The Psychopath’s Choice: Investigating the Neural Correlates of
Decision-making in Psychopaths.
Sebastian Speer
Master Brain and Cognitive Science, University of Amsterdam
Abstract
Psychopathy can be conceptualized as a decision-making disorder (Koenigs & Newman, 2013). It is assumed that impairments in brain areas involved in decision-making cause impulsivity, risk-taking and antisocial behavior frequently observed in psychopaths. However, despite decades of research, the underlying mechanisms of the decision-making deficits in psychopathy remain controversial and the findings across studies are inconsistent. The current review combines fMRI studies on psychopathy and decision-making research to investigate the underlying mechanisms and neural correlates of decision-making deficits in psychopaths. Integrating insights from cognitive neuroscience and forensic psychology, it is argued that dysfunctional response
modulation associated with structural abnormalities, aberrant functional activity and connectivity in prefrontal and paralimbic areas underlies deficient decision-making in psychopathy. Deficits in attention allocation result in dysfunctional avoidance learning, impaired reversal learning and disadvantageous behavior in decision-making tasks such as the Ultimatum Game and the Prisoners Dilemma. Furthermore, it is argued that inconsistency in decision-making research is due to several methodological concerns, including lack of consideration of subtypes of
psychopathy, small sample sizes and different conceptualizations of psychopathy. Future research may benefit from further investigating neurological differences between subtypes and applying more advanced neuroimaging techniques.
Keywords
Psychopathy, Decision-making, Response modulation, Neuroimaging, Subtypes, OFC, Amygdala, Functional Connectivity
1. Introduction
Psychopathy is among the most difficult disorders to identify and there are few disorders more malignant. Superficially, the psychopath appears normal, confident and even charming. On closer inspection, however, it becomes clear that he lacks conscience and empathy. Psychopaths are emotionally detached and often engage in manipulative, exploitative and often criminal behaviour (Hare, 1991). They commit a disproportionate amount of violent crimes and consequently place a considerable economic and emotional burden on society.
Psychopathy is a personality disorder defined by the failure to experience many emotions that are essential for appropriate social behaviour (R. D. Hare & Neumann, 2008). The disorder has been associated with impairments in decision-making, impulsivity and empathy. It essentially involves two components, namely emotional dysfunction and antisocial behaviour (Hare, 1991; Harpur, Hakstian, & Hare, 1988). The first component or factor represents interpersonal and affective traits and is expressed in reduced empathy, remorse, guilt and a lack of close
relationships to others. This first factor can be subdivided into two facets, namely interpersonal and affective traits. Psychopaths are often disinterested in and fail to establish long lasting bonds, which represents the interpersonal traits (Facet 1) (Hare, 1991). Further, psychopaths are often described as emotionally detached and callous, which corresponds to the affective traits (Facet 2). Secondly, the antisocial behaviour component (Factor 2) reflects a predisposition to antisocial behaviour from an early age. This is expressed by impulsivity, irresponsibility, lack of
behavioural control and promiscuity (Hare, 1988). Factor 2 is also often divided in lifestyle traits (e.g. impulsivity, irresponsibility & sensation seeking; Facet 3) and antisocial traits (Facet 4). As a consequence, psychopaths frequently violate social norms and cause significant financial, emotional or even physical harm to others. They are characterized by engaging in selfish and manipulative behaviour to achieve their goals (Hare, 1978).
It has been proposed that psychopathy is essentially a decision-making disorder (Koenigs & Newman, 2013). This seems intuitive as a psychopath’s decision-making clearly differs from the norm, in terms of impulsivity and lack of consideration of the consequences. Psychopathy research has had a long history. For decades researchers have attempted to identify the underlying psychobiological mechanisms that may cause the psychopath’s deviant behavior. Until now, however, most of the studies remain inconclusive (M Koenigs, Baskin-Sommers, Zeier, & Newman, 2011) and only recently more consistent neural correlates of psychopathy have been
found (Wolf et al. 2015). Further, there are multiple different theoretical models and approaches, such as the somatic marker hypothesis (Bechara & Damasio, 2005) or the low fear hypothesis (D T Lykken, 1995), none of which have comprehensively explained psychopathy and its inherent decision-making deficit. This may partially be explained by the methodological differences across studies investigating neural underpinnings of psychopathy. The studies differ significantly in the samples used, the definition of psychopathy, how subtypes are considered and in many more domains, which will be discussed in more detail later.
An alternative view has proposed that a psychopath’s decision-making may be beneficial and may lead to competitive advantages in certain contexts (Andrea L. Glenn, Kurzban, & Raine, 2011). A cunning and manipulative person who is emotionally detached and does not care about the feelings of others may more easily act selfishly, which could lead to a significant advantage in contexts like politics, business and even academics. It has been shown that psychopathic
individuals more often choose economic utility (Osumi & Ohira, 2010). Further, individuals in executive positions often exhibit psychopathic characteristics (Babiak, 2010). Hence, when expressed moderately as in the so-called “successful psychopath”(Cleckly, 1951; Hall & Benning, 2006), the disorder may enable a successful and efficient form of decision-making in some contexts.
Investigating the underlying mechanisms of decision-making in psychopathy is of great importance for several disciplines. The callous and antisocial behaviour of psychopaths and their often-violent crimes place a great emotional burden on their social surroundings. Besides that, their criminal tendencies lead to a high conviction rate. While psychopathy affects around 1% of the general population, it is observed in 20% of the prison inmates (Hare, 2003). As a
consequence, psychopaths do not only cause emotional but also considerable economical costs. Therefore, a better understanding and efficient treatment is urgently needed. However, at the present moment there is no known effective form of treatment or therapy for psychopathy (Harris & Rice, 2006). Hence, understanding the neural and psychological mechanisms could be helpful in clinical psychology to develop more effective forms of treatment. Insights into the neural underpinnings of decision-making in psychopathy could be used to develop pharmacological forms of treatment that may target the deficient brain areas. In addition, more knowledge about neural underpinnings may help to distinguish between different subtypes of psychopaths, which may partially explain the inconclusive findings of previous studies. Lastly, psychopathy is among the most important psychological constructs in legal settings as it has strong predictive utility for
recidivism, institutional adjustment and treatment outcomes (Walters, 2003). Advances in
research on decision-making in psychopaths may contribute to solving legal questions such as the culpability of psychopaths, the likelihood of future offense and the prospects of rehabilitation. Therefore, investigating the underlying psychobiological mechanisms of decision-making in neuroscience may benefit several different disciplines such as clinical and forensic psychology, neuroscience and criminology. Advances in these fields can deepen our understanding about the disorder and help reduce the economic and emotional burden psychopaths place on society.
As a consequence, the aim of this thesis will be to investigate the psychobiological
underpinnings of decision-making in psychopaths. In this thesis I will review existing research in psychopathy and decision-making and will critically evaluate the findings of these studies. Furthermore, this thesis will attempt to explain the inconsistency of previous behavioural and neuroimaging research. Finally, it will propose ways in which future research could be improved to avoid inconsistencies and to establish a more coherent way of studying decision-making in psychopathy.
