Personality traits are elaborations of what once were relatively simple reflexive mechanisms.

“The Twilight Zone” between Genotype and Social Phenotype

The nineteenth-century “science” of phrenology proposed that each personality trait had a particular locus in the brain that shaped the skull above it. Today, we view this kind of brain localization as fallacious. But in the search for simplicity we may be creating a new kind of phrenology, one more in accord with real brain entities and modern neu­rophysiology, but still inaccurate. The new “phrenology” suggests that each personality trait is based on one particular brain structure or system or one biochemical.

Personality traits are elaborations of what once were relatively simple reflexive mechanisms. A simple organism like a paramecium has two basic “personality traits”: approach and withdrawal. While these are largely a function of external stimuli they are conditionable, and one could conceive of individual differences in the traits based on variations in the stimulus-response mechanisms or “life experiences.” Analogous ten­dencies like impulsivity and anxiety traits in higher organisms may have origins in early evolution, but the mechanisms mediating them have become quite complex. Anxiety, for instance, requires “anticipation,” including abstraction from the common elements of past situations of danger, storage of them in memory, their control over behavior in the appropriate circumstances, and a rapid shift in the physiology of the body toward the demands of defense and survival. Any personality trait involves a variety of behavioral mechanisms and each behavioral mechanism is likely to be mediated by a number of biological mechanisms. It is these biological mechanisms that are most directly under the control of the genotype through its assembling of chemical components.

Given this kind of complexity in brain behavior relationships, a “top down” rather than a “bottom up” approach to defining basic personality traits may be best. A ”top down” approach (e.g., Eysenck) is (1) definition of personality dimensions at the highest or broadest level; (2) delineation of narrower traits composing them; (3) identification of behavioral mechanisms involved in the personality traits; and (4) finding the biological mechanisms controlling the behavioral ones and thereby the personality traits. A “bot­tom up” approach (e.g., Gray) starts with the biological bases of behavioral mecha­nisms, as developed from comparative studies of other species, followed by extrapola­tion to the behavior and personality traits of humans. Theoretically, both approaches could yield the same isomorphic solution. Because of the problem of finding the appropriate animal models for human traits, it is more likely that different solutions will be reached by those starting from the top and those working up from the bottom. Gray (in press), for instance, has attempted to redefine the basic dimensions of personality within Eysenck’s (Eysenck, 1967; Eysenck & Eysenck, 1985) model by drawing the axes at peculiar angles through the three-dimension space without regard for the empirically derived structure of personality traits as revealed in factor-analytic studies. While Gray has identified basic biobehavioral mechanisms in rats, their involvement in human personality traits is open to question. What will be described in this chapter is a more complex hypothetical relationship between basic personality traits and their biological substrates at several levels between the genotype and the social phenotype as shown in Figure 8.1.

Personality Traits

Personality trait dimensions, based on self-report questionnaires or ratings by oth­ers, represent the most abstract level of description. A recent factor analysis of question­naire scales (Zuckerman, Kuhlman, & Camac, 1988) has shown that the three broad personality dimensions postulated by Eysenck (1967) can be identified in both men and women. Eyscnck’s scales, Extraversion (E), Neuroticism (N), and Psychoticism (P) are good markers for these dimensions. Data in this study and a more recent one suggest that Aggression (Agg) and Hostility may constitute an equally reliable fourth factor, located between the P and N dimensions, although somewhat closer to N than to P.  Activity, suggested by Buss and Plomin (1984) as a primary dimension of temperament, particu­larly in children, can also be reliably identified in a five-factor analysis; but at higher levels it tends to divide up into the other factors. The major component of the E factor is consistently Sociability. The N factor at the broadest level consists of scales measuring negative affectivity: anxiety, hostility, and anger. Although markers for trait depression were not included they would probably fall into this factor. While the factor is labeled General Emotionality, it should be understood that this pertains primarily to dysphoric emotions, not to positive affect. The latter emotional trait is primarily related to the E dimension, and when it is related to the N dimension, the relationship is an inverse one. “Personality in the third dimension” (Zuckerman, 1989) or P is more complex in terms of its constituent personality traits.

