Commentary: I don’t second that emotion: subjective experience of fear in adolescents with psychopathic traits – reflections on


In an important early characterisation of the psychopathic personality, Cleckley wrote that the psychopath: ‘...appears almost as incapable of anxiety as of profound remorse.’ He also remarked that the psychopath shows ‘...a relative immunity from such anxiety as might be judged normal or appropriate in disturbing situations’. These clinical descriptions set the stage for experimental investigations of fear and anxiety in psychopaths.

In one of the most famous doctoral theses ever published, Lykken (1957) demonstrated that adult psychopaths show deficits in fear conditioning, as evidenced by impaired skin conductance responses to conditioned stimuli that predicted electric shocks. The psychopaths also displayed deficits in passive avoidance learning when negotiating a mental maze in which certain choices led to electric shock. Lastly, Lykken found that psychopaths expressed a preference for highly arousing or dangerous experiences such as parachute-jumping over mundane and repetitive, but safe, activities, whereas control subjects showed the reverse pattern. Lykken’s work thus provided an empirical basis for the view that psychopaths are deficient in their experience of fear and anxiety, and his results were subsequently replicated and extended by researchers such as Robert Hare, Christopher Patrick and Joseph Newman.

The study by Marsh et al. (2011) in this issue of the Journal used a self-report method developed by Scherer and Wallbott to provide further evidence for a deficit in the experience of fear in adolescents with psychopathic traits. Participants were asked to recall real-life events in which they had felt specific emotions, such as happiness, anger, sadness or fear, and to articulate and quantify the symptoms of physiological arousal and other cognitions that had accompanied these affective states. While a control group of typically-developing adolescents reported greater physiological arousal during the experience of fear relative to other negative emotional states, adolescents with psychopathic traits and disruptive behaviour disorders failed to show heightened arousal for fear, thereby failing to differentiate between the negative emotions. The adolescents with psychopathic traits did, however, show increased arousal for negative, compared with positive, emotions indicating that they were able to recall instances in which they had experienced strong emotions accompanied by the appropriate physiological sensations. This runs counter to the view that psychopathic individuals are ‘unemotional’ or show a ‘general poverty in major affective reactions’. Marsh et al. claim that they were able to measure symptoms of sympathetic and parasympathetic arousal by self-report, and that the deficits in arousal in adolescents with psychopathic traits were specific to indices of sympathetic activity (e.g., faster heartbeat) during fear. Finally, the participants were asked to estimate how strongly and how frequently they felt each of the emotions. The adolescents with psychopathic traits reported feeling fear less often and less strongly than typically-developing adolescents. Importantly, this deficit in subjective experience again appeared to be selective to fear, rather than being manifest as a generalised emotion deficit.

The closest we can get to ethically inducing fear in the laboratory is to study aversive conditioning or present visual images conveying threat cues, so this work provides a useful bridge between research on these processes and ecologically valid assessments of fear in real-life situations. It also neatly complements a study by Patrick and colleagues (1994), in which adult criminal psychopaths, non-psychopathic criminals and a control group were asked to imagine highly unpleasant or frightening situations, and rate the intensity of their resulting subjective states. Patrick et al. (1994) also measured autonomic responses while the subjects imagined the aversive situations, using psychophysiological recording equipment. Interestingly, while subjective reports of emotional experience did not differ between groups, heart rate and skin conductance responses were attenuated in the psychopaths while they imagined fear-inducing scenarios. However, this attenuated autonomic response was not specific to the psychopathic group, but was also observed in the non-psychopathic criminals. Consequently, Factor 2 (antisocial behaviour) scores on the Psychopathy Checklist appeared to be more influential than Factor 1 (emotional detachment) scores in determining physiological responses to fear imagery. More recently, Fung et al. (2005) reported that skin conductance responses occurring in anticipation of an aversive stimulus were similarly reduced in psychopathy-prone and non-psychopathy-prone adolescents with high levels of antisocial behaviour, relative to healthy controls. The results of these two studies raise the possibility that fear deficits (or at least impairments in the autonomic correlates of fear) apply to non-psychopathic forms of antisocial behaviour as well. Since all of the adolescents with psychopathic traits included in Marsh et al. (2011) also had disruptive behaviour disorder diagnoses, in future research it will be important to examine whether deficits in the subjective experience of fear are specific to individuals with psychopathic traits, or are observed more generally in those with disruptive behaviour disorders (with or without psychopathic traits).

Intriguingly, the results of the current study by Marsh et al. also provide a potential mechanistic explanation for psychopaths’ deficits in empathy and emotion recognition, which appear to be particularly marked for states of fear in others (Marsh & Blair, 2008). According to perception-action models of empathy (e.g., Preston & de Waal, 2002), in order to understand how someone else is feeling, we have to model their affective state. Supporting this view, a network of brain regions, including the anterior insular cortex, is activated both when experiencing pain ourselves and observing someone else in pain. However, this matching of representations may be impaired if one has little experience of the affective state that is being modelled. According to Preston and de Waal (2002): ‘If a subject needs to access representations of a particular internal state to understand the object’s situation, then one would expect more empathizing for situations or states that the subject has experienced.’ (p. 17). In the context of this theory, the findings of Marsh et al. may prompt researchers to assess the extent to which impairments in subjective experience of fear relate to deficits in the recognition of facial or vocal expressions of fear in others. This would seem to be an important next step in terms of extending the current findings. If an association between these deficits can be confirmed, this would have clear implications for perception-action or simulation models of empathy as well as enhancing our understanding of psychopathy. It may also prompt investigations of the integrity of mirror neurone systems for emotions in psychopaths.

While this commentary has highlighted the positive features and interesting results of the article by Marsh et al., it should also be noted that the study is subject to a few limitations. First, measuring subjective reports of physiological arousal as elicited by past life events relies on introspection and assumes that the respondents were not only aware of their physiological arousal and subjective states during those situations, but that they were able to commit these experiences to memory and then accurately recall them. Consequently, deficits in emotional memory, rather than the subjective experience of emotion, could explain the findings. However, as the authors justifiably argue, on this account it is not clear why fear should be selectively impaired, while the other emotions remain intact. Second, given that it is difficult to accurately differentiate between sympathetic and parasympathetic nervous system activity even when using standardised psychophysiological recording techniques (for example, it is well-established that heart rate is subject to both sympathetic and parasympathetic influences), the argument that certain physiological phenomena map neatly onto sympathetic or parasympathetic activity should be treated with caution. Nevertheless, the study by Marsh et al. provides novel and important findings showing deficits in fear in those with psychopathic traits, demonstrates that they are already present in children and adolescents (and therefore could have consequences for socialisation processes), and highlights several issues worthy of investigation in future research, such as the relationship between impairments in subjective experience of emotion and recognition of affective states in others.

Correspondence to

Graeme Fairchild, School of Psychology, University of Southampton, Southampton, SO17 1BJ, United Kingdom. E-mail: