Psychopathy is a serious mental health disorder defined by a cluster of interpersonal, affective, and behavioral characteristics including impulsivity, grandiosity, and callousness (Hare, 2003). A core feature of psychopathy is an inability to experience a normal range and depth of emotions, including “moral emotions” such as guilt and empathy. These characteristics are early-emerging, severe, and persist into adulthood (Lynam et al., 2007).

Deficits in emotion processing related to psychopathy have been demonstrated across multiple modalities. Psychopathic males show reduced physiological responses to unpleasant and fear-inducing events (Patrick et al., 1993, 1994), impaired ability to identify multiple types of facial and vocal expressions (Blair et al., 2002; Kosson et al., 2002; Glass and Newman, 2006; Bagley et al., 2009; Dawel et al., 2012), and reduced response facilitation to emotional words in lexical decision tasks (Lorenz and Newman, 2002). Fewer investigations of these processes have been conducted in female psychopathy. One study found that low-anxious psychopathic females showed reduced startle potentiation to unpleasant images (Sutton et al., 2002). Another study found that psychopathic and non-psychopathic females performed similarly on tasks evaluating response facilitation to emotional words in a lexical decision task, and response inhibition in a passive avoidance learning task (Vitale et al., 2011). These results suggest that female and male psychopathy share some, but not all, of the same aberrant emotional responses.

Neuroimaging studies have provided additional support of emotional dysfunction in psychopathy. Psychopathic males show reduced engagement of brain regions associated with emotion processing including the amygdala, anterior cingulate/ACC, and ventromedial prefrontal cortex/vmPFC when viewing emotional words, conditioned fear stimuli, statements, or pictures depicting immoral behavior, and during emotional perspective taking tasks (Kiehl et al., 2001; Birbaumer et al., 2005; Glenn et al., 2009; Harenski et al., 2010a; Decety et al., 2013). Individuals with high levels of callous-unemotional traits show reduced amygdala activation to fearful facial expressions (Marsh and Blair, 2008; Jones et al., 2009), and psychopathic males also show reduced vmPFC and orbitofrontal cortex activation to several types of emotional facial expressions (Decety et al., 2014). Regarding females, neuroimaging studies have reported negative correlations between self-reported psychopathic traits and amygdala responses to fearful facial expressions (Carre et al., 2013) and unpleasant pictures (Harenski et al., 2009). These results suggest that female psychopathy, like male psychopathy, may be associated with reduced engagement of limbic regions during affective tasks. However, this hypothesis has not been tested in females with clinical levels of psychopathy.

Emotion deficits and associated brain dysfunction in psychopathy are believed to impair moral development. Blair (2007) proposed a neurodevelopmental account of psychopathy in which early dysfunction within the amygdala and vmPFC contributes to impaired moral socialization. The proposal is based on the importance of stimulus-reinforcement associations in moral socialization (learning that certain behaviors are harmful to others and should be avoided) and the role of the amygdala (valence, i.e., “good”/“bad” representation) and vmPFC (outcome expectancy) in these processes. Consistent with this hypothesis, psychopathic males have shown reduced vmPFC and amygdala engagement when presented with pictures depicting immoral behaviors and complex moral dilemmas (Glenn et al., 2009; Harenski et al., 2010a).

Whether the neuroimaging findings in male psychopathy summarized above extend to female psychopathy is unknown. We used fMRI to scan a large sample of incarcerated females with varying levels of psychopathy, and a comparison group of non-incarcerated females, while they viewed unpleasant pictures that did or did not depict moral transgressions and thematically-matched neutral pictures. They also rated the severity of moral transgressions. We predicted that psychopathy would be negatively correlated with brain activation while viewing unpleasant relative to neutral pictures in the ACC, amygdala, and vmPFC. We further predicted that activation within the latter two regions would be negatively correlated with psychopathy when viewing unpleasant pictures depicting moral transgressions relative to pictures without moral transgressions.

Regarding the moral salience aspect of the picture task used in this study, the task has been shown to reliably engage brain regions frequently implicated in moral judgment in non-clinical populations (Harenski and Hamann, 2006; Harenski et al., 2008, 2012). Such regions include the amygdala and vmPFC, and additionally the temporo-parietal junction [which supports several component cognitive processes related to inferring mental states (Decety and Lamm, 2007)], posterior cingulate and precuneus (Young and Dungan, 2012). Whether engagement of these latter regions (i.e., beyond vmPFC and amygdala) would show psychopathy-related differences to moral stimuli was an open question.