I have argued that Eysenck’s label “Psychoticism” is not an accurate description of the trait. If a clinical term must be used, “Psychopathy” or “Antisocial Personality” would be more appropriate, since this disorder incorporates the traits and many of the biological constituents of the dimension better than psychotic disorders like schizo­phrenia. I have called the dimension “Impulsive, Unsocialized Sensation-Seeking” (ImpUSS) to summarize the narrower traits involved. Impulsivity can be distinguished from the other traits in the     P-ImpUSS dimension in factor analyses rotating six or seven factors, but at these levels there is less reliability of the factors across gender or samples. The clustering of narrower factors within broader ones does not mean that there is no point in assessing narrow as well as broad ones because some of them seem to be more closely related to behavior mechanisms or biological factors, as will be shown.

 Figure 8.1. A psychobiological model for personality. Dopamine refers particularly to the A10 dopaminergic pathway from the ventral tegmental area to the nucleus accumbens via the medial forebrain bundle and the A9 pathway from the substantia nigra to the caudate putamen. Low levels of type B monoamine oxidase (MAO) may deregulate these systems. High levels of gonadal hormones, particular­ly testosterone, may furnish a basis for both sociability and disinhibition. High levels of serotonin in conjunction with high levels of both type A and B MAO may provide the basis for strong inhibition; low levels of serotonin together with high activity of dopaminergic systems may be involved in disinhibition, impulsivity, and aggression or hostility. Regions of the septal area are particularly involved in inhibition­-disinhibition of behavior. Norepinephrine (Norepi), particularly in the dorsal ascending noradrenergic pathways from the locus coeruleus, is also involved in the adrenergic arousal found in both anxiety and anger. Low levels of norepinephrine, perhaps related to low levels of the enzyme dopamine-beta-hydroxylase (DBH), may be involved in the traits of disinhibition and impulsivity. Stimulation from the central nucleus of the amygdala to the ventral tegmental areas and the locus coeruleus may increase activity in dopaminergic and noradrenergic systems. At low levels this catecholamine system activity may be rewarding and facilitating, but at high levels may be associated with anxiety, distractibility, inhibition, and adrenergic arousal. When adrenergic arousal is combined with high activity of ben­zodiazepine receptor inverse agonists and low levels of GABA inhibition, the result may be anxiety. Specific combinations of these biological traits may underlie the disposition of trait anxiety and emo­tionality in general.

Cognitive Affective and Behavioral Mechanisms

The model I am proposing would characterize the mechanisms underlying the E and N dimensions in cognitive terms rather than conditioning ones. Extraverts can be charac­terized in terms of a strong “Generalized Reward Expectancy” and introverts in terms of a weaker one. The normal (average N and P) extravert is generally an optimist, with high self-esteem and self-efficacy beliefs, particularly in search of social reinforcements. Neurotics or high anxiety-trait persons can be characterized in terms of a strong “Gener­alized Punishment Expectancy.” Generalized anxiety is associated with a cognitive component (worry) which involves excessive apprehension about possible negative out­comes and a feeling low efficacy (helplessness) in coping with stress. Recent data correlating scales of “Generalized Reward and Punishment Expectancies” (GRAPES) with the E and N scales from the Eysenck Personality Questionnaire (EPQ) tend to support these hypotheses (Ball & Zuckerman, 1990).

The basic mechanism underlying the P dimension is hypothesized to be one of Disinhibition vs. Inhibition. The combination of sensation-seeking (incautious risk ­taking in pursuit of reward), impulsivity (inability to restrain behavior even where it might lead to punishment) and a lack of socialization (need or desire to follow the rules and abide by the values of society) all suggest a deficit in inhibition and a low threshold fur disinhibition. The defect in behavioral restraint may operate at a brain level that short-circuits cognitive analyses. The impulsive and incautious person typically thinks after acting, rather than before. The characteristic learning problem in psychopaths is one of learning when not to act (passive avoidance), rather than learning when to act (Newman & Kosson, 1986). Their failure to learn from punishment experience may reflect a classical conditioning deficit when the unconditioned stimulus is aversive (Lykken, 1957; Hemming, 1981). Originally, Eysenck (1965) proposed a deficiency in general “conditionability” as the basis for extraversion. At that time extraversion was conceived of and assessed as a dual component trait including sociability and im­pulsivity. Eyelid-conditioning studies (Barratt, 1971; Eysenck & Levey, 1972) showed that conditioning was only related to the impulsivity component of E, particularly to a measure of narrow impulsivity (Frcka & Martin, 1987). However, the introduction of the P dimension into the personality instrument resulted in a dropping of impulsivity type items from the E scale and the closer alignment of impulsivity with P. The P scale itself predicted conditioning in a study where paraorbital shock was used instead of an air puff as the unconditioned stimulus (Bytes, Frcka, Martin, & Levey, 1983). Perhaps the more aversive nature of the UCS provided clearer evidence of the influence of P than an earlier study (Frcka, Beytes, Levey, & Martin, 1983). Impulsive and high P individuals seem to be poor conditioners in response to aversive UCSs. This would mean that they would have a difficult time in learning inhibition from physical punishment of the type that is often used in an attempt to discipline them.

