This week’s blog post is a little different. In part, it is a response to my students’ requests for a demonstration of the “critical thinking” I keep demanding from them. Fair enough. It is important for people to be as clear as possible about what they mean when they use ‘technical’ words. Accordingly, this post provides a partial critique of a recent paper relevant to this week on my altruism module. That paper, by Meffert et al. (2013), has been widely reported as showing that:
The central claim
Meffert and her colleagues claim to demonstrate a need for “a profound reconceptualization of psychopathy” (p. 2559). They note that psychopaths are usually portrayed as having “a profound lack of empathy” (p. 2550) due to their deficient brains making them incapable of caring about others’ welfare. Against this, Meffert et al. (2013) claim to show that, while psychopaths usually don’t care about others’ welfare, they can do so when motivated to try.
See how I can force myself to share your joy at this stunning news!
There are two main parts to Meffert et al.’s evidence.
The first they claim shows that, as commonly believed, non-psychopaths typically have more pronounced reactions to others’ experiences than do psychopaths. The second they claim shows that psychopaths and non-psychopaths have (more) similar reactions to others’ experiences when each is deliberately trying to empathise.
For key similarities and differences across groups to be meaningfully interpreted it is important to know how comparable the groups are. Meffert et al.’s paper does not make it easy to determine such things.
Almost nothing is said about who the control group were (e.g., members of the general public) or about how either the control group or the psychopathic (“target”) group were recruited.
In the “Participants” section on p. 2551, the paper says that the control participants were male and had the same handedness as members the target sample. (Many neuroimaging studies exclude left-handed people.) It also says that the target group and the control group had, on average, the same age and IQ. I could not find a report of how low or high that average IQ was.
Later in the paper, Meffert et al. reveal that the control group also differed from the target group on a range of “factors linked to criminal lifestyle (e.g., lower level of education, history of drug abuse and length of incarceration)” (p. 2558).
Given this, attributing found differences between the target and the control groups specifically to their differences in psychopathy levels seems somewhat bold, especially if they could plausibly stem from other differences between the groups.
As just seen, the term “psychopaths” does not capture all that was distinctive about Meffert’s target sample. As well as being diagnosed as psychopathic, the eighteen members of that sample were also Dutch, adult, male, poorly educated, right-handed, incarcerated criminals, the vast majority of whom had a recognised dependence on drugs. It seems at best questionable that this tiny and highly distinct sample is strongly representative of all people high in psychopathy. Even most media headlines reporting the study took care to be clear that the findings relate most obviously to criminal psychopaths.
Of course, if the study compellingly shows that any psychopaths (perhaps especially criminal ones) can care about others’ welfare (perhaps especially that of anonymous strangers), that may indeed suggest that a profound reconceptualization of psychopathy is needed.
Confirmation of a psychopathy difference across samples
To attribute other differences between the two groups to their difference in psychopathy, one clearly needs to be pretty sure that there are differences in psychopathy across groups. In Meffert et al.’s study members of the target sample were all diagnosed as psychopaths and members of the control group (as far as I can tell) were not. Nevertheless, it makes sense to check and Meffert et al. did. All participants completed a 187-item self-report measure of psychopathy (p. 2551).
As expected, the control group obtained a lower average score on this Psychopathic Personality Inventory (PPI) than did the target group: about 399 vs. 422 (Table 1, p. 2555), a statistically significant difference (p < .05, p. 2554).
If I had more time I would examine this result a lot more closely. Like many psychiatric diagnoses, “psychopathy” is somewhat an umbrella condition and two people or samples with identical scores might obtain those scores because of very different qualities. Moreover, Meffert’s control group psychoticism score of 399 is much higher than the average score of 268 obtained by some very scary criminals sampled in a study by Copestake et al. (2011).
For the sake of argument, though, let’s accept that Meffert et al.’s target group members were psychopaths (among other things) and that their control group members were not.
The “spontaneous empathy” task
All participants watched short video clips of two hands interacting. The hands interacted in several ways, two of which are more important than the others. In the “love” interactions an “approaching” hand stroked the back of someone else’s “receiving” hand, which gave a tender squeeze in response. In the “pain” interactions the approaching hand pulled back one of the receiving hand’s fingers to a point that would be painful for most people.
Participants in this spontaneous empathy task were given the rather strange instruction to view the film clips “as if they were watching one of their favourite videos” (p. 2552). Call me strange but I’d find it quite unsettling pretending that my favourite videos comprised largely of clips in which people were inexplicably deliberately hurting others!