This study examined the neurobiology of emotion and moral judgment in female psychopathy. In accordance with hypotheses and prior studies of male psychopathy, PCL-R scores were negatively correlated with activation to unpleasant pictures in the right amygdala and rostral ACC. PCL-R scores were also negatively correlated with activation to unpleasant pictures depicting moral transgressions in the right TPJ. These results extend prior findings regarding emotion processing in adult male psychopathy to female psychopathy, and reveal aberrant neural responses to morally-salient stimuli that may be unique to female psychopathy.

Reduced amygdala engagement during emotional processing is a consistent finding in male psychopathy, in such tasks as viewing emotional words, conditioned fear, making moral judgments, and during affective perspective taking (Kiehl et al., 2001; Birbaumer et al., 2005; Glenn et al., 2009; Harenski et al., 2010a; Decety et al., 2013). Anatomical imaging studies of male psychopathy have also reported reduced amygdala gray matter volume (Yang et al., 2009; Ermer et al., 2012). While the anatomical findings implicated bilateral amygdala, the present findings and those of prior fMRI studies have often been lateralized to the right (Harenski et al., 2010a; Decety et al., 2013) or left (Birbaumer et al., 2005; Glenn et al., 2009) amygdala, though one study reported bilateral effects (Kiehl et al., 2001). Because this is the first fMRI study of female psychopathy in an incarcerated population, it would be premature to conclude that reduced amygdala activation to emotional stimuli is right-lateralized, especially given the focus of this study on moral judgment. Whether these findings extend to other emotion paradigms is a question for future studies.

The analysis of PCL-R factor scores revealed a negative correlation between right amygdala activation to unpleasant pictures and Factor 2 scores. This is in contrast to prior male psychopathy studies in which negative associations with amygdala activation were primarily related to Factor 1 scores (Harenski et al., 2010a; Decety et al., 2013). However, the correlation was no longer significant when substance dependence was included as a covariate. This is likely due to predictor criterion overlap (i.e., substance abuse is a symptom of psychopathy and measured most directly by PCL-R Factor 2 items). Additionally, the negative correlation between right amygdala activation and total PCL-R scores was significant regardless of the inclusion of substance dependence as a covariate.

Relative to the amygdala, the status of ACC function in psychopathy is less clear. ACC dysfunction is proposed in some theoretical models of psychopathy (Kiehl, 2006) but not others (Blair, 2005). Anatomical and functional imaging studies of the ACC in male psychopathy have produced mixed findings. Anatomical studies have not reported reductions in ACC volume (Glenn et al., 2010; Ermer et al., 2012), whereas functional imaging studies have found reduced rostral ACC responses during emotion tasks (Kiehl et al., 2001; Birbaumer et al., 2005). Here we observed a negative correlation between PCL-R scores and rostral ACC activation to unpleasant pictures. The region of rostral ACC that was correlated with psychopathy corresponds to the rostral section of the anterior mid cingulate cortex/aMCC, bordering the pregenual anterior cingulate. The aMCC has been implicated in studies examining a wide variety of negative emotion and cognitive control processes and has extensive connections with subcortical brain regions including the amygdala and ventral striatum (Shackman et al., 2011). One possible explanation for functional but not anatomical psychopathy-related differences in this region is that it may be anatomically intact in psychopathy but less responsive when processing emotional stimuli due to reduced input from one or more of the emotional processing brain regions with which it is connected (e.g., amygdala).

Similar to the findings in the right amygdala, the analysis of PCL-R factor scores revealed a negative correlation between rostral ACC activation to unpleasant pictures and Factor 2 scores. Unlike the findings in the amygdala, this result was unaffected by substance dependence. Prior psychopathy studies that reported reduced ACC function did not examine associations with Factor scores (e.g., Kiehl et al., 2001; Birbaumer et al., 2005). Thus, it is unknown whether the Factor 2-ACC negative correlation during emotion processing is unique to female psychopathy or also present in male psychopathy. More research in both male and female psychopathy is needed to explore this possibility. Overall, the current results suggest that reduced ACC function during emotion processing is related primarily to the lifestyle/antisocial aspects of female psychopathy. It is also worth noting that PCL-R Factor 2 contains items relevant to the developmental nature of psychopathy, including childhood, adolescent, and adult-related disruptive behavior; thus the effects related to Factor 2 may be developmental in nature.