2. The different subtypes of Psychopathy
Although psychopathy has mostly been conceptualized and treated as a unitary construct, accumulative evidence from theoretical and empirical research suggests that there are subtypes of psychopathy that differ significantly from each other (J. L. Skeem, Poythress, Edens, Lilienfeld, & Cale, 2003). One distinction that can be made is the division between primary and secondary psychopathy. Previous research has suggested that primary psychopathy arises from a genetic deficit, which leads to innate dysfunction in basic and affective attentional mechanisms (J. L. Skeem et al., 2003). Secondary psychopathy, in contrast, is proposed to be an acquired affective disturbance, caused by environmental factors such as parental rejection or abuse or
socioeconomic disadvantage (J. Skeem, Johansson, Andershed, Kerr, & Louden, 2007; J. L. Skeem et al., 2003). Several studies have demonstrated that a major difference between primary and secondary psychopaths is their level of anxiety (R. Blackburn, 1975; Hicks, Markon, Patrick, Krueger, & Newman, 2004; D T Lykken, 1957; J. Skeem et al., 2007; J. L. Skeem et al., 2003). Whereas primary psychopaths exhibit a profound lack of anxiety, secondary psychopaths are typically characterized by highly elevated levels of anxiety (J. L. Skeem et al., 2003). Hare’s psychopathy checklist (PCL-R; Hare, 2003), the most used assessment tool for psychopathy, does
not include a measure for anxiety and is uncorrelated with various measures of anxiety (Schmitt & Newman, 1999). As a consequence, psychopathic subtypes cannot be detected by the PCL-R. Secondly, psychopaths also differ in their interpersonal behaviour. Primary psychopaths are characterized as confident, socially dominant, extraverted and sometimes even superficially charming as opposed to secondary psychopaths who are often described as moody, submissive, emotionally disturbed and low in self-confidence (Ronald Blackburn & Coid, 1998).
Moreover, the two subtypes can also be distinguished based on their clinical features; secondary psychopaths are considerably more psychopathological and lower functioning than primary psychopaths (Ronald Blackburn & Coid, 1998). In other words, secondary psychopaths are more impaired and negatively affected by their symptoms than primary psychopaths.
Secondary psychopaths have often been associated with borderline personality disorder and they exhibit many traits that are common for individuals suffering from this disorder (Kane, Fawcett, & Blackburn, 1996). The hostile and callous behaviour of secondary psychopaths can be seen as an emotional adaption to environmental factors such as parental abuse or socioeconomic
disadvantage, which renders them to be considered more amenable to treatment than more inherited deficits in the primary psychopaths (J. L. Skeem et al., 2003).
Notably, some of the studies investigating the differences between psychopathic subtypes suffer from some methodological limitations. The majority of studies apply traditional cluster analytic methods that are known to find clusters where none exist (J. Skeem et al., 2007). For several cluster analysis methods the number of clusters is determined before the analysis is run, which may lead to non-meaningful clusters. Further, the studies are based on heterogeneous samples of which only a small proportion is psychopathic (J. Skeem et al., 2007). As a result, it is not clear to which extent the differences identified are differences between psychopaths, per se. More recent studies tackled these issues by applying model-based cluster analyses on
psychopathic prisoners exclusively. A study by Hicks and colleagues (2004) supported previous findings by identifying two sub-groups of psychopaths with one cluster representing
‘’emotionally stable psychopaths, characterized as low stress reactant, playful, socially dominant and sensation seeking” and a second cluster that was characterized by high stress reaction, high irritability, aggression and disinhibition. The second cluster was labelled as ‘’aggressive
psychopaths’’ (Hicks et al., 2004). The emotionally stable and the aggressive psychopath closely match the previously defined characteristics of primary secondary psychopaths. Another study by Skeem et al. (2007) further confirmed this distinction showing that secondary psychopaths
exhibit higher anxiety levels and manifest more features of borderline personality disorder (i.e., poorer interpersonal functioning).
To sum up, converging evidence consistently identifies two different subtypes of
psychopaths that differ significantly in their personality traits. Underestimation of psychopathic subtypes in research may obscure or completely abolish important results due to high inter-group variability. Therefore, in addition to the PCL-R standardised measures assessing anxiety and emotional stability should be applied when assessing psychopaths in order to distinguish these subtypes.
Besides that, another distinction can be made between the successful and the unsuccessful psychopath. Although criminal activity is often associated with psychopathy, it is not a necessary component of the disorder (Skeem, 2010). The successful psychopath possesses all the core personality traits of a psychopath, but he typically either refrains from traditional criminal activity or avoids conviction (Gao & Raine, 2007). Until now, however, it is unknown whether the successful psychopath and the unsuccessful psychopath differ in their genetic and
neurobiological mechanisms or whether they are determined by environmental factors such as parenting, socioeconomic status and education.
In sum, considering differences between subtypes is highly important as they have profound consequences for risk assessment, management and treatment. If a behaviourally similar group of psychopaths consists of etiologically different subtypes, studies on psychopathy may be confounded and effects may be obscured due to the variability within the group. Thus, the different psychopathy subtypes may partially explain the unreliable and contradictory findings in previous psychopathy research.
3. Psychopathy as an Emotion-processing versus an Attention Deficit - A Brief Review on Influential Theories
3.1 Emotion-Processing Theories
Psychopaths engage in impulsive manipulative and often criminal behaviour without consideration for further consequences (Hare, 1991). This is why psychopathy can be conceptualized as a decision-making disorder. However, there has been a long-standing discussion on whether the deficit in decision-making is due to a lack of emotion or deficient
allocation of attention. In the following paragraph the most relevant theories of both streams of psychopathy research are reviewed.
Theories in the emotion-processing stream of psychopathy research usually emphasize affective deficiencies. One of the most influential theories is the low fear hypothesis by Lykken (1995). The theory proposes that psychopathic traits are due to a below average endowment of innate fearfulness. The lack of experience and learning from fear results in poor fear conditioning and poor passive avoidance learning (D T Lykken, 1957). The lack of fear also complicates socializing psychopathic individuals as many parenting styles heavily rely on learning from responses to punishment. The theory is supported by studies confirming that psychopathic individuals perform poorly on fear conditioning tasks (Birbaumer et al., 2005) and passive avoidance learning tasks (J. P. Newman & Schmitt, 1998) and exhibit lower electrodermal response (in older studies often referred to as galvanic skin response, GSR) in reaction to
conditioned stimuli (D T Lykken, 1957). The electrodermal response is the phenomenon that the skin momentarily becomes a better conductor of electricity when stimuli are perceived that are physiologically arousing (Boucsein, 1992). Arousal is commonly considered to be one of the two main dimensions of emotional response and is thus not the same as measuring emotion but is an essential aspect of it.
The theory, however, does not account for the vast variety of manifestations of restricted affect, such as reduced guilt, embarrassment, love and empathy. It is unlikely that all of these emotions can be explained by a lack of fear. Further, it has been shown that psychopathic individuals show normal affective reactions when instructed to focus on threat relevant cues directly. When emotional stimuli are the primary focus of attention instead of peripheral cues deficits in passive avoidance learning, electrodermal responses to threat cues and fear potentiated startle responses are not observed (Arnett, Smith, & Newman, 1997; Baskin-Sommers, Curtin, & Newman, 2011). Hence, these deficits are specific to certain contexts and task paradigms.
Further, psychopaths exhibit deficits in processing non-emotional stimuli when they are
presented as peripheral cues not directly related to the goal (Newman & Brinkley, 1997), which cannot be explained by the low fear hypothesis.
