similar data for traumatised individuals.
It seems not unlikely that these deficiencies in attention in those with somatisation disorder are related to their common experience of pain, which might be seen in some cases as a manifestation of failure to develop adequate systems of stimulus intensity control [114] akin to the failure of affective control. Once again, it seems reasonable to suppose that such systems depend upon descending inhibitory mechanisms.
Dissociation
Another of the phenomena of BPD, dissociation, can be understood in terms of Jackson's notion of the hierarchial organisation of the CNS, the highest and most complex state of function being the most fragile and easily disrupted [115]. Pierre Janet, who provides us with the most authoritative description of dissociation, is likely to have been influenced by Jackson. Ribot, Janet's mentor, was a strong protagonist of the Jacksonian viewpoint.
Dissociation refers to a mental state which is sequestered from ordinary consciousness. Waller, Putnam, and Carlson [116], who have written extensively on this subject, identify five key features of dissociation: disturbances of memory; depersonalisation; derealisation; a discontinuity of personal existence; and hallucinatory phenomena. To this list must be added constriction of the field of consciousness which was central to Janet's description. We consider each of these features in turn.
Self as the stream of consciousness involves not only an awareness of the movements of inner life but also a certain kind of memory which concerns the ability to remember episodes from one's life occurring in the distant past. This memory is termed remote episodic or autobiographical.
It has recently become apparent that memory is not a unitary function but is made up of a number of ‘modules’ [117] which do not necessarily all function at one time. One module may be lost as a consequence of brain damage while others continue to function. These modules do not develop all at once but emerge serially during the child's development. The last to develop is remote episodic which appears at about 4 or 5 years [118]. An earlier form of episodic memory emerges at about 2.5 years of age. It concerns recent events. Episodic memory depends upon prefrontal activity [119,120].
Typically, in states of severe dissociation, the autobiographical and episodic forms of memory are lost, while an earlier form of memory which concerns knowledge of the world (i.e. facts) is retained [121]. Tulving called this more primitive form of memory, which appears late in the first year of life [122,123], semantic [124,125]. (It becomes declarative when the child begins to use words in the latter half of the second year of life.) Autobiographical memory has not yet been investigated in BPD. However, it is disrupted in subjects who have been abused early in life [88].
Another of the cardinal features of dissociation, depersonalisation, can be understood as a consequence of the loss of the stream of consciousness. Depersonalisation involves a diminished sense of personal reality, or ‘self’. The stream of consciousness which involves a core of personal feeling which might be called ‘me-ness’ (p.107 [126]). This feeling fades as the stream of consciousness is lost. Its loss is accompanied by a sense of estrangement.
A second consequence of the loss of the stream of consciousness is predicted from Jacksonian theory. Since this experience is a manifestation of ‘coordination’ of the components of CNS function, its loss will result in a relative dis-coordination, and a failure adequately to synthesise sensory data. There will be a tendency for consciousness to break up, ‘split’, into constituent parts. The continuity and coherence of self will be lost. This is a central feature of BPD. It is also found in victims of massive trauma who experience a fragmentation of ‘self’ [127, p.370].
The hallucinosis which is sometimes an aspect of dissociative states and which may accompany BPD, can also be explained in terms of Jacksonian theories. He considered the phenomenon in terms of ‘faces in the fire’ phenomenon (Vol. II, p.24 [98]). We all have the capacity to see shapes in the flames of a fire, in clouds, and in rocks. They are often faces. However, a higher order monitoring and evaluating function results in these experiences being judged illusory. Loss of this function, which might correspond to Janet's ‘fonction du réel’, leads to the reality of these illusory perceptions remaining unquestioned. This approach to the origin of hallucinosis has recently been explored, in a number of brain imaging studies (e.g. [57]).
Dissociation also involves changes in attention, which, as previously remarked, also depends upon prefrontal activity. Recent evidence suggests that the prefrontal cortex is part of a system which is involved in holding representations of sensory information on line, in working memory [128]. It gives us the capacity to keep a number of channels of information open at one time. If this system fails, the number of channels of information kept open will be reduced. There will be a ‘constriction of consciousness’, a cardinal feature of dissociation as Janet [129] described it.
Discussion
Jackson's colleagues considered him a genius. He pioneered the theory of lateralisation of hemispheric function (Vol. II, pp.129–145 [98]), anticipated MacLean's notion of the triune brain (Vol. II, p.39 [98] and [132]), and suggested that dreaming has the function of integrating sensory data (Vol. II, p.71 [98]). In this paper, we use his ideas to provide a preliminary hypothetical framework for approaching borderline-related pheomena.
Jackson considered that the higher psychological functions, which appear most recently in evolutionary history, are less ‘complete’ than those which emerge earlier. He implied, then, that suitable environmental circumstances are necessary for their maturation. This idea was developed by Vygotsky and Luria [133], who proposed that maturation involves the coordination of a genetically given biological program, which they called ‘biogenesis’, with appropriate environmental provision which they called ‘sociogenesis’. Where ‘sociogenesis’ is deficient, maturation is impaired. The highest levels of psychological function, which, in Jacksonian theory, include those upon which the experience of self depends, are the most vulnerable to maturational disturbance. We are suggesting the borderline and related phenomena are a consequence of a failure of ‘sociogenesis’, causing an impairment of late developing neural networks, of largely prefrontal origin, which link a number of brain regions including the amygdala and hippocampus. The deficient maturation of this system has effects on the proper establishment of higher order functions which involve attention, memory, and affect. Disturbances in these spheres will be subtle and depend upon mechanisms which involve ‘fine tuning’ and an increased voluntary control over aspects of psychic life.
This proposal is testable using current brain imaging and electrophysiological [107] technologies. These technologies will increasingly allow evaluation of more refined brain models, concerning more explicit network disturbances of attention, memory and emotion. The highly interconnected nature of the brain, demands models (and measures) that more specifically reflect the multiple and parallel network interactions across scale, that underlie dysfunctions association with trauma, memory and self [130,131].
Finally, since the suggested maturation failure is conceived as ‘experience-dependent’, it is considered to be reversible, at least in part, so that psychotherapeutic approaches, which are focused on restoration of the lost functions [126,134], particularly the reflective function, offer the hope of beneficial change in these difficult conditions.