The comparison of moral vs. non-moral pictures yielded results that were less consistent with prior findings in male psychopathy. PCL-R scores were negatively correlated with right TPJ activation. In our prior study of male psychopathy using the same moral judgment task, psychopathy was not associated with reduced (or increased) TPJ responses when viewing moral pictures (Harenski et al., 2010a). However, psychopathic males showed a positive correlation between right TPJ activation and severity of moral transgression ratings. Non-psychopathic males, in contrast, showed no correlation between right TPJ and severity ratings. Drawing from neuroimaging and lesion evidence regarding the role of the right TPJ in moral judgment (Greene and Haidt, 2002; Moll et al., 2005; Raine and Yang, 2006), particularly in attributing intentions to moral transgressions (Young et al., 2010; Koster-Hale et al., 2013), we proposed that intention attributions influenced severity of moral transgression ratings more in psychopathic males relative to non-psychopathic males. In the present study, female psychopathy was associated with less engagement of the right TPJ when viewing all moral pictures (regardless of how moral transgressions were rated) which could indicate a generally reduced sensitivity to the intentionality aspect of moral transgressions.

Although the role of the TPJ in making intentionality attributions during moral judgment has been well-demonstrated, studies have highlighted other functions of this region. For example, the right TPJ has been implicated in sense of agency, self-other discrimination, and directing attention to task-relevant stimuli (Decety and Lamm, 2007; Mitchell, 2008). Decety and Lamm (2007) proposed that rTPJ is associated with a variety of lower-level processes (e.g., redirection of attention) that contribute to higher level functions such as mentalizing. Thus, lower TPJ activation in psychopathic females could be related to intention attribution and/or other TPJ-related processes. The moral pictures contained multiple cues that could be differentially attended to by high and low psychopathy females. In other words, specific components of the pictures that are considered relevant to moral judgment, and the amount of attention paid to those components, may differ as a function of psychopathy. Future studies employing methods such as eye tracking could shed light on this possibility. Right TPJ activation has also shown elevated activation in individuals with heightened justice sensitivity as well as in individuals more likely to exculpate harmful actions resulting from innocent intentions (Young and Saxe, 2009; Yoder and Decety, 2014). It is possible that females higher in psychopathy have reduced justice sensitivity, which may be emphasized when evaluating moral transgressions and related to TPJ activation. This possibility requires further study. Regarding the forgiveness of unintentional moral transgressions, psychopathic males are in fact more likely than non-psychopathic males to exhibit this behavior (Young et al., 2012). However, whether psychopathic females are similar to psychopathic males in this respect has not been investigated.

PCL-R scores were unrelated to behavioral moral judgments. This result is in line with prior findings in male psychopathy using the same task (Harenski et al., 2010a), and other studies showing similar moral judgments in psychopathic and non-psychopathic males (Cima et al., 2010; Aharoni et al., 2012). However, male psychopathy-related differences in certain types of moral judgment have been reported (Koenigs et al., 2012; Young et al., 2012). The moral task used in this study was relatively straightforward compared to other tasks such as moral dilemmas. Female psychopathy may be associated with abnormalities in the latter type of moral judgment.

Limitations of this study should be considered. The standard PCL-R cutoff score for psychopathy is 30/40. Eight participants in this study scored 30 or higher on the PCL-R. We might have observed stronger and/or additional correlations between brain activation and PCL-R scores if more 30+ participants were included. Previous studies used lower PCL-R cutoff scores (e.g., 24) to identify female psychopathy (Vitale et al., 2007, 2011) due to lower base rates of psychopathy relative to males at the 30 cutoff score (Vitale et al., 2002) and lower levels of criminal behavior in females than males (Goldstein et al., 1996). Regarding our emotion task, it should be emphasized that participants made moral judgments of all pictures. We previously demonstrated that judging morally-salient pictures on severity of moral transgression, relative to making indoor/outdoor judgments of the same pictures, was associated with greater activation in the vmPFC, a region prominent in neurodevelopmental theories of psychopathy (Harenski et al., 2010b). It is possible that using a more “implicit” emotional task not related to moral judgment could yield additional/different findings.

In conclusion, we demonstrated that females with clinical levels of psychopathy show reduced amygdala and ACC activation when processing emotional stimuli, similar to findings in male psychopathy. Female psychopathy was also associated with aberrant right TPJ activation related to moral judgments, but the results differed from prior findings in male psychopathy. Overall, the results provide new evidence of neurobiological dysfunction in female psychopathy and support theories of limbic and paralimbic dysfunction in psychopathy.