Neuroendocrinology of Personality Disorders
Roger T. Mulder and Peter R. Joyce
Disorders of personality are currently defined as categories of deeply ingrained and enduring behaviour patterns that are persistently maladaptive and encompass multiple domains of behaviour and psychological functioning. Personality types were described in ancient Greece, yet, despite this lengthy history, it may be argued that we are still far from adequate in describing the phenomena that constitute the clinical core of personality disturbance (Tyrer, 1995). Personality pathology has been conceptualized in a number of ways, including categories, dimensions and clusters, but none of the classification systems have been validated from an aetiological or a biological perspective. We begin by discussing ways of classifying personality pathology and measurement of central nervous system monoamines since the patterns of association are critically influenced by these factors.
The separation of personality disorders (PDs) as a discrete axis of classification by the American Psychiatric Association (APA) in DSM-III (American Psychiatric Association, 1980) has focused attention on this neglected group of disorders and stimulated research in this area. However, the large increase in literature (Gorton and Akhtar, 1990) has not been accompanied by a corresponding increase in systematic models to aid in understanding the underlying behavioural abnormalities. The 10 currentDSM-IV categories—paranoid, schizoid, schizotypal, antisocial, borderline, narcissistic, histrionic, avoidant, dependent and obsessive compulsive—are derived from a mixture of theory, opinion and historical precedent. The 10th Revision of The International Classification of Diseases (World Health Organization [WHO], 1992) has replicated most features of DSM-IV. Diagnoses in both systems involve a list of operational criteria, of which a specific number must be present for an individual to receive a diagnosis of personality disorder.
Although it is implied that the categories group patients into mutually exclusive diagnostic entities, most studies have reported high rates of co-occurrence (e.g., Pfohl et al., 1986; Joffe and Regan, 1988; Mulder et al., 1994), and the measured behaviours appear to be distributed dimensionally with no evidence of the discontinuity a categorical model would imply (Zimmerman and Coryell, 1990). The overlap and clinical heterogeneity which result hamper attempts to link personality disorder diagnoses with specific neurobiological processes.
Unfortunately, the dimensional models have fared little better. Many have emerged from academic psychology, the most important being the work of Eysenck and the five-factor model currently being refined by Costa and McCrae (1992). Eysenck's model initially consisted of two dimensions labelled introversion/extraversion (E) (sociability, stability, activity) and neuroticism (N) (temperamental sensitivity to negative stimuli). His model included a biological explanatory schema based on autonomic nervous system reactivity and ease ofconditionability. He suggested that arousal stems from the ascending reticular activating system and is linked with E, while N reflects activation from the limbic system (Eysenck, 1967). Later critics have noted that these two systems of brain circuitry are interconnected and are thus unlikely to explain two supposedly independent dimensions (e.g., Claridge, 1986; Gray, 1982), but it remains an initial attempt to link behavioural dimensions to underlying biological processes. In 1976, Eysenck added a third dimension, psychoticism (P), defined as tenderminded versus tough-minded (high P is similar to psychopathy) (Eysenck et al., 1976).
Costa and McCrae's (1992) five-factor model consists of five dimensions—neuroticism, extraversion (both similar to Eysenck's dimensions), agreeableness (trust, altruism), conscientiousness (selfdiscipline, competence, order) and openness to experience (aesthetics, fantasy, values). Although this model is being increasingly related to the DSM PD classification (Nestadt et al., 1994; Widiger and Costa, 1994), it has only recently been linked with biological processes, and then mainly genetics.
More recently, a model of temperament with specific theoretical relationships to neurobiology was proposed by Cloninger (1986; 1987). This model began by postulating brain systems based on animal and genetic studies and then worked out their behavioural manifestations (Carey and DiLalla, 1994). There are three temperament dimensions, namely, novelty seeking (NS), harm avoidance (HA) and reward dependence (RD). Initially, NS was hypothesized to reflect heritable differences in the behavioural activation system and to be associated with central nervous system (CNS) dopaminergic activity, HA was said to reflect individual differences in behavioural inhibition and to be associated with serotonergic neural firing, and RD was held to reflect differences in behavioural maintenance and was associated with noradrenergic activity. More recently, Cloninger's theory has been modified, so that HA is related to gamma-aminobutyric acid (GABA), as well as serotonin; RD is related to serotonin (median raphe) as well as noradrenaline; and persistence (a new temperament dimension based on a reward-dependence subscale) is related to serotonin (dorsal raphe). The tridimensional model of