A more comprehensive theory of emotion in decision-making has been the somatic marker theory (Damasio, 2001). The theory posits that physiological processes, including those that constitute emotions, act as signal to influence decisions and behaviour. Through experience we develop associations between various situations and the corresponding somatic states. If a
certain situation happens again the same neural patterns are triggered, which marks the situation as good or bad. Weak somatic markers as a result of deficient emotional responses may lead to poor decision-making. The advantage of this theory is that it is not limited to fear only, but can explain abnormalities in wider ranges of emotions. However, it also fails to explain the context specificity of deficits and the deficits related to non-affective stimuli.
3.2 Attention-based theories
On the other hand, attention-based models propose that decision-making deficits in psychopathy are caused by a broader information-processing deficit, which is not specific to affective information. These theories posit that psychopathy arises from the inability to flexibly allocate attention away from a dominant goal.
The response modulation theory has been one of the most relevant and clearly delineated attention-based theories (Gorenstein & Newman, 1980; Joseph P. Newman, Schmitt, & Voss, 1997). Response modulation is defined as the temporary suspension of a dominant response set and a brief concurrent shift of attention from the organization and implementation of goal-directed responding to (stimulus) evaluation (Lorenz & Newman, 2002). It refers to the flexible allocation of attention from a dominant response set to attend to unanticipated and peripheral cues (Lorenz & Newman, 2002). It has been suggested that psychopathic deficits in decision-making could be due to dysfunctional response modulation, which leads to the inability to integrate contextual information that contraindicates goal-directed behaviour (J. P. Newman & Schmitt, 1998). Stated differently, secondary stimuli, such as moral conventions, legal
requirements and past experience, cannot be taken into consideration and only primary considerations, such as maximizing immediate reward, or motivations based on immediate concerns guide the decisions. Due to the focus on primary consideration it is highly difficult to resist impulsive behaviour and instead attend to long-term goals. Hence, the response modulation theory proposes that the disinhibition observed in psychopaths is caused by a failure to stop and reflect on the potentially maladaptive nature of their behaviour. Thus, the response modulation theory accounts for the context dependence of psychopathic deficits in emotion processing (Joseph P. Newman et al., 1997). Based on this theory it can be assumed that if stimuli that initiate response evaluation and self-regulation are the focus of attention the decision-making deficits reported in previous studies will not be observed. The theory is supported by empirical
evidence from a word-color Stroop task, which points to a dysfunction in processing or attending to certain types of contextual information (Hiatt, Lorenz, & Newman, 2002). Importantly, these behavioural effects have been observed for non-affective stimuli, which demonstrates that the deficits are not restricted to the emotional domain. To date, however, the underlying neural mechanisms of the response modulation theory have not been identified.
Another highly influential attention based theory is the attention bottleneck model of psychopathy put forward by Baskin-Somers, Curtin and Newman (2011), which argues that dysfunctional response modulation can be explained by an early attention bottleneck. According to this model, abnormalities in early selective attention interfere with processing of information unrelated to the dominant response set. As a consequence, the deficiencies in information processing can be conceptualized as a disorder of early selective attention. Baskin-Somers and colleagues (2011) propose that early attention deficits reduce the scope of attention to the extent that everything except the main focus of attention cannot be processed. A study by Zeier et al. (2009) using a modified version of the Flanker task provides evidence for this theory. They demonstrated that psychopaths exhibited less interference to response incongruent information than non-psychopathic individuals, when attention was cued to the target location and the
response incongruent information was peripheral to the target location. Psychopathic participants displayed normal interference, however, when there was no primary focus of attention (Zeier et al., 2009). Hence, as proposed by Baskin and Somers (2011), psychopathic individuals fail to integrate unexpected or response incongruent cues with an ongoing attentional set due to an early attentional bottleneck that prevents processing of this information. As a result, these cues cannot be used to re-evaluate and regulate the dominant response set. The attention bottleneck model further strengthens the response modulation theory by providing an explanation as to how the attention deficits occur. As a result, the attention bottleneck model helps to make clearer predictions about performance of psychopaths’ cognitive and affective tasks. One limitation, however, is that it has not yet been supported by neurobiological evidence.
4. The Decision-making network
In several economic decision-making theories, such as the expected utility theory (von Neumann & Morgenstern, 2007), the expected value of a choice outcome is the main factor in decision-making. The expected value of one of several options is defined as the sum of the utility
of every possible outcome, each multiplied by the probability of its occurrence (von Neumann & Morgenstern, 2007). Recent fMRI research has shown that the expected reward value and the subjective pleasantness of specific reinforcers are represented in the orbitofrontal cortex (OFC) (Ursu & Carter, 2005). Further, several neuroimaging studies have reported that the
OFC/ventromedial prefrontal cortex (vmPFC) encodes subject’s expected values in a wide range of economic decision-making tasks (T. a. Hare, O’Doherty, Camerer, Schultz, & Rangel, 2008; Todd a Hare, Camerer, & Rangel, 2009; Plassmann, O’Doherty, & Rangel, 2010, 2007). Besides that, the ventral striatum is central in value computations across different WTP tasks (Clithero & Rangel, 2014). In addition, neuroimaging research has revealed that the ventral striatum is anatomically and functionally connected to the reward circuit, including the globus pallidus, amygdala, insula and vmPFC/OFC (Cauda et al., 2011; Di Martino et al., 2008). Hence, the vmPFC and the ventral striatum are assumed to be critically involved in computing the expected value.
It has been proposed that in addition to the expected value, each possible action is simulated to anticipate the consequences (Bell, 1982; Loomes & Sugden, 1982). In healthy individuals decisions are influenced by the anticipated consequences induced by the simulation of each action. It should be noted that the expected value and anticipation of consequences heavily rely on each other, as the expected value is needed to anticipate the consequences and this in turn updates the expected value. The relevance of simulated consequences and possible regret is emphasized by research that found that the same pattern of activity observed during experienced regret, was also observed just before the choice, when the brain was anticipating possible future consequences of decisions and its counterfactuals (Coricelli et al., 2007). It is therefore possible that the same neural circuitry mediates both anticipated consequences and experienced regret. Consequently, decision outcomes may be simulated and evaluated before the decision is actually made. Different possible actions are simultaneously planned before deciding among them, and the unselected actions are not always suppressed before the actual choice is made (Klaes et al., 2011). Anticipated consequences can evoke strong feelings, which in turn have a large impact on decision-making (Bechara et al., 2000). Regarding the neural underpinnings of anticipated consequences, it has been found that hypothetical outcomes of unmade decisions are reflected by neuronal ensembles in the OFC as well as in the dorsolateral prefrontal cortex (dlPFC) and ventral striatum (Abe & Lee, 2011; Takahashi et al., 2013; Van der Meer & Redish, 2009). This importance is also supported by fMRI findings that show the OFC is involved in representing the
anticipated feelings associated with the outcomes of actions (Ursu & Carter, 2005). In addition, Hare and colleagues (2009) have found that exercising self-control in decision-making tasks involves the modulation of the decision-value signal in the OFC by the dlPFC. The dlPFC is assumed to encode higher order values (e.g. long term consequences) into the decision value signal (Hare et al., 2009). Furthermore, findings from connectivity research indicate, in line with known anatomical pathways, that OFC activity during decision-making is correlated with activity in dorsal and medial PFC, parietal and temporal cortices, as well as the striatum (Heekeren et al., 2003). This connectivity is associated with planning and working memory (PFC and parietal cortex), and reward calculation and prediction in the striatum (Cavada et al., 2000). The OFC integrates the cognitive information from these different brain regions with sensory input to simulate outcomes before a decision is made (Wallis, 2007). Hence, the connectivity between these regions is highly relevant for imagining possible consequences of decisions. The expected value seems to be computed in the OFC and ventral striatum, whereas the anticipated
consequences of one’s decision seem to be mainly associated with the OFC and dlPFC. Together these two processes influence decision-making.