The two groups had differing brain activation in “a large network of regions” (p. 2555). In each of these regions the control group showed higher brain activation than did the psychopath group (Table 2 and Supplementary eTable 2). Some of these Meffert et al. considered regions of particular interest because past research suggests that they are sometimes similarly activated both by personal experience and by observation of others having similar experiences (p. 2553). These included “the dorsal and ventral premotor cortex, primary and secondary somatosensory cortices, the anterior insula and the cingulate cortex” (p. 2555).
Accepting for the sake of argument that these findings and the authors’ interpretations of them are valid, “individuals with psychopathy showed … reduced activation” compared to control participants in brain areas associated with (among other things) vicarious experience, sensations, and emotions (p. 2558).
To report these results slightly differently, when instructed to watch short clips of interacting hands as if they were watching their favourite videos, poorly educated, incarcerated, drug-dependent, male criminals with relatively high levels of psychopathy appeared not to react as emotionally as did relatively well-educated, non-addicted, non-criminals with relatively low levels of psychopathy.
Maybe this indicates a difference in spontaneous empathy and maybe it results from a difference in levels of psychopathy across groups. Many other possibilities seem at least plausible.
My guess is that members of the target group were less motivated and anyway less able to comply with the task. I also doubt that this was primarily due to differences in psychopathy.
It seems at least possible that the control group members were more or less willing and able to comply with the instructions and they accordingly became relatively engrossed in the task of trying to conceive of the emotionally-evocative hand-theatre as high-drama. I suspect that the uneducated, drug-addled, career criminals would have really struggled (if they even tried) to successfully enter into the same spirit. If so, the brain activation differences across groups stemmed from the different engagement in the task due to different compliance with the viewing instructions, which itself stemmed mainly from the factors that happened in this study to be associated with psychopathy.
This is of course only my guess. The much more important point is that there is no compelling evidence which enables anyone to choose between my guess and Meffert et al.’s interpretation of their findings. The first part of their study may or may not tell us anything about psychopathy and if it does, there is no telling what, exactly. The two groups differed on too many potentially important characteristics, the instructions given to participants before the main task plausibly promoted different responses in each group, and no checks were made for differences between the groups in such things as task-engagement (e.g., how much participants tried and succeeded in viewing the clips as though they were their favourite movies), task-accuracy (i.e., how well each group recognised the emotional content of the viewed videos), subjective experience during the task (e.g., how emotionally aroused participants felt and what specific emotions were involved), or how they felt “on behalf of” the hands/people in the video clips (e.g., "empathic" joy, distress, anger, etc.).
In some ways, perhaps it doesn’t matter. In this part of the study, Meffert et al. are pushing at an open door. As they note themselves, the claim that psychopaths empathise less with others’ experiences than do non-psychopaths is the standard view – as well it might be because having a lack of empathy is part of the diagnosis for being a psychopath!
Let us therefore move on the second strand of evidence considered by Meffert et al.
The “deliberate empathy” task
Participants next watched a slightly different series of hand-interaction video clips, some showing nothing but “loving” interactions and some showing nothing but “painful” interactions. This time, participants received instructions to “feel with” one or other of the hands.
In this task, there were still many differences in brain activation across the groups (Supplementary eTable 3) but (a) it varied which group had the higher level of activation in a given area, (b) almost none of the areas of differential activation occurred within the researchers’ identified regions of interest (Table 3), and (c) there was less difference in brain activity between the target and control groups than had been the case in the “spontaneous empathy” condition.
Thus, in brain areas relevant to having similar experiences to those others seem to be having, psychopaths were not that different to non-psychopaths during this “deliberate empathy” task. The authors conclude that psychopaths are almost as able to experience vicarious emotions as non-psychopaths when each is “following instructions to empathize”.
Let’s phrase this conclusion a little more conservatively. If valid, this part of Meffert et al.’s study finds relatively little difference in brain activation when (druggy, criminal, etc.) psychopaths and (non-druggy, non-criminal, etc.) non-psychopaths are instructed to “feel with” someone else’s hand which is giving or receiving either a loving caress or a painful smack with a ruler.
Meffert et al. provide one possible explanation for this, i.e., that all participants were feeling love or pain as though they were actually experiencing caresses or smacks themselves. Another possibility is that no one was feeling very much, in either group. No measure was used to check this either way.
