Lindquist et al. demonstrate that there is little evidence for a locationist approach to emotion. We, too, think that the evidence points toward psychological constructionism, the view that emotions of a certain kind are constructed out of more general brain structures whose function is not specific to emotions of that kind, or even to emotions at all. There is not much evidence that emotions are localized in specific, unique regions of the brain dedicated to a particular kind of emotion alone and no other cognitive function. There is even less evidence for the modularity that has often accompanied the localization hypothesis.
This commentary aims to add two things. First, we suggest that there are other routes to the same constructionist conclusion. We sketch two. They are complementary to the meta-analyses of Lindquist et al. and have the potential to add a considerable amount of new evidence. Second, we look at what is perhaps the most challenging type of case for constructionism, namely, the existence of emotional anomalies where there is evidence for specific, highly correlated, and probably relevant brain anomalies. We have chosen psychopaths as our example.
One additional route to evidence concerning constructionism would be to look at “task specifics,” that is, the specifics of the context of individual occurrences of a kind of emotion. Consider anger. Anger can be produced by many input modes: seeing, hearing, recalling, being told. Some situations leading to anger are more difficult to understand than others (i.e., they impose a greater cognitive load; Lindquist et al. touch on this issue in section 5.4, para. 4). At the output end, the manner in which anger is expressed varies enormously from case to case, especially across cultures, genders, and socio-economic groups (Ekman Reference Ekman1993; Cheng et al. Reference Cheng, Mallinckrodt and Wu2005). Sometimes anger expresses itself in physical aggression, sometimes in verbal aggression, sometimes in contempt (of which there are in turn many forms), sometimes in passivity, sometimes in deliberately blunted affect (“coldness”). (Lindquist et al. touch on this as well [sect. 6.1, para. 3], but do nothing with the observation. Indeed, it is not even clear whether for them expression is part of emotion.) Therefore, our question is: Since these different ways of expressing anger are almost certainly associated with widely different regions in the brain, are there any regions or sub-regions activated throughout the different kinds and contexts of anger? If no region or sub-region is activated across all the differences of input, cognitive demand, and output, that would be important further evidence for constructionism.
One could use task specifics in a number of other ways. For example, one could look at cases where expression is roughly constant across occurrences but underlain by very different emotions. Or one could try to factor out differences in brain activation associated with different reasons for a given kind of emotion, or differences resulting from personality type or psychological disorder.
Another way to generate data relevant to constructionism would be to look at cases in which a stimulus generates a judgment about an emotion rather than the emotion itself; for example, where a face is interpreted as expressing anger but does not make the subject feel angry. In the relatively unemotional judgment about the stimulus, are the same areas activated as when the subject actually feels angry?
In the same vein, there are cases where we would expect a face to be interpreted as expressing anger, but some subjects interpret it instead as expressing pain. Would the same areas be activated as when subjects interpreted faces in line with our expectations as expressing anger? And what about activation associated with differences of experimental protocol (reaction time or decoding, masking or no masking, variations in the emotional intensity of the stimulus, and/or differences among different kinds of stimulus set, visual modality [seeing vs. hearing], and level of cognitive load)? The research questions behind task specific studies are: (1) Concerning a given emotion, what are the activation patterns for different patterns of vision, cognition, and expression? And, (2) Do these patterns lend any support to the localization hypothesis?
Psychopaths, in whom striking emotional anomalies are strongly correlated with specific brain anomalies, appear to challenge constructionism. The emotional anomalies associated with psychopathy include callousness or lack of empathy, lack of remorse and guilt, and shallow affect (Cleckley Reference Cleckley1982; Hare Reference Hare1991). Turning to task specifics, psychopaths often react differently from non-psychopaths to emotionally laden stimuli. The emotional anomalies in psychopaths are usually specific to fear (Blair et al. Reference Blair, Colledge, Murray and Mitchell2001). They show little startle reaction in fear situations (Patrick et al. Reference Patrick, Bradley and Lang1993) and poor decoding of fearful facial expressions (Blair et al. Reference Blair, Mitchell, Peschardt, Colledge, Leonard, Shine, Murray and Perrett2004) in comparison to non-psychopaths. They also present with a distorted ability to allocate attention in a way that reflects the emotional content of stimuli (Lorenz & Newman Reference Lorenz and Newman2002; Newman et al. Reference Newman, Curtin, Bertsch and Baskin-Sommers2010). (The latter already points to constructionism.)
What about the underlying, correlated brain activation? While the anomalies suggest amygdala dysfunction and there is indeed evidence for this (Blair Reference Blair2005; Reference Blair2008), the behavioral evidence strongly supports constructionism – indeed, more than Lindquist et al. fully recognize. They associate the psychopath's aggression with anger (sect. 5.3, para. 1). In fact, there is far more variability in psychopaths than this. In psychopaths, aggression tends to play an instrumental role as opposed to the reactive role that it usually plays in non-psychopaths (Hart & Hare Reference Hart and Hare1996). Aggression appears to be how psychopaths respond to a wide variety of situations about which they have a wide variety of feelings. If this is true, aggression in them is likely to be associated with a wide range of emotions, not just anger. This conclusion supports constructionism.
Thus, not only does psychopathy not threaten psychological constructionism, it offers support for the view. We expect that the same will be true for other syndromes that have seemed to some to offer evidence for localization.