Decision-making will be influenced by the quality of a person’s perception, the accessibility of relevant memories and prior learning, and the integration of all these considerations during response selection. In psychopaths, however, these decision-making processes are assumed to be deficient (A L Glenn, Raine, & Schug, 2009; Michael Koenigs, Kruepke, & Newman, 2010; Michael Koenigs, 2013). According to the response modulation hypothesis, impaired decision-making in psychopathy is due to the failure to stop and consider the consequences of one’s decisions and actions (Lorenz & Newman, 2002). Moral conventions, legal requirement and past experiences are not reflected upon before making decision and the implication are not anticipated. According to the emotion-based theories, psychopaths’ dysfunctional decision-making is caused by a lack of fear and a reduced emotional response leading to weak somatic markers. As a consequence, the question arises as to what are the underlying neural mechanisms that could explain the deficits in allocation of attention and anticipation of consequences, which result in disinhibited behaviour in psychopaths. Further, I will review which neural abnormalities could cause the impaired integration of affective information into the decision-making process.
5. The neural mechanisms underlying Psychopathy
In the last two decades, an increasing number of studies have applied neuroimaging techniques to investigate the neurobiological basis of psychopathy. These studies can provide insights about the neurological abnormalities observed in psychopaths that could not be explained by the attention based theories mentioned above. The most commonly reported structural and functional abnormalities are distributed across the frontal and temporal regions (R. J. R. Blair, 2010a). Of particular importance are limbic structures such as the amygdala as well as frontal regions such as the vmPFC. These findings have led to the development of two highly influential neurobiological theories of psychopathy.
Blair’s (2006) model primarily highlights the dysfunction of the amygdala as cause of development of psychopathy. The amygdala plays a central role in establishing connections between affective states and emotional cues and the development of basic threat circuits (R. J. R. Blair, 2006). More recently, Blair has also emphasized the importance of the vmPFC in
behavioural regulation (R. J. . Blair, 2008; R. J. R. Blair, 2010b). In contrast to Blair, Kiehl’s paralimbic dysfunction model (Kiehl, 2006) proposes that attentional and affective deficiencies observed in psychopaths are due to abnormalities in a more extensive set of brain regions. Kiehl’s model also accounts for abnormalities in paralimbic structures such as the amygdala, septal region and substantia innominate, which are strongly connected to the anterior cingulate cortex (ACC), posterior cingulate and vmPFC. The paralimbic structures are critically involved in linking cognition, emotions and visceral states (Eslinger, 2011). An advantage of Kiehl’s model is that it accommodates the broad neurological abnormalities and non-affective deficits seen in psychopathy. In the following paragraphs the individual brain regions associated with
psychopathy and their connectivity will be discussed in more detail. 5.1 Prefrontal Cortex
Brain regions in the prefrontal cortex have been found to be essential for monitoring ongoing behaviour, considering consequences and incorporating emotional learning into decision-making (Salzmann & Fusi, 2010). The OFC/vmPFC has been the most commonly studied prefrontal brain region in relation to psychopathy. Based on lesion studies, damage to the vmPFC has been associated with patterns of poor moral judgement and increased impulsivity
(Anderson, 2000). Individuals with lesions to the vmPFC exhibit emotional disturbances
including a diminished sense of empathy and guilt and blunted affect while manifesting preserved intellectual abilities (Bechara & Damasio, 2005). These clinical similarities, and several
laboratory paradigms, such as reversal learning (Budhani, Richell & Blair, 2006), smell identification (Lapierre, Braun, & Hodgins, 1995) and autonomic physiological responses to emotional stimuli (Blair, 1997), have demonstrated parallel deficits between psychopaths and vmPFC lesion patients. These parallels suggest that dysfunction in vmPFC plays a critical role in psychopathic behaviour.
Findings from neuroimaging research demonstrate that vmPFC damage is also associated with specific types of utilitarian moral judgements such as sacrificing one life to save two others (Koenigs et al., 2007), which emphasizes the non-empathic rationality characterizing decision-making in psychopaths. Concerning the anatomical structure, orbitofrontal grey matter reductions have been consistently reported when comparing psychopaths to controls (Tiihonen et al.,2008; Boccardi et al, 2011). Further, reduced cortical thickness in the OFC has also been observed to be related to response preservation (Yang, 2011), which is a well-established behavioural correlate of psychopathy (Newman et al., 1987; Vitale & Newman, 2001).
In line with the structural findings, functional imaging research has found reduced activity in the OFC in adolescents with psychopathic traits during the reinforcement stage of a passive avoidance learning task (Elizabeth C. Finger et al., 2011). In addition, in an emotional Simon paradigm, which involves integrating emotional information into ongoing behavioural outcomes, adult criminal psychopaths exhibited no prefrontal cortical activation during emotional
integration scenarios (Müller, 2008). Besides that, a PET study by Pietrini et al. (2000) has demonstrated OFC involvement in the expression of aggression in healthy people and suggested that OFC dysfunction may correlate with dyscontrol of aggressive and impulsive behaviour.
However, psychopathy is not always associated with reduced OFC/vmPFC activity. Sommer and colleagues (2010) found that while healthy individuals use the mirror neuron system when attributing emotional states to others, psychopathic patients (psychopathy cut-off was a PCL-R score of 28 or higher) utilize the OFC and the temporoparietal junction. Hence, dysfunction does not necessarily mean that these brain regions are not active at all in psychopaths. They might just be used differently than in healthy individuals.
1). The vmPFC is uniquely placed as a hub that integrates input from a variety of different regions to compute the decision value signal (Basten et al., 2010; Kahnt et al., 2010). Patients with vmPFC lesions have been found to conduct disadvantageous decision-making in personal and social realms. These individuals seem to be oblivious to the consequences of their decision due to the fact that they fail to learn from repeated mistakes (Bechara and Damasio, 2005). Conclusively, in line with Blair’s as well as Kiehl’s neurobiological theories, it can be said that structural abnormalities and atypical function of the vmPFC/OFC likely contribute to the deficits in decision-making observed in psychopaths. As the OFC is essential in anticipating the
consequences of one’s decisions, abnormalities in the OFC could explain why psychopaths do not consider the implications of their action (see Figure 1).
Figure 1. Goal-directed Action-model. Figure adopted from Kim (2013).
According to this model The OFC is involved in value judgment and sends the value signal to either the ventral striatum for reward anticipation or to the amygdala for punishment anticipation. Subsequently the ACC and OFC are involved in outcome evaluation. Further, the ACC and dlPFC are involved in cognitive control. This graph highlights the central role the OFC plays in goal directed action. However it also emphasizes the involvement of other areas such as the ACC, amygdala and dlPFC, which have all been reported to be dysfunctional in psychopathy.
5.2 The Amygdala
Numerous studies have shown that the amygdala plays a central role in the acquisition and conditioning of fear responses (Bechara et al., 1995; LaBar, LeDoux, Spencer, & Phelps, 1995). This is in line with the low fear hypothesis by Lykken (1957), which argues that many of the deficits observed in psychopathy can be attributed to the lack of fear. Further, the amygdala has been reported to be involved in stimulus reinforcement learning, anticipation of punishment (see Figure 1) and the detection and identification of affective information, such as the recognition of threatening stimuli (Davis & Whalen, 2001), which renders it highly relevant to the study of psychopathy. Converging evidence from fMRI studies indicates that psychopathy is linked to dysfunctional hemodynamic activity in the amygdala (Kiehl et al., 2001). More recent
neuroimaging research has demonstrated that psychopaths exhibit lower levels of amygdala activity when viewing pictures illustrating moral violations or fearful faces (Harenski, Harenski, Shane, & Kiehl, 2010). Higher levels of psychopathy have also been associated with lower levels of amygdala activation in youths, when engaged in passive avoidance learning (E C Finger et al., 2008). With regard to anatomical structure, it has been found that high levels of psychopathy are associated with reduced grey matter volume in the amygdala (Ermer, Cope, Nyalakanti, Calhoun, & Kiehl, 2013), while controlling for substance abuse, brain size and age. More specifically, Boccardi and colleagues (2011) found significant volume reductions in the basolateral amygdala, which shares reciprocal connections with the orbitofrontal cortex (OFC) and has been associated with updating reinforcement values (Schoenbaum, 2003). The amygdala is also highly important for decision-making, as it evokes emotional responses to affective stimuli such as monetary reward or punishment (see Figure 1). Patients with amygdala damage or individuals with a dysfunctional amygdala (i.e., psychopaths) lack these autonomic responses to reward and
punishment (Rupa Gupta, Timothy R. Koscik, Antoine Bechara, 2012). According to the somatic marker hypothesis, these individuals cannot use somatic markers, such as increased heart rate, and associate emotional stimuli with specific situations to guide future decision-making (Bechara & Damasio, 2005). Empirical support is provided by studies that found deficient decision-making in the Iowa Gambling Task in patients with bilateral amygdala damage (Bechara, Damasio & Damasio, 2003).
5.3 Paralimbic Structures
Based on Kiehl’s paralimbic dysfunction model (2006), which highlights the relevance of paralimbic structures, these brain regions will also be reviewed. These areas often have more subtle roles in emotional processing and integration of affective information in decision-making, but they are nevertheless critically involved. Kiehl (2006) especially emphasized the relevance of the parahippocampal gyrus, insula and anterior and posterior cingulate cortex. Concerning the structural imaging research, several studies have confirmed his theory. Applying radial and cortical distance mapping, Boccardi et al. (2011) reported reduced grey matter volumes in the ACC, parahippocampal gyrus and superior frontal gyrus. In a similar study with a large sample, Ermer and colleagues (2013) found that psychopathic individuals had reduced gray matter volumes in the posterior cingulate, parahippocampal region and the temporal lobe, which are all part of the paralimbic system. Importantly, this study controlled for effects of substance abuse, age, brain size, used more than 40 subjects and the psychopathy cut-off was above 30 on the PCL-R, which make these findings more reliable and less likely to be confounded by other factors.
With regard to functional imaging, psychopaths show reduced activity in the hippocampal formation and the cingulate cortex, while engaging in moral dilemmas (Pujol, 2012). Another study has found reduced cingulate cortex activity in psychopaths during defection in the prisoner’s dilemma (Rilling et al., 2007). These findings indicate that structural and functional abnormalities also exist beyond the amygdala and the prefrontal cortex and that several regions in the paralimbic system can be associated with decision-making and other deficits in psychopathy. Due to the intricate and strong connectivity between the paralimbic system and the prefrontal cortex, functional abnormalities may be partially attributed to dysfunction in the OFC and primary limbic regions. However, converging evidence suggests that damage to these paralimbic regions can lead to symptoms closely resembling psychopathy. Examples are lesions in the ACC leading to impairments in affective recognition (Hornak, 2003) and abnormally risky decision-making after damage to the insula (Weller, 2009).
decision-making process. Particularly the ACC has been found to monitor action outcomes and guide decision-making (Botvinick, 2007). It further serves to detect conflicts in information processing (see Figure 1). This response to conflict is proposed to serve as a teaching signal, which might drive a form of avoidance learning (Botvinick, 2007). Further, Clithero and Rangel (2014) demonstrated that the PCC in addition to the ventral striatum and vmPFC is central in computation of value across several tasks, reward modalities and stages of the decision making process. Based on these findings it can be concluded that abnormalities in paralimbic structures may critically contribute to decision-making deficits in psychopaths.
5.4 Differences in Structural and Functional Connectivity in Psychopaths
Recently, psychiatric disorders have more frequently been conceptualized as disorders of neural networks (Morris and Cuthbert, 2012). It has been proposed that dysfunction in structural and functional connectivity within or between major cortical networks such as the default mode network or the central executive network may underlie various psychiatric and neurological disorders (Morris and Cuthbert, 2012). Due to the fact that psychopathy involves severe deficits in integration of social and affective information into the decision making process (R. J. R. Blair & Mitchell, 2009; Lorenz & Newman, 2002), which presumably heavily depends on connectivity between the major cortical networks, it can be assumed that psychopaths exhibit abnormalities in connectivity. As a consequence, neuroimaging has started to investigate alteration in connectivity between neural networks in psychopaths.
5.4.1. Structural connectivity in psychopathy & Decision-Making
With regard to structural connectivity, a study by Craig and colleagues (2009), using diffusion tensor imaging (DTI), revealed that psychopathy is associated with reduced structural connectivity between the vmPFC and the amygdala. More specifically, they found decreased structural integrity of the uncinated fasiculus (UF), which is the primary white matter pathway connecting the vmPFC to the anterior temporal lobe structures, such as the amygdala. Notably, this study consisted of a sample of nine participants only and the PCL-R cut off used was 25 or greater. Another study by Motzkin et al. (2011) with a bigger sample (N=27, 14 psychopaths and 13 non psychopaths) and a more stringent PCL-R cut off (30 or greater) confirmed these results,
which supports the robustness of the association between psychopathy and decreased structural connectivity between the vmPFC and the amygdala. Finally, these findings were confirmed by the largest DTI study on incarcerated criminal offenders to date (N=147, PCL-R cut off = 30 or greater) (Wolf et al., 2015), showing that total PCL-R score was inversely related to structural connectivity in the right uncinated fasciculus. More specifically, they found that reduced connectivity in the UF was associated with interpersonal features of psychopathy, such as superficial charm, glib, manipulativeness, rather than affective, antisocial or lifestyle features (Wolf et al., 2015). There results provide a neurobiological marker of a specific and delineated factor of psychopathy.
5.4.2 Functional Connectivity
In line with the structural connectivity findings, it has been reported that psychopaths exhibit aberrant resting state functional connectivity between the vmPFC and the amygdala (Motzkin et al., 2011). These findings are in accordance with numerous studies emphasizing the importance of vmPFC and amygdala connectivity in the regulation of emotion and social
behaviour (Davidson, 2002; Delgado, 2008). Motzkin and colleagues (2011) also found reduced resting state functional connectivity between the vmPFC and posterior cingulate cortex (PCC), which has been suggested to be involved in self-reflective cognition (Qin & Northoff, 2011). Psychopathy has been proposed to be associated with deficient self-insight and self-reflection in relation to the affective state especially in decision-making (Koenigs, 2010), which will be discussed in more detail later. Another important finding from this study is that reduced connectivity can distinguish not only psychopaths from non-psychopaths but also primary psychopaths from secondary psychopaths (Motzkin et al., 2011). Hence, this study provides the first evidence of different neural underpinnings of the different subtypes of psychopathy.