Tellegen (1985) has suggested that basic dimensions of personality are strongly related to independent dimensions of positive and negative affect. Meyer and Shack (1989) have shown that both trait and state positive affect are aligned with extraversion, while negative affect items fall on a dimension defined by neuroticism. As shown in Figure l, the present model accepts this idea that there is a tendency of sociable persons to experience frequent states of positive affect which may account for their high reward expectancy. Given the same positive reinforcement, reward effects (positive affect) would be stronger in the extravert than in the introvert. The punishment expectancy in neurotics would be based on their more frequent experience of states of negative affect. One could cogently argue that the generalized expectancies affect the occurrence of positive or negative affects more than the other way around. Affect and expectancy undoubtedly influence each other. However, emotions stem from more basic and earlier evolved parts of the nervous system than cognitions. Emotional discriminations may be made on the basis of partial, incomplete stimulus processing. While cognitions may modify and direct emotional responses, the initial emotional reactivity occurs at a lower level of the nervous system (Le Doux, 1987). As Zajonc (l980) says, “Preferences need no inferences.” This question will be addressed again when I discuss a neuropsychologi­cal basis fur the N dimension.


Most of the research with humans on the biological basis of personality has come from the field of psychophysiology, primarily because these non-invasive bioelectrical recording techniques are more accessible to psychologists than biochemical methods. Much of the research has been based on arousal theories of personality (Strelau & Eysenck, 1987) which have survived and even flourished despite criticisms. Eyscnck’s (1967) theory postulates that arousal and arousability of the reticulocortical activation system are the biological basis of extraversion and introversion.

The most relevant evidence for a cortical arousal theory of E-I comes from EEG studies since these provide the most direct psychophysiological measure of cortical arousal. The results from these studies have been equivocal (Gale & Edwards, 1986; O’Gorman, 1984). There are many methodological problems in this type of research, like defining the conditions in which to measure tonic arousal. But even when the analysis of the literature is limited to the most conceptually and methodologically sound studies, as agreed on by Gale and O’Gorman, the results are still inconclusive (Zucker­man, 1991).

Recent evidence suggests that the cortical arousal hypothesis may be relevant for some traits involved in the P dimension rather then the E dimension. O’Gorman and Lloyd (1987) used the EPQ-E scale, which unlike previous versions of E is largely devoid of impulsivity items, and the Eysenck and Eysenck’s (1977) broad Impulsivity scale. They measured EEG in two conditions, one of which was suggested by Gale to be optimal for revealing differences between introverts and extroverts. While E and broad impulsivity measures were not related to EEG arousal, narrow impulsivity was related: impulsives were less aroused. Narrow impulsivity (responding quickly without restraint) was also the type related to eyelid conditioning, as discussed previously. Since arousal may be a primary basis for conditionability, the findings are consistent, but the emphasis must be shifted from E, as currently defined, to a specific component of the P dimen­sion. Goldring and Richards (1985) also report that the P scale itself is related to low cortical arousal.

 Eysenck’s theory suggests an interaction between stimulus intensity and cortical arousability due to the sensitivity of introverts to stimulation in the lower range of intensity and their transmarginal (cortical) inhibition in response to high intensity stim­uli. The cortical evoked potential (EP) augmenting-reducing paradigm developed by Buchsbaum and Silverman (1968) provides an ideal way of testing this hypothesis, since it measures the cortical responsivity at each of several stimulus intensities covering a range of intensity. Augmenting describes the tendency for a linear increase in EP ampli­tude with increasing stimulus intensity, while reducing refers to either the lack of increase or a reduction of EP amplitude at the highest stimulus intensities.