Being asked to “feel with” a hand that one is watching on a video is an odd instruction. Imagine being a participant in this study yourself. If you were told to “feel with” a videoed hand hitting someone else’s hand with a ruler, do you imagine that it would feel like your hand was swooshing down and reverberating with the shock of contact? Or do you imagine that you would try to work out what sorts of experiences an owner of such a hand might be having, perhaps by imagining how you might feel in such a situation? If like me you find the ‘working out’ possibility plausible, it seems at least possible that both groups were engaging in quite a lot of thinking about others’ feelings – without necessarily having comparable (“vicarious”) feelings themselves.
Such a possibility may seem particularly credible when we learn that during this task the target group had relatively high activity in “mentalising” areas of the brain, i.e., those thought to be involved when trying to work out what others are experiencing (p. 2559). It is widely acknowledged that psychopaths can be good at identifying others’emotions, all the better to exploit them for personal gain.
As in the first part of the study, Meffert et al. do too little to work out what participants in each condition are doing and experiencing to be confident about how best to interpret the differences in the brain activation they find – still less the differences they don’t find.
The most important of Meffert et al.’s findings for their key claim is that uneducated, criminal, drug-addicted psychopaths and educated, non-criminal, non-addicted, non-psychopaths show relatively similar levels of activation in brain regions associated with vicarious emotions when instructed to “feel with” one or other hand in short videos of hands interacting in apparently loving or painful ways.
One possible reason for this is that the uneducated, criminal, addicted psychopaths had subjective experiences that corresponded in some way to those portrayed in the videos (to about the same extent that anyone else would) and one possible and scientifically dramatic implication of this is that psychopaths can care about the welfare of others when they choose to do so.
Another possibility is that, regardless of levels of psychopathy, all participants in this part of the study were trying to work out what experiences others might be having and this did not markedly involve anyone actually having “vicarious” feelings (or being concerned about the welfare of anyone else having similar feelings). This possibility would do nothing whatsoever to threaten standard conceptions of psychopathy.
Given the multiple impediments to confidently interpreting Meffert et al.’s results, much more and much better evidence is needed before there is good reason to suppose that psychopaths are able to share or truly care about others’ emotional experiences.
Meffert et al. (2013) is much more complicated than I have let on here. I am not a neuropsychologist and I am not familiar with many of the methodological and statistical procedures reported in Meffert (2013). This means that you should be particularly sceptical of any of my claims above which might be hindered by these inadequacies on my part. On the other hand, there are problems with the paper that, because of time and presentation issues, I did not make explicit here (e.g., the potentially rather suspect practice of comparing brain activation in the “spontaneous” and the “deliberate” empathy parts of the study) and there may be other – perhaps more serious – problems with the paper that I am simply not skilled enough to notice. If anyone would like to correct my claims or make new ones in the comments section below, your contribution to my and my students’ education will be most welcome.
References and further reading
Chambers, C. (2013, Nov 25). Could a brain scan diagnose you as a psychopath? The Guardian [Link]
Copestake, S., Gray, N. S., & Snowden, R. J. (2011). A comparison of a self-report measure of psychopathy with the psychopathy checklist-revised in a UK sample of offenders. Journal of Forensic Psychiatry & Psychology 22(2): 169–182.
Jarrett, C. (2013, Dec 13). A calm look at the most hyped concept in neuroscience – Mirror neurons. Wired. [Link]
Jarrett, C. (2012, Dec 10). Mirror neurons: The most hyped concept in neuroscience? Psychology Today [Link]
Lamm, C., & Majdandžić, J. (2014). The role of shared neural activations, mirror neurons, and morality in empathy-a critical comment. Neuroscience Research. [Link]
Neuroskeptic (2014, July 21). Functional neuroimaging’s Neymar problem. Discover. [Link]
Neuroskeptic (2014, June 6). fMRI motion correction: The quick and the dead. Discover. [Link]
Neuroskeptic (2013, May 13). Looking Askance At Cognitive Neuroscience. Discover. [Link]
Neuroskeptic (2013, April 27). The (sigh) psychopath brain. Discover. [Link]
Neuroskeptic (2012, April 25). Does brain scanning show just the tip of the iceberg? Discover. [Link]
Neuroskeptic (2009, Sept 16). fMRI gets slap in the face with a dead fish. Discover. [Link]
Vincent, J. (2013, July 26). An 'empathy switch' allows psychopaths to feel at will. Discover. The Independent. [Link]
Wikihow. How to identify a psychopath. [Link]
Psychotic empathy doll [Link]
Forced smile [Link]
Sherlock Holmes [Link]
Snakes in suits [Link]
How to cite this blog post using APA Style
T. Farsides. (2014, November 17). Psychopathic empathy. Retrieved from http://tomfarsides.blogspot.com/2014/11/psychopathic-empathy.html