Furthermore, Pujol and colleagues (2012) observed impaired integration between cognitive and affective circuits in psychopaths during a moral decision-making task, which was also present during rest. Specifically, they found that frontal areas, namely the vmPFC and dlPFC, exhibited reduced connectivity with the PCC, which is in accordance with the results form Motzkin et al. (2011). Besides that, reduced resting state functional connectivity between the left dorsal ACC and left insular cortex in psychopaths compared to controls has been reported by Ly et al. (2012) and Contreras-Rodriguez (2014). Reduced parietal-dorsal ACC connectivity
has been proposed to be a main correlate of psychopathy (Philippi et al., 2015). It has been proposed that the impaired communication between the parietal-dACC connectivity could lead to deficits in directing attention to salient stimuli (Philippi et al., 2015). This could explain
impairments in the allocation of attention to peripheral emotional cues observed in psychopaths (Joseph P. Newman et al., 1997; Hiatt, 2004).
With regard to task based functional connectivity, a recent neuroimaging study by Yoder and colleagues (2015), using 88 prison inmates, investigated the influence of psychopathy on neural networks in a moral decision-making task. More specifically, they examined the influence of task contingencies on neural activity in the salience network while participants were engaged in moral decision-making. The salience network is anchored by the dACC and the anterior insula and includes the superior temporal pole, supplementary motor area and hypothalamus. This network has been associated with attention allocation to personally or motivationally salient information (Harsay et al. 2013). In the context of moral decision-making, it is also involved in orchestrating shifts between cognitive control and default mode networks (Yoder et al., 2015). During the implicit moral task high levels of psychopathy were associated with reduced activity in the dlPFC and the caudate nucleus (located in the striatum) when participants were viewing harmful compared to helpful social interactions (Yoder et al., 2015). Moreover, functional connectivity analyses with seeds in the amygdala showed decreased coupling with the ACC, anterior insula, which are part of the salience network, but also the vmPFC and the striatum. In the explicit task psychopathy was characterized by decreased activity in several nodes of the salience network, including the amygdala and ACC (Yoder et al., 2015). These results suggest that activity and neural coupling in the salience network are disrupted in psychopaths. Further, these findings show that the effects of psychopathy on moral decision-making are influenced by attentional demands. These findings may provide neurological evidence for the response
modulation hypothesis proposing that the decision-making deficits in psychopaths are due to dysfunctional allocation of attention. Hence, disruption in the salience network may result in an inability to integrate contextual information that contraindicates goal-directed behaviour (Lorenz & Newman, 2002). The decreased connectivity in the salience network and the OFC may explain the psychopath’s deficit in anticipating the consequences of his actions and decisions.
Another study by Juarez and colleagues (2013) applied independent component analysis (ICA) to identify functional networks responding to the oddball task correlated with psychopathy scores on the PCL. Their results indicated that Factor 1 scores, interpersonal and affective traits,
were strongly associated with abnormalities in functional connectivity in the default mode network, particularly between the anterior and posterior cingulate. These abnormalities could contribute to deficits in the internal monitoring of cognitive and attentional processes (Juarez et al., 2013). In contrast, Factor 2 traits, antisocial and lifestyle traits, were more strongly linked to abnormalities in the frontoparietal network. This network has been associated with the allocation of attention to emotional stimuli (Bently et al., 2003). Further, dysfunction in the frontoparietal network has been linked to increased prevalence of reactive and impulsive aggression (Coccaro et al, 2011). Hence, the frontoparietal network seems to be strongly linked to the impulsive
dimension (Factor 2) of psychopathy. In general, these findings provide evidence that abnormal functional connectivity within cortico-limbic and paralimbic networks may underlie deficits in decision-making in psyhopaths. These findings further support Kiehl’s paralimbic dysfunction model emphasizing the importance of not only the amygdala and the vmPFC but also paralimbic structures. The study, in line with the findings of Yoder and colleagues (2015), further
strengthens the response modulation hypothesis by linking abnormalities in the frontoparietal network, associated with allocation of attention to emotional stimuli to impulsivity in
psychopathy.
With regard to decision-making research, a study by Cohen and colleagues (2008) used DTI and fMRI to assesses structural and functional connectivity during a feedback guided decision-making task. They found that both structural and functional connectivity between the amygdala, OFC and ventral striatum predicted how well participants adapted to rule reversals. Hence, based on converging evidence from psychopathy and decision-making research it can be assumed that abnormalities in connectivity between the salience network, the frontoparietal network, the OFC and the ventral striatum may lead to deficient allocation of attention, reduced flexibility and decreased ability to adapt to changing task demands. More specifically, it can be assumed that abnormalities in functional connectivity in the salience network leads to an
impairment in shifting between the DMN and the frontoparietal network, which results in deficits in cognitive control and attention allocation.
Conclusively, the evidence from neuroimaging research including structural, resting state and task-based neuroimaging research, consistently indicates that several brain regions involved in decision-making exhibit significant abnormalities in psychopaths. In particular, frontal and paralimbic regions show structural and functional dysfunctions and exhibit abnormal connectivity amongst each other. The amygdala, vmPFC and their connectivity are consistently reported by
neuroimaging studies, which provide support for the psychological perspective that deficient decision-making is due to impaired integration of affective information (Hiatt & Newman, 2006). These two regions are the central components of the somatic marker hypothesis, which argues that decision-making deficit is due to impaired weak physiological responses and consequently impaired avoidance learning. As mentioned above, the amygdala is assumed to encode affective information of environmental stimuli (Davis & Whalen, 2010), whereas the vmPFC integrates the input from different areas and discriminates the value of options and outcomes during decision-making (Basten et al., 2010, Grabenhorst & Rolls, 2011). Hence, reduced structural and
functional connectivity could lead to reduced communication between these regions, which results in deficient integration of emotional information and abnormal decision-making.
The results reported, however, also provide compelling evidence for the response modulation hypothesis. Abnormalities in several regions in the salience network and the frontoparietal network, which have been strongly associated with allocation of attention and cognitive control, could explain the deficits in response modulation and consequently impaired decision-making. As a result of this impaired flexibility of attention and reduced ability to adapt to changing task demands, the psychopath may be incapable of anticipating and evaluating the consequences of decision and action. The convergent evidence from these neuroimaging studies reveals that the areas centrally involved in decision-making all have been observed to be
structurally abnormal, and dysfunctional, which emphasizes that psychopathy clearly is a decision making disorder. As a consequence the next section will focus on studies investigating the decision-making impairment in psychopaths in order to further explore the underlying mechanism of the discussed theories.
6. Decision-making in Psychopaths: Evidence from Economic Decision-Making Tasks 6.1 Avoidance Learning
One well-studied and often reported cognitive process that may underlie deficient decision-making in psychopathy is avoidance learning. It refers to the process in which
individuals learn how to respond to avoid a stressful or unpleasant outcome (Archer et al., 1989). Avoidance learning has played a prominent role in early theories of psychopathy. A study of Lykken (1957) demonstrated that psychopaths exhibit deficient avoidance learning and poor fear
conditioning. In this study, psychopaths showed significantly less anxiety on a questionnaire particularly designed for this study and significantly less electrodermal response (GSR) to a conditioned shock stimuli, which were part of an avoidance learning task. Several studies have confirmed this deficit in passive avoidance learning (Schachter, 1964; Gorenstein & Newman, 1980). A study by Newman and Kosson (1986) then revealed that psychopaths only showed significantly more avoidance errors compared to controls when the task combined competing goals but performed equally well as non-psychopaths when subject’s only goal was to avoid punishment. Hence, the avoidance-learning deficit is dependent on task demands. Based on this finding the response modulation hypothesis was formulated, stating that psychopaths suffer from an attentional deficit that entails the inability to switch attention once it has been allocated to a motivationally significant goal (Joseph P. Newman, Patterson, Howland, & Nichols, 1990). Stated differently, it has been proposed that deficient avoidance learning in psychopaths is due to impaired response modulation (Joseph P. Newman et al., 1997). The difficulty in switching attention away from a dominant response set to process nondominant peripheral cues that are important for guiding behaviour causes poor self-regulation.