No relationship has been found between extraversion and EP augmenting-reducing, but a very robust relationship has been found between the Disinhibition sensation-­seeking scale and augmenting of visual and auditory EPs. This literature involving 15 studies has been recently summarized (Zuckerman, 1990). In 5 of 7 analyses of the visual EP and 7 of 9 analyses of the auditory EP there was a significant relationship between at least one of the sensation-seeking scales (usually Disinhibition) and augment­ing: high disinhibitors tend to be augmenters, while low disinhibitors tend to be re­ducers. Barratt, Pritchard, Faulk, and Brandt (1987) found a similar relationship be­tween impulsivity, particularly “cognitive impulsiveness,” and augmenting.

The augmenting-reducing source of individual differences has also been demon­strated in cats (Hall, Rappaport, Hopkins, Griffin, & Silverman, 1970; Lukas & Siegel, 1977; Saxton, Siegel, & Lukas, 1987) where it has been related to natural behavior characteristics. Augmenter cats tend to be active and exploratory while reducers are inhibited and tend to withdraw from novel stimuli. The augmenter cats performed poorly on an experimental task requiring the animal to delay or inhibit response in order to get reward, despite the fact that they performed better than reducers on a simple fixed interval reward schedule. The pattern of performance seen in augmenter cats is also one seen in septal-lesioned rats and Gorenstein and Newman (1980) have suggested a paral­lel between the behavior of these rats and the “disinhibitory psychopathology” in hu­mans, particularly in the antisocial personality. Another part of the pattern in septal-­lesioned rats is enhanced “stimulus-seeking” behavior.

The evidence strongly suggests that the cortical augmenting-reducing paradigm is a marker for at least some of the traits in the P dimension, particularly disinhibition and impulsivity. Reducing represents the capacity for behavioral inhibition and augmenting is associated with a deficit in this capacity. This interpretation is consistent with the clinical correlates of EP augmenting including alcoholism, delinquency, drug use, and bipolar disorders (Zuckerman, Buchsbaum, & Murphy, 1980).

A particularly challenging line of research has been conducted by Pivik, Stelmack, and Bylsma (1988). They measured excitability of a spinal motoneuronal reflex with stimulation applied to the leg. High scorers on both E and disinhibition scale showed reduced motoneuronal excitability as assessed by reflex recovery functions. The arousal hypothesis has been centered on the cortex, but there is a possibility that it might apply to neurons at subcortical levels as well. While the functional significance of reduced motoneuronal excitability is not clear, it has been associated with increased dopaminergic function, suggesting a possible link with the biochemical level.

Autonomic Arousal and Arousability


Eysenck (1967) hypothesized that the N dimension is biologically based on the limbic brain which regulates emotionality and the peripheral adjustments of the autono­mic nervous system in reaction to stress. Adrenergic arousal is reflected in a variety of measurable psychophysiological changes such as heart rate, blood pressure, peripheral vasoconstriction, respiration rate, and skin conductance fluctuations and level. How­ever, must studies have not shown a relationship between levels of adrenergic phys­iological activity and N in normal populations (e.g., Fahrenberg, 1987). Eysenck and Eysenck (1985) acknowledged the general failure of the hypothesis linking N with sympathetic-autonomic arousal, but said that this might be due to the fact that most studies did not expose subjects to stress or aversive stimuli. However, Fahrenberg (1987) describes large-scale studies where physiological measures were taken during resting, basal conditions, and during a variety of stressor situations, including physical stress (blood taking, cold-pressor tests) and social stress (interview, performance). While the stressors were effective in increasing physiological responsivity in all subjects, there was no evidence of a relationship with N.

However, a survey of studies (Zuckerman, 1991) comparing controls and patients shows that all groups of anxiety disorders, except those with simple phobias, have higher levels of basal heart rate and skin conductance fluctuations than normals. Physiological reactivity in response to general kinds of stress does not differentiate normals from anxiety patients; in fact, normals tend to show greater response to stress situations simply because they start at a lower level than patients. However, when phobic patients are exposed to objects or situations which normally elicit their fears, they do show heart rate reactions that are greater than those of normals exposed to the same stimuli. Unlike patients with anxiety disorders, high N subjects from the normal popula­tion cannot be characterized by a general dysregulation of the adrenergic arousal sys­tems. Like the phobic patients among the anxiety disorders, their anxiety response may be only to specific kinds of stressors or persons, or the trait of neuroticism in normals may be more related to cognitive mechanisms than to arousal ones.


Catecholamines (Dopamine, Norepinephrine, and Epinephrine)

Many theorists have suggested that the neurotransmitter dopamine is the basis of some kinds of general motivational trait involving exploration directed toward primary rewards in animals (Gray, in press; Stein, 1978) and novelty and sensation seeking in humans (Cloninger, 1987; Zuckerman, 1979). Moderate doses of dopamine agonists, like stimulant drugs, increase social behavior and activity; higher doses may reverse these effects (Zuckerman, 1984). Dopamine is vital to the intrinsic reward effects pro­duced by self-stimulation or self-infusion of stimulant drugs (Bozarth, 1987). When dopamine is depleted, as in Parkinsonism, the result is an anhedonic, apathetic personality, not interested in the environment and lacking in positive emotionality or joy. All of the effects of dopamine depletion or release suggest that it must be involved in the positive affect, high activity, and sociability of extraversion.

However, there are also suggestions of a link to sensation-seeking and impulsivity found in the P dimension. Drug abusers tend to show high levels of sensation-seeking and antisocial tendencies (Zuckerman, 1987) and many of the drugs abused act through the dopaminergic systems, stimulants like cocaine and amphetamine having their pri­mary effects on the nucleus accumbens, while opiate reward is mediated in the ventral tegmental area (Bozarth, 1987). Perhaps the non-drug using extravert has a high level of tonic activity in one or more of the dopamine systems, while the disinhibiting or boredom susceptible individual has a low level and therefore is particularly attracted to drugs or exciting activities that act on brain dopamine systems. But the only empirical correlations found thus far are between norepinephrine (in cerebrospinal fluid) and plasma dopamine-beta-hydroxylase (the enzyme which converts dopamine to epine­phrine in the neuron) and sensation-seeking. Both correlations are negative, suggesting that high sensation-seekers have low levels of both the neurotransmitter and the enzyme involved in its production.

While adrenergic arousal may be an essential component of clinical anxiety, it is not the entire story. Adrenergic arousal can be pleasurable when it occurs at an optimal level in sensation-seeking or sexual activities, neutral when it occurs in physical exercise, or displeasurable when it occurs during a panic attack. Drugs, like yohimbine, that stimu­late activity of the norepinephrine system in the brain, produce anxiety and panic attacks in persons who already have these disorders. But other drugs like lactate and caffeine, which do not stimulate catecholamine activity, also produce anxiety in these patients (Gorman, Fyer, Liebowitz, & Klein, 1987). All of these drugs do not produce panic and major anxiety in most normals, even though they do increase their physiological arousal.

The common denominator of all drugs which have an anxiogenic effect is that they produce peripheral sympathetic nervous system effects such as tachycardia. However, such arousal is not intrinsically associated with the subjective dysphoria characterizing anxiety. Perhaps recurrent arousal may result in the internal sensations of arousal becoming conditioned cues for the full panic or anxiety attacks. This would produce a positive feedback in which apprehension of arousal would increase arousal.

Benzodiazepine-GABA System

Something else must dispose the person to perceive internal arousal as a sign of threat. Is there a particular mechanism for the emotion of fear or anxiety, as distin­guished from general emotionality? The benzodiazepines seem to reduce the subjective sense of anxiety without the generalized, intensive, sedative effects of barbiturates or alcohol. They act on recently discovered receptors in the brain called “benzodiazepine receptors.” These receptors work by potentiating the effects of GABA, an inhibitory neurotransmitter widely distributed in the nervous system. However, the ben­zodiazepines (BZs) do not work by general sedation, so the GABA effects must be specific to certain pathways. These will be discussed in a subsequent section. The very existence of the BZ receptors suggests that there must be natural receptor agonists (which would dampen anxiety) or inverse agonists (which would be anxiogenic). A natural polypeptide produced from rat brain, called diazepam-binding inhibitor (DBI) has an affinity for the BZ receptor and also facilitates suppression of behavior in a conflict situation (Guidotti, Forchetti, Corda, Konkel, Bennett, & Costa, 1983). Beta­carbolines produced in the laboratory, but with a natural affinity for benzodiazepine receptors, have been shown to produce “apprehension” in normal subjects (Dorow, Duka, Holler, & Sauerbrey, 1987).

These early studies suggest that a balance between natural BZ receptor agonists and inverse agonists, when combined with catecholamine-mediated arousal may produce the full-blown phenomenon of anxiety. An inverse agonist like DBI could be what “tags” arousal as “fear.” Another possibility is that the number and distribution of BZ receptors may be what underlies the vulnerability to anxiety or N trait. Decreased concentrations of BZ receptors have been found in an anxious strain of mice and BZ-binding is higher in an emotionally nonreactive strain of rats than in one characterized by high emotionality (Robertson, Martin, & Candy, 1978). Whatever their precise role in emotionality in general or anxiety in particular, endogenous biochemicals acting on the BZ receptor sites are likely to playa crucial role in the generalized apprehensiveness characterizing the N dimension of personality.



The case for serotonin as a mediator of anxiety (Cloninger, 1987; Gray, 1982, in press) is largely based on its role in inhibition of approach behavior in conflict situations (Soubrie, 1986) or emotional systems in general (Panksepp, 1982). The case for anxioly­tic effects of serotoninergic drugs is far less conclusive (File, 1988). As a matter of fact, the comparative and human clinical literature suggests that it is low levels of serotonin that are related to anxiety and more so to depression. But serotonin is primarily corre­lated with the P dimension, impulsivity and aggressiveness. Persons with low levels of the serotonin metabolite 5 hydruxyindoleacetic acid (5-HIAA) tend to score high on Eysenck’s P scale (Schalling, Asberg, & Edman, 1984), and on hostility and psychopa­thy scales (Brown, Ebert, Goyer, Jimerson, Klein, Bunney, & Goodwin, 1982). Such persons also are found among those personality disorders who are behaviorally aggres­sive (Brown, Goodwin, Ballenger, Goyer, & Major, 1979) in contrast to more passive forms of disorders. Low levels of 5-HIAA are found in parsons who have attempted or committed suicide in impulsive, violent ways, and in impulsive murderers (Van Praag, 1986). Van Praag, Kahn, Asnis, Wetzler, Brown, Bleich, and Korn (1987) have sug­gested that low 5-HIAA is more indicative of aggressive disregulation than depression. The human data suggest that low serotonin is related primarily to the disinhibition of behavioral impulse associated with the P dimension and only secondarily to anxiety and depression characteristic of the N dimension. These findings are consistent with the animal data suggesting that serotonin regulates impulsive behavior associated with the possibility of punishment.


Eysenck (1967) proposed that testosterone may be involved in the p dimension of personality, largely on the basis of the human sex difference in violent aggressiveness and the well-demonstrated association of aggressiveness with testosterone in other species. Studies have demonstrated a direct relationship between testosterone in male, and the sensation-seeking trait of Disinhibition (Daitzman & Zuckerman, 1980; Daitzman, Zuckerman, Sammlelwitz, & Ganjam, 1978), and Monotony Avoidance (Schalling, 1987). Testosterone also correlates with social extraversion (Daitzman & Zuckerman, 1980; Schalling, 1987).

The evidence from both normals and prisoner samples suggest that testosterone in both sexes is related to dominant sociability and interest in sex as well as sexual experience. In the prisoner samples, testosterone also seems to be associated with a high degree of unprovoked violence (Dabbs, Ruback, Frady, Hopper, & Sgoutas, 1988; Ehrenkrantz, Bliss, & Sheard, 1974; Mattson, Schalling, Olweus, Low, & Svensson, 1980; Rada, Laws, & Kellner, 1876). But in the normal population the association with aggressiveness of this type is not found; instead testosterone is associated with both the E and P dimensions, the latter through sensation-seeking and the capability for normal aggressiveness as a defensive reaction in adolescent boys (Olweus, 1987).