Lindquist et al. demonstrate that there is little evidence for a locationist approach to emotion. We, too, think that the evidence points toward psychological constructionism, the view that emotions of a certain kind are constructed out of more general brain structures whose function is not specific to emotions of that kind, or even to emotions at all. There is not much evidence that emotions are localized in specific, unique regions of the brain dedicated to a particular kind of emotion alone and no other cognitive function. There is even less evidence for the modularity that has often accompanied the localization hypothesis.
This commentary aims to add two things. First, we suggest that there are other routes to the same constructionist conclusion. We sketch two. They are complementary to the meta-analyses of Lindquist et al. and have the potential to add a considerable amount of new evidence. Second, we look at what is perhaps the most challenging type of case for constructionism, namely, the existence of emotional anomalies where there is evidence for specific, highly correlated, and probably relevant brain anomalies. We have chosen psychopaths as our example.
One additional route to evidence concerning constructionism would be to look at “task specifics,” that is, the specifics of the context of individual occurrences of a kind of emotion. Consider anger. Anger can be produced by many input modes: seeing, hearing, recalling, being told. Some situations leading to anger are more difficult to understand than others (i.e., they impose a greater cognitive load; Lindquist et al. touch on this issue in section 5.4, para. 4). At the output end, the manner in which anger is expressed varies enormously from case to case, especially across cultures, genders, and socio-economic groups (Ekman Reference Ekman1993; Cheng et al. Reference Cheng, Mallinckrodt and Wu2005). Sometimes anger expresses itself in physical aggression, sometimes in verbal aggression, sometimes in contempt (of which there are in turn many forms), sometimes in passivity, sometimes in deliberately blunted affect (“coldness”). (Lindquist et al. touch on this as well [sect. 6.1, para. 3], but do nothing with the observation. Indeed, it is not even clear whether for them expression is part of emotion.) Therefore, our question is: Since these different ways of expressing anger are almost certainly associated with widely different regions in the brain, are there any regions or sub-regions activated throughout the different kinds and contexts of anger? If no region or sub-region is activated across all the differences of input, cognitive demand, and output, that would be important further evidence for constructionism.
One could use task specifics in a number of other ways. For example, one could look at cases where expression is roughly constant across occurrences but underlain by very different emotions. Or one could try to factor out differences in brain activation associated with different reasons for a given kind of emotion, or differences resulting from personality type or psychological disorder.
Another way to generate data relevant to constructionism would be to look at cases in which a stimulus generates a judgment about an emotion rather than the emotion itself; for example, where a face is interpreted as expressing anger but does not make the subject feel angry. In the relatively unemotional judgment about the stimulus, are the same areas activated as when the subject actually feels angry?
In the same vein, there are cases where we would expect a face to be interpreted as expressing anger, but some subjects interpret it instead as expressing pain. Would the same areas be activated as when subjects interpreted faces in line with our expectations as expressing anger? And what about activation associated with differences of experimental protocol (reaction time or decoding, masking or no masking, variations in the emotional intensity of the stimulus, and/or differences among different kinds of stimulus set, visual modality [seeing vs. hearing], and level of cognitive load)? The research questions behind task specific studies are: (1) Concerning a given emotion, what are the activation patterns for different patterns of vision, cognition, and expression? And, (2) Do these patterns lend any support to the localization hypothesis?
Psychopaths, in whom striking emotional anomalies are strongly correlated with specific brain anomalies, appear to challenge constructionism. The emotional anomalies associated with psychopathy include callousness or lack of empathy, lack of remorse and guilt, and shallow affect (Cleckley Reference Cleckley1982; Hare Reference Hare1991). Turning to task specifics, psychopaths often react differently from non-psychopaths to emotionally laden stimuli. The emotional anomalies in psychopaths are usually specific to fear (Blair et al. Reference Blair, Colledge, Murray and Mitchell2001). They show little startle reaction in fear situations (Patrick et al. Reference Patrick, Bradley and Lang1993) and poor decoding of fearful facial expressions (Blair et al. Reference Blair, Mitchell, Peschardt, Colledge, Leonard, Shine, Murray and Perrett2004) in comparison to non-psychopaths. They also present with a distorted ability to allocate attention in a way that reflects the emotional content of stimuli (Lorenz & Newman Reference Lorenz and Newman2002; Newman et al. Reference Newman, Curtin, Bertsch and Baskin-Sommers2010). (The latter already points to constructionism.)
What about the underlying, correlated brain activation? While the anomalies suggest amygdala dysfunction and there is indeed evidence for this (Blair Reference Blair2005; Reference Blair2008), the behavioral evidence strongly supports constructionism – indeed, more than Lindquist et al. fully recognize. They associate the psychopath's aggression with anger (sect. 5.3, para. 1). In fact, there is far more variability in psychopaths than this. In psychopaths, aggression tends to play an instrumental role as opposed to the reactive role that it usually plays in non-psychopaths (Hart & Hare Reference Hart and Hare1996). Aggression appears to be how psychopaths respond to a wide variety of situations about which they have a wide variety of feelings. If this is true, aggression in them is likely to be associated with a wide range of emotions, not just anger. This conclusion supports constructionism.
Thus, not only does psychopathy not threaten psychological constructionism, it offers support for the view. We expect that the same will be true for other syndromes that have seemed to some to offer evidence for localization.