6.2 Reversal Learning
Besides that, impaired reversal learning may contribute to the decision-making deficits observed in psychopaths. Reversal learning tasks measures behavioral flexibility by requiring subjects to switch from one learned response set to a different response set when task demands and contingencies change. Hence, reversal learning entails both the processing of negative feedback once a learned response set is no longer reinforced, and the ability of flexible response selection (D’Cruz et al, 2011). A study by Mitchell and colleagues (2002) investigated the performance of psychopaths on the four-pack gambling task and the
intradimensional/extradimensional (ID/ED) task. Whereas the four-pack gambling task assesses reversal learning, the ID/ED task indexes extradimensional set shifting as well as response reversal performance, which are two dissociable function of the frontal cortex. The results of this study revealed that psychopaths were significantly less likely to avoid risky gambles and learn from mistakes in the gambling tasks. Further, psychopaths exhibited selective response reversal impairment in the ID/ED task, while performing similar to controls in the other components of the task. These results, in line with suggestions of a previous study (LaPierre et al., 1995),
indicate that psychopathy is associated with deficits in reversal learning. This provides further support for the response modulation hypothesis. Response reversal requires participants to reallocate their attention from a dominant response set to contraindicative peripheral cues in order to switch to a more beneficial response set once task contingencies change. As proposed by the response modulation hypothesis psychopath fail to consider the relevant peripheral information and persist in responding in a previously rewarded response, although punishment to this
response increases. Stated differently, due to dysfunctional response modulation psychopaths fail to attend to information indicative of changing task demands and consequently fail to adapt their responses. This impairment may reflect dysfunction in the OFC and the amygdala and the connectivity between these two regions. As mentioned above, the amygdala is involved in stimulus-reinforcement associations (LeDoux et al., 1998) and the OFC encodes motivational cues and the expected value of outcomes (Plassmann et al., 2007; Rangel, Camerer, & Montague, 2008), the connectivity between these regions is essential in encoding associative information about the motivational significance of stimuli. Impaired reversal learning in adult individuals with psychopathic traits has been confirmed by subsequent research using a probabilistic reversal-learning task (Budhani et al., 2006). A neuroimaging study by Finger and colleagues (2008) investigated reversal learning and its neural underpinnings in psychopathic adolescents as well as in adolescents with ADHD and conduct disorder. Their research revealed that children with psychopathic traits exhibit aberrant activity in the vmPFC during punished reversal errors compared to controls and adolescents with ADHD. Adolescents with ADHD and controls did not differ in their vmPFC activity which suggests that vmPFC dysfunction during reversal learning is specific to psychopaths. As mentioned above, numerous studies have demonstrated the critical involvement of the OFC/vmPFC in the generation and representation of reinforcement
expectation and expected value across species (Hampton et al., 2007; Schoenebaum et al., 2002, Plassmann et al., 2010, Grabenhorst et al., 2011). Hence, these findings further emphasizes the vmPFC’s essential role in reversal learning is the processing of reinforcement expectation. Psychopathic adolescents lack the typical reduction in vmPFC activity during punished reversal learning errors, which indicates that these individuals may not appropriately process violations of reinforcement expectation. This may impair their ability to detect contingency change during reversal learning (Finger et al., 2008). As a result of their impairments in the processing of reinforcement information and expectation, psychopathic adolescents and adults may be
predisposed to impaired decision-making, which may lead to a high prevalence of frustration and reactive aggression.
These findings are in accordance with previous neuroimaging research showing abnormal vmPFC function in psychopaths in a wide range of tasks (Finger et al., 2010; Müller, 2008; Pujol 2012) and provide further support for central role of vmPFC dysfunction in deficient decision-making in psychopaths. It has to be noted, however, that the study by Finger and colleagues (2008) included only 14 psychopathic individuals, which leads to low power and reduced generalizability of these findings.
To sum up, these findings suggest that deficits in reversal learning, frequently observed in psychopaths, may be due to dysfunctional response modulation. The failure to attend to relevant peripheral information results in response perseveration in spite of changing task demands. These deficits in reversal learning may contribute to risky decision-making.
6.3 Social cooperation in the prisoner’s dilemma
In order to investigate how information and emotion processing deficits in psychopaths lead to disturbances in social decision-making, Rilling and colleagues (2007) examined the neural correlates of cognitive processes during interactions in an iterated prisoner’s dilemma (PD) game. Specifically, thirty subjects were scanned using fMRI while they were engaged in interactions with human confederates who were outside the scanner. The PD game is an experimental paradigm in which two participants both have two options whose outcome crucially depends on the simultaneous choice made by the other, formulated in terms of two prisoners separately deciding whether to confess to a crime. If both players decide to cooperate (CC outcome) then both players receive 2 Dollars. If Player A defects and player B cooperates player A receives 3 Dollars and player B receives Dollars and vice versa. Finally, if both players defect both players receive 1 Dollar. The iterated version of this game models relationships that are based on reciprocal exchange of favors (Rilling, 2007). There are different emotions related with the different outcomes. Whereas mutual cooperation is often associated with friendship, love and trust mutual defection is associated with rejection and hatred. Cooperation by one and defection by the other player leads to anger or indignation in the cooperator and results in the defector experiencing anxiety, guilt or elation from successfully exploiting the other player (Rilling et al., 2007). The PD game is unique among decision-making tasks in its quality that optimal choices
can arise from either emotional or cognitive motivations. Rilling et al. (2007) found that
individuals with high levels of psychopathy more frequently chose to defect and often did so after a mutually cooperative interaction, which can be interpreted as aiming for immediate payoffs at the expense of long-term success. This is in accordance with psychopaths’ impaired impulse control and inability to form long lasting social relationships. It was also observed that
psychopathic individuals also experienced more CD outcomes, which might be a result of their choice to defect frequently. Interestingly, CD outcomes lead to significantly less amygdala activity than in controls, which suggests impaired aversive learning expressed in the failure to learn to avoid these outcomes (Rilling et al, 2007). The analysis of neural activity during the decision-making stage revealed that controls exhibit emotional biases towards cooperation, reflected by OFC activity, which can be overcome by increased cognitive effort, associated with activity in the dlPFC. The decision not to cooperate also leads to emotional conflict, which is reflected in activity in the ACC, commonly associated with emotional and cognitive conflict (Kerns et al., 2004). Contrarily, subjects with psychopathic traits showed biases towards defection, represented in the OFC, which have to be overcome by exerting cognitive effort, reflected in dlPFC activity, in order to cooperate. Stated differently, in order to make a morally appropriate decision and to delay gratification in favour of long-term success, psychopaths have to exert increased cognitive effort, which is associated with activity in the dlPFC.