Monoamine Oxidase (MAO)

While the implications of platelet measures MAO for activity in the three mono­amine systems are not certain, its relationships to at least two major dimensions of personality are clear from a wealth of correlational data. Low MAO levels have been related to high levels of general activity in neonates during the first three days of life (Sostek, Sostek, Murphy, Martin, & Born, 1981), and high levels of social activity in adult humans (Coursey, Buchsbaum, & Murphy, 1979), and general and social activity in colony-dwelling monkeys (Redmond, Murphy, & Baulu, 1979). MAO has been found to be negatively correlated with extraversion or positively correlated with introversion scales in several studies, but not in some others. If we weigh the behavioral data more highly than the questionnaire findings, there does seem to be a relationship between MAO and the E or sociability dimension.

There also appears to be a relationship between MAO and some traits within the P dimension. In the Coursey et al. (1979) study contrasting high and low MAO types in the normal population, the low MAO group reported more convictions for criminal offenses and more alcohol and drug use. The low-MAO male monkeys in the Redmond et al. (1979) study engaged in more aggressive and sexual activity than the high MAO ones. General sensation-seeking and Monotony Avoidance scales have been found to be negatively correlated with MAO in males in a number of studies (see Zuckerman, 1987, for review). While the results are not always significant, and the correlations tend to be low, the total pattern confirms an inverse relationship between MAO and sensation-seeking. To these personality trait findings we may add the fact that low MAO levels are also found in alcoholics and chronic marijuana users.

While the relation of MAO to sensation-seeking in particular and the E and P dimensions in general is clear, the mechanism is not. MAO is not a direct behavioral inhibitor or activator but only affects behavior through its effects on the monoamine systems. All we can infer is that the monoamine systems are somehow involved in the biological substrates for personality. Perhaps the rule of MAO is one of regulating or stabilizing these systems in response to environmental stimulation. Bipolars, who already have low levels of MAO even in the depressed state, tend to shift to the manic state when given monoamine oxidase inhibitors, perhaps due to a buildup of dopamine with insufficient MAO to metabolize the neurotransmitters accumulating in the neurons. Dopamine-beta-hydroxlase (DBH) is an enzyme involved in the conversion of dopamine to norepinephrine in the neuron. Plasma DBH has been found to be negatively related to sensation-seeking in several studies. Low DBH has been associated with severe psychopathic disorder in alcoholics (Major, Lerner, Goodwin, Ballenger, Brown, & Lovenberg, 1980) and emotionally disturbed boys (Rogeness, 1984). The relationship of DBH to the P dimension must be mediated through its limiting effect on production of norepinephrine. Perhaps there is a link with the low adrenergic levels which are predic­tive of adult criminality and aggressiveness (Olweus, 1987).


The designation of particular brain structures as the locus of personality traits is a dangerous flirtation with the type of thinking that produced 19th century phrenology. A structure like the amygdala contains many types of neurotransmitters, and different nuclei within the discernible structure mediate different functions.

Reward and Activity

On the assumption that the trait of extraversion is specifically associated with reward sensitivity or expectancy and activity, three dopamine systems in the brain are likely candidates to be involved with this trait, as well as having some involvement in sensation-seeking (Zuckerman, 1979). The A l0 system originates in the ventral tegmen­tum and projects to the nucleus accumbens via the medial forebrain bundle (MFB). About 85% of the projections from ventral tegmen­tum to accumbens are dopaminergic (Stellar & Stellar, 1985). The MFB is a highly active site of self-stimulation reward effect, and the nucleus accumbens is the primary site of action for reward by stimulant drugs (Bozarth, 1987). Another dopamine system originates in the subtantia nigra (A9) and projects to the neostriatum, caudate nucleus, and putamen. This system is necessary fur regulation of activity and is the one severally damaged in Parkinson’s disease. It is also largely dopaminergic. The subtantia nigra and caudate also support brain stimula­tion (Stellar & Stellar, 1985). Projections from both systems reach the lateral and medial prefrontal cortex as well as limbic areas such as the septum and amygdala. Since the Al0 system is vital in reward and the A9 in motivated activity, individual differences in their physiology could very well be tile source for the activity and search for reward typical of extraversion and sensation-seeking.

Behavioral Inhibition

Gray suggested that the core of anxiety is a “behavioral inhibition system” (BIS) in which the underlying neurological substrate is the septohippocampal system. The func­tion of the BIS is to check incoming stimuli against the memory of the previous experi­ence with those stimuli. If the stimuli are novel or associated with past punishment, the system is activated producing arousal, inhibition of ongoing behavior, and orienting (diversion of attention) to the stimulus.

