A resurgence of
interest in the area of personality disorder has been seen in the past 20
years. Researchers have continued along the biological line of investigation
working from diverse theoretical standpoints and using increasingly
sophisticated approaches. In this review these aspects are set against the
background of current debate in the field and placed in their historical
context.
Given that much related work has
been carried out using a variety of terms such as episodic dyscontrol, impulse
control and impulsive aggression, the computer-assisted search was
augmented by a manual search.
The aims of this overview are to
outline current debates and
to trace the develop-ment of biological approaches in this field. In so doing,
conditions mimicking per-sonality disorders will be discussed.
The standard
classifications of personality disorder have undergone major revision in the
past 20 years. Differences still exist between those in current use, namely the
International Classification of Diseases, ninth revision (ICD-9) (World Health
Organization,1978) and the Diagnostic and Statistical Manual of Mental Disorders,
third edition and its later revision (DSM-III and DSM-IIIR) (American
Psych-iatric Association, 1980, 1987). Changes that have been made during the
development of 10th revision of the ICD (World Health Organization, 1989) have
resulted in some convergence between the two systems and it has been suggested
that the remaining differences are largely seman-tic (Tyrer et al, 1991).
However, debate continues as to how personality and its disorders are best
described.
The categories of
abnormal personality have long been challenged. Much criticism has resulted
from the poor agreement between clinicians as to what kind of personality
disorder an individual has and whether or not the individual has one at all.
Although DSM-III with its provision of operational definitions has led to some
improvement, many continue to advocate a so called dimensional approach as an
alternative (Widiger et al, 1988).
The essential difference between
the two approaches lies in the questions they pose. The categorical approach
asks 'Do this patient's characteristics meet the diagnostic criteria or not?',
whereas the dimensional approach asks 'How much of the given characteristic
does this patient display?' Perhaps the best known advocate of the dimensional
approach in this country is Eysenck, who has suggested that much personality
variance can be accounted for by the dimensions of extroversion-introversion
(E), neuroticism-stability (N) and psychoticism-normality (P). These and other
dimensions referred to below have their origins in experiments conducted from
the perspective of behavioural psychology and have been described in detail
elsewhere (Lewis, 1991; Siever and Davis, 1991).
The categorical and
dimensional approaches are not mutually exclusive but trying to combine them
poses two further questions. First, which dimensions (or behaviours) are
most relevant to personality disorder? Second, are such behaviours abnormal in
themselves? Dimensional approaches tend to assume that the characteristics
under consideration are present in all individuals and exaggerated in those
with personality disorders.
From a slightly different
perspective it has been argued that personality disorders represent attenuated
forms of frank psychiatric illness. Thus, the personality disorders in
DSM-III cluster A may consti-tute part of a schizophrenia spectrum, those in
cluster B an affective disorder spectrum and those in cluster C neurotic
conditions. As yet the evidence seems most clear for the cluster A relationship
and less clear for the other two.
Overall, the biological approaches
to re-solving these and other issues of aetiology have become more
sophisticated as techno-logical advances have been made. In keeping with this,
the focus of attention has shifted over the last century from signs that are
detectable on examination to models of CNS neurotransmitter systems.
One 19th century
formulation of moral insanity, upheld by a number of influential authors, was
that it represented a degenera-tive process arising in 'poor stock', which was
handed on to successive generations, ultimately leading to extinction of the
line (Lewis, 1974). Attempts were made to iden-tify and describe the outward
manifesta-tions of the condition but this process was more successful in
excluding other physical conditions (such as encephalitis lethargica) from the
category than in establishing objective inclusion criteria (Schneider, 1958).
Subsequently, Kretschmer (1936) and
Sheldon (1954) attempted to relate physical characteristics and body build to
various personality types. However, this approach attracted little empirical
support ('1"rimble, 1981) and is of historical rather than prac-tical
interest.
The genetic aspects of personality
dis-order have recently been extensively re-viewed (McGuffin and Thapar, 1992).
Prob-lems of case definition hamper such work and modes of transmission are
undoubtedly complex. However, these authors suggested that on balance there is
reasonable evidence to support the idea of a genetic component in many
categories of abnormal personality and felt that the future localization of
quantitative trait loci (loci that relate to some continuously distributed
character-istics) may shed some light on this area.
The role of temperament in the
development of personality and personality dis-order has been reviewed in some
depth by Rutter (1987). Given that the character-istics which differentiate
children are not necessarily those which differentiate adults, it is difficult
to establish broad clear links between specific childhood behaviours and adult
personality disorders. Perhaps the strongest suggestions are those relating
childhood hyperactivity to later antisocial behaviour.
Undesirable or even aggressive behaviour is not synonymous with personality
disorder and the false attribution of such behaviour may lead the clinician
into overlooking other important diagnoses (Lewis and Appleby, 1988). In the
general hospital setting it is important that organic causes of aggressive
behaviour are considered, and information obtained from the patient 5 relatives
or close friends may clarify the situation greatly. In addition to a full
history and examination, several points are worth considering as follows.
The personality
characteristics of those with cluster A and C diagnoses tend to be stable over
time. Cluster B characteristics tend to become less pronounced in the patients'
third or fourth decade (Tyrer et al, 1991). Therefore aggressive or antisocial
behaviour arising after that age is less likely to be related to personality.
By definition the
behaviours associated with personality disorder are evident in early life and
persist into adulthood. Their develop-ment against the background of an
unre-markable Premorbid personality may well point to either functional or
organic illness.
In general,
behaviour that is frankly bizarre (e.g. micturition in inappropriate places) is
unlikely to represent personality dysfunc-tion, especially when it is
associated with other mental state abnormalities (e.g. delu-sions or
hallucinations).
A pervasive nature
for abnormal personality traits is assumed in the diagnosis of person-ality
disorder and so behaviour that is highly specific to a given situation does not
usually come under this heading.
Disturbances in
cognitive function such as disorientation and memory impairment point to an
organic or conceivably a func-tional aetiology. Those with learning
diffi-culties are excluded from the category of personality disorder.
The range of possible organic
causes of aggressive behaviour is illustrated in Table 3 and the reader is
referred to Lishman (1987) for a more detailed account. The historical
relationship between personality disorder and epilepsy is complex and is dealt
with separately.
The episodic nature
of the violent outbursts displayed by patients led writers to propose that
epilepsy was important in the aetiology of moral insanity. Schneider later
refuted this and used the term explosive personality to describe a similar
group (Partridge, 1930). However, the suspicion of a link between the two
persisted, and gained credence from the electroencephalographic (EEG) studies
of the 1940s (Gurvitz, 1951). These and subsequent studies provided evidence of
delayed cortical maturation in 'psychopaths' on the basis that slow waves
particularly in the temporal region, with sudden bursts of activity) were more
frequent in this group than in controls. This work was limited by numerous
methodological diffi-culties and further weakened by the later finding that
psychopaths were not over-represented among violent offenders with posterior
temporal EEG abnormalities (Fenwick, 1981).
Attempts to relate clinical or
subclinical seizure activity on the one hand and impul-sive or aggressive
personality characteristics on the other have continued with the description of
the episodic dyscontrol syn-drome (Bach-y-Rita et al, 1971) and the delineation
of borderline personality dis-order subtypes (Andrulonis et al, 1982). The
former relates to patients with histories of episodic aggression in the absence
of provo-cation, who have soft neurological signs suggestive of what has been
described as minimal brain damage. This remains a con-troversial formulation.
Andrulonis et al (1982) reported histories of epilepsy in a small but
significant subgroup of those with borderline personality
disorder. This diagnosis, which has attracted much atten-tion in the past
20 years (Tarnopolsky and Berelowitz, 1987), shares some of the characteristics
associated with the term psychopathy.
Overall, it can be said that epilepsy is rarely the sole cause of episodic
aggression, although lesions and seizures in the limbic system can lead to a
lowering of impulse control and to aggression (Fenwick, 1986).
Historically, the
so-called epileptic person-ality has been described using a number of
adjectives (Fenton, 1981). For convenience, these can be grouped under
obsessional (e.g. pedantic and circumstantial), paranoid (e.g. suspicious and
touchy) and organic (e.g. concrete thinking and emotional viscosity).
Attempts to define the category
have been hampered by the problems of obtaining representative samples and
achieving a con-sensus on appropriate case definition.
Behavioural disturbance secondary
to temporal lobe epilepsy has been more formally described as the Geschwind
syn-drome (Geschwind, 1979). This constella-tion of features includes
hypergraphia and hyposexuality in addition to some of the earlier
characteristics.
Even if specific categories can be
described, identifying the cause remains a formidable task which has to take
into account such diverse factors as: age; sex; socioeconomic status;
aetiology, location and chronicity of the seizure activity; stig-matization
associated with the diagnosis and its impact on the individual's self-esteem;
adherence to medication; interper-sonal relationships; and the effects of
medication on cognitive performance and schooling. As such the exact
relationship between epilepsy and personality is likely to remain unknown for
some time.
The search for
anatomical substrates of antisocial behaviour can be traced back at least to
19th century faculty psychology. The belief that various mental faculties
(including the moral faculty) could be linked to specific brain areas generated
the hypo-thesis that moral insanity resulted from a lesion of that area
controlling the moral faculty. A variant of this hypothesis was that the
impulsive behaviour exhibited by these patients arose from an innate
defi-ciency of moral inhibition secondary to a lesion in the main control
centre. This was believed to be located in the frontal lobe (Burt, 1926).