It can be concluded that amygdala dysfunction, causing aversive conditioning deficits, and reduced OFC response when deciding to cooperate, reflecting absence of biases toward minimizing distress of others, lead to impaired social decision-making in the PD game. The resulting negative impact on long-term success can only be overcome by increased cognitive effort and reasoning, which is associated with increased dlPFC activity. The focus on immediate reward instead of long-term success is in line with the response modulation hypothesis. The failure to stop and reflect on the consequences of a decision (e.g., defection and immediate reward) can be attributed to deficient response modulation. However, these findings also suggest that psychopathic individuals can compensate for their deficient emotional processing by means of cognitive effort and abstract reasoning. This more rational reasoning is reflected in increased neural activity in the dlPFC during social decision-making. It has to be noted that the sample did not contain diagnosed psychopaths but consisted of college students, some of which were found to have high scores of psychopathy on the Levenson self-report psychopathy scale (LSRP, Levenson, 1995). The use of college students may explain why some of the participants with
psychopathic traits were able to engage in more rational decision-making focusing on long-term rewards. These observations might not be replicated in diagnosed psychopaths. Further research is needed for clarification.
6.4 The Ultimatum Game
Whereas the emotional deficit in psychopaths that leads to the disinhibition of selfish behaviors is generally considered an impediment to a successful life in human society, Osumi and colleagues (2010) propose that psychopathy may have a positive effect on decision-making in social settings. In order to provide evidence for their proposal they examined the decision-making of college students with high and low levels of primary psychopathy in the ultimatum game (UG). The UG is an experimental paradigm that illustrates conflict between fairness and economic utility (Osumi, 2010). In this paradigm the first player receives a certain amount of money and proposes how to divide the sum between himself and the other player. The second player chooses to either accept the proposal or reject it; in the latter case none of the player gets any money. Based on this rule the second player should accept any offer, however, actual
responders often turn down this rational decision due to negative emotions such as frustration and anger in response to an unfair offer. Hence, the UG represents an interpersonal situation in which emotional function (e.g., perceived fairness) can violate economic rationality (Güth et al. 1982). In addition, they also assessed electrodermal responses (GSR) in response to fair and unfair offers. Skin conductance has been found to be a predictor of irrational rejection in the UG (van Wout et al., 2006). It is assumed that the SCR associated with rejection is caused by negative emotions such as anger and frustration (Osumi & Ohira, 2010). Psychopaths often exhibit deficient SCR in aversive learning tasks (D T Lykken, 1957) which implies that they are less sensitive to negative emotional events.
In line with this research, Osumi and colleagues (2010) found that students with higher levels of primary psychopathy more often accepted unfair offers than students with low levels of primary psychopathy. This study also showed that while individuals with low psychopathy exhibited higher SCR in response to unfair offers, individuals with high levels of psychopathy showed equal amplitudes of SCR for fair and unfair offers. Moreover, lower SCR was associated with higher acceptance rates of unfair offers, which strongly suggests an emotional motivation for rejection of unfair offers. As a consequence, it can be assumed that individuals with high
levels of psychopathy are insensitive to violations of fairness, which helps them to make more rational economic decisions. These findings support the view that psychopathy is due to affective dysfunction.
Notably, these interpretation only hold true due to the fact that only one round of the UG was played. If an iterated UG would have been used the rejection of unfair offers in the beginning may lead to fairer offers in subsequent rounds and may consequently be the better long-term strategy (Osumi, 2010). In this sense accepting unfair offers can be seen as a fixation on immediate rewards at the expense of long-term success, which has also been observed in
psychopathic individuals in the previously discussed study using the Prisoner’s Dilemma (Rilling et al., 2007). Further research would have to investigate the behaviour of psychopaths in an iterated UG to confirm these hypotheses. One considerable limitation in this study is that again no diagnosed psychopaths were used. Instead, the sample consisted of college students with psychopathic tendencies identified by the Levenson scale (Levenson, 1995). As a consequence, the behaviour reported in this study may differ considerably in a sample of convicted
psychopaths due to differences in level of education, IQ, age and other latent variables.
In a related study by Koenigs and colleagues (2010) the UG and the Dictator Game (DG) were used to investigate whether primary and secondary psychopaths exhibit different patterns of economic decision-making. Similar to the UG, the dictator game consists of two players who have to split a certain amount of money. In contrast to the UG, the responder has to accept whatever amount he is offered. Hence, the amount that is offered represents a prosocial sentiment such as empathy or guilt. Further, this study tested whether either of the subtypes shows similar decision-making to patients with vmPFC lesions. As pointed out before, several clinical and laboratory studies have reported that patients with vmPFC lesions and psychopaths exhibit several clinical similarities (Bechara & Damasio, 2005) and exhibit the same deficits in reversal-learning (Budhani et al., 2006). A previous study by Koenigs and colleagues (2007) has
demonstrated that patients with vmPFC lesions, known for their high irritability and low frustration tolerance despite their general hypo-emotionality (Bechara et al., 2005), reject a significantly higher proportion of unfair offers than controls. With regard to the DG, vmPFC lesions patients, known for abnormally low levels of empathy and guilt, have been observed to offer considerably lower amounts of money as compared to controls (Koenigs et al., 2007). Using a sample of prison inmates Koenigs and colleagues (2010) found that primary psychopaths (n=6) exhibited significantly lower acceptance rates in the UG as well as significantly lower offers in
the DG than secondary psychopaths or non-psychopathic individuals. In contrast, secondary psychopaths did not differ in any aspect of the UG or DG from the non-psychopathic controls. It has to be noted that possible confounds such as age, IQ, psychopathy or Factor 1 and 2 severities or were accounted for and thus cannot explain the reported differences (Michael Koenigs et al., 2010). These findings reveal that primary psychopath decision-making performance closely resemble the decision-making of patients with vmPFC lesions patients. Further, this study shows that primary and secondary psychopaths do differ in their economic decision-making
performance, which may be due to different neurobiological impairments underlying these different subtypes. In addition, further evidence is provided for the central role of vmPFC dysfunction in decision-making impairments in psychopaths. On closer inspection it becomes apparent that decision-making performance of primary psychopaths and vmPFC lesion patients is somewhat contradictory. On the one hand, reduced emotionality, expressed in low levels of guilt and empathy, are assumed to underlie the low offers in the DG as higher offers are at least partially determined by prosocial sentiment as there is no finiancial interest in offering high amounts. On the other hand, the high rate of rejection of unfair offers is assumed to be due to higher irritability, exaggerated anger and lower frustration tolerance (Anderson et al., 2006). This hyper-emotionality and the inability to regulate these emotions could explain the lower
acceptance rates. Koenigs and colleagues (2010) explain this contradiction between hypo-emotionality in the DG (low empathy and guilt) and hyper-hypo-emotionality in the UG (anger and frustration) in terms of the response modulation hypothesis. As stated above, deficient response modulation in psychopaths results in a failure to stop and reflect on potentially maladaptive nature of their behavior. In decision-making this applies to self-insight and self-reflection on affective states. Frustrating situations such as unfair offers in the UG, evoke the urge to respond aggressively (e.g., reject an offer). Whereas healthy individuals would reflect on the
consequences of such impulsive behaviour on social relationships and modulate their response to some degree, psychopaths and vmPFC lesion patients do not engage in self-reflection, fail to resist the urge and thus react aggressively. Koenigs and colleagues also propose that the
employment of prosocial emotions (e.g,. empathy and guilt) that motivate higher offers in the DG are dependent on self-reflection and insight. Due to deficient response modulation psychopaths are unable to reflect on the effects of their actions on their social environment. As a consequence, the prosocial emotions appear considerably diminished (Koenigs et al., 2010). Hence, deficient