While inhibition is an immediate reaction involved in anxiety, it is also involved in other kinds of activity including approach behavior. Orienting to novel stimuli is not necessarily associated with anxiety. Strong orienting responses (ORs) have been positively associated with sensation-seeking and state anxiety seems to dampen ORs to neutral but novel stimuli (Zuckerman, 1990). The inhibition of behavior in at) approach­-avoidance conflict situation may be a function of anxiety, but it is also a function of the disinhibition tendency postulated to be the core of the P dimension. This is why Gray (in press) regards anxiety and psychopathy traits as the two ends of a bipolar dimension of personality.

The septohippocampal system may be more relevant for the disinhibition vs. inhibi­tion mechanism than for an anxiety mechanism and therefore more related to the P dimension than to the N dimension of personality. The inhibition mechanism, triggered by signals of punishment, would be weakened in persons with high P traits like im­pulsivity and sensation-seeking. Serotonin pathways may be the main ones involved in inhibition of behavioral approach.


Where then should we look for a locus for the system underlying general emo­tionality or N? The amygdala seems to be at the center of such a system. The amygdala has been called the “sensory gateway to the emotions” by Aggleton and Mishkin (1986). This term is used because this structure serves as the central target for converging inputs from several cortical processing areas involving all of the sensory modalities. The olfactory input is even more direct, reflecting the early evolutionary control of emotional and behavioral response from this modality. LeDoux (1987) points out that there are direct pathways between the thalamus and amygdala which would allow emotional reactions to stimuli before they are fully processed by higher centers in the hippocampus and cortex. The amygdala seems to serve as a comparator, as does the hippocampus, but probably begins the process at an earlier stage than the hippocampus. Human anxiety of the panic and generalized type is often triggered by unknown stimuli. The amygdala may respond to partial cues that can not be identified in consciousness, perhaps accounting for its important rule in classical fear-conditioning.

The temporal lobe is a major source of input to the amygdala. Reiman, Raichie, Robins, Mintun, Fusselman, Fox, Price, and Heikman (1989) used positron emission topography on patients with panic disorders before and after lactate infusion. Those who panicked showed significant increases in regional blood-flow in the bilateral tem­peropolar cortex as well as deeper limbic structures. The original Kluver-Bucy (1939) effect of removal of the temporal lobes was produced by damage to the amygdala lying within them. The syndrome was one of “psychic blindness”; animals could perceive stimuli but seemed ignorant of their emotional significance. The operated animals were also usually tame (no fear of handlers) and did not show fear of snakes. The amygdala receives such information from the inferotemporal cortex via the entorhinal cortex, a region of limbic cortex that is the major source of input into the hippocampus. Many of the input sources for the amygdala are also sources of input for the hippucampus. LeDoux’s (1987) view is that the hippocampus mediates the more cognitive aspects of emotions transmitting thoughts or memories to the amygdala for reappraisal of emotion­al significance. But is this role of the hippocampus in emotions a primary one, or just one of its memory-related functions? The amygdala seems to be more central to an emotion-generating system, and therefore a likely basis for the dimension of personality based on emotionality.

Behavior Genetics

The direction of this chapter has been downward from the social trait through the different levels of the biotypes to the genotypes. The usual question asked by behavior geneticists is the extent to which a behavioral or biological phenomenon is determined by heredity and to what extent by environment. Questions asked by the more sophisti­cated analyses concern the kinds of genetic mechanisms and environmental factors involved. Given that we do not inherit personality traits as such, what is it (the biological characteristic) that we do inherit that influences them?

Personality Traits

Various large-scale studies of identical twins (ITs) and fraternal twins (FTs) reared together, come involving thousands of each type, have shown a fair uniformity of results. For most broad traits, such as E (Extraversion), N (Introversion), and P (Psychotocism), the estimate of heritablity range from 40% to 60% with the typical figure around 50%. Usually the correlation for ITs is about .50, while that for FTs ranges from ) to .3. For E there is some evidence of nonadditive genetic factors, as indicated by very low FT correlations. The heritability is the same for the N and P dimensions traits, but the genetic mechanism is purely additive. The results suggest little effect of shared environment; the main environmental effects seem to be specific