German writers also favoured an
organic explanation, with Kleist focusing more attention on subcortical
structures. He described three aspects of self which equate with various
clinical features of personality disorder, and related them to three anatomical
sites. The somatopsyche or body self (governing hysterical impulses and sexual
behaviour) was located in the grey matter around the third ventricle; the
thymopsyche or feeling self (governing emotional lability, anxiety, lack of
affect and touchiness) was related to the thalamus; and the autopsyche or
character self (which was expansive, contentious, fanatical and compulsive) was
ascribed to the pallido-striatum (Schneider, 1958).
The work described above in
relation to epilepsy implicated the temporal lobe in the causation of
antisocial behaviour. More rec-ently the mesolimbic and septal regions have
been viewed as relevant in the genesis of episodic aggressive behaviour. The
frontal lobe continues to attract much attention but the results of
neurophysical investiga-tions of frontal lobe function have been conflicting
(Zuckerman, 1991).
On the basis of the
dimensional description outlined above, Eysenck has characterized psychopaths
as displaying high levels of neuroticism and extroversion resulting from the
associated variability in autonomic nervous system reactivity and arousal (as
mediated by the reticular activating system). As a consequence of this,
the group should show some difficulty in learning, or more precisely in
conditioning, e.g. the development of fear in response to an aver-sive
stimulus. Indeed, psychopaths have been shown to exhibit lower anticipatory
fear as demonstrated by lower galvanic skin responses. However, as Lewis (1991)
points out in quoting Hare (1978), the nature of the aversive stimulus is
important, and the appropriate autonomic responses are
obtained if the stimulus is changed to loss of reward.
Children with conduct disorders
show an increased reactivity (measured as heart rate and skin conductance) in
response to reward, which is relatively slow to fade when the stimulus is
withdrawn, and a flattened response to an intense but neutral stimulus. These features
distinguish this group from emotionally disordered children and con-trols. A
tentative interpretation suggested that an inability to modulate arousal was
evident, perhaps analogous to the high levels of sensation seeking displayed by
adults with antisocial characteristics (Garralda et al, 1991).
The proposition that personality
dis-orders in the DSM-IIIR cluster B are attenuated forms of affective
disorders has aroused interest in a possible role for sug-gested biological
markers of depression in borderline personality disorder. The dexa-methasone
suppression test, the TRH/TSH (thyrotrophin-releasing
hormone/thyroid-stimulating hormone) test and sleep EEGs have all been used in
this group (Steiner et al, 1988). The neuroendocrine investigations are a
little difficult to interpret because of the uncertain specifi-city of the
tests, and the low sensitivity of the dexamethasone suppression test in
particular, in this population. The blunted TSH response to TRH in some
borderline patients and the shortened rapid eye move-ment latency in others may
indicate an overlap with affective disorders, but definite conclusions would be
premature.
Similarly, those with schizotypal
per-sonality disorder may share patterns of abnormal eye tracking movements and
impaired information processing with schizophrenics (Siever and Davis,
1991). This would be in keeping with the notion of spectrum schizophrenia, such
that the per-sonality disorder would be a forme fruste.
Increasing evidence
has accumulated in recent years to support the role of 5-hydroxytryptamine
(5-HT, serotonin) in the central control of aggressive behaviour. It has been
suggested that this form of behaviour includes both aggression towards others
and suicide and that the association is independent of diagnostic category (van
Praag, 1991). The evidence has been taken from studies in animals and humans.
Many human experiments have revealed lowered values of 5-hydroxyindoleacetic
acid (the main metabolite of 5-HT) in the CSF of those exhibiting a variety of
aggressive behaviours. The assumption that CSF metabolite levels are an
accurate reflection of central 5-HT activity has been ques-tioned (Soubrie,
1986) but overall the data seem to support the notion of reduced serotonergic activity
being at least relevant to impulse aggression, particularly when postmortem and
other evidence is taken into account (Coccaro, 1989).
The role of the noradrenergic
system is much less clear. Although alpha2-receptor binding sites have been
reported to be decreased in anxious, unmedicated border-line patients
(Southwick et al, 1990), for example, the results of metabolite studies of
aggressive behaviour (which is important in these patients) are as yet
unconvincing (Soubrie, 1986). The use of platelet mono-amine oxidase activity
as a marker remains equivocal.
One recent proposal has been that
the personality dimensions of behavioural acti-vation (novelty seeking),
behavioural inhibi-tion (harm avoidance) and behavioural maintenance
(reward dependence) are governed by the dopaminergic, serotonergic
and noradrenergic systems respectively. It is suggested that these dimensions
are genetic-ally determined and can be related to the more traditional
categories of personality disorder (Cloninger, 1987). As yet it is too early to
evaluate this hypothesis.