Differences in self- and other-induced Mu suppression are correlated with empathic abilities

Abstract

Recent research suggests a role of the human mirror neuron system in empathic processing. Electroencephalographic (EEG) mu suppression in the 8–13 Hz band-range has been proposed to reflect mirror-like activation of sensorimotor cortices. We therefore investigated the relationship between suppression in the 8–13 Hz range and empathic abilities as measured by the Interpersonal Reactivity Index (IRI). Based on previous work showing significant differences between execution- and observation-related mu suppression and on theories of empathy pointing to the importance of discriminating self from other, we predicted that the size of this difference within individuals would correlate with participants' perspective-taking abilities. Larger differences in execution- and observation-induced mu suppression were associated with greater scores on the perspective-taking subscale of the IRI. The current data are therefore consistent with the claim that neural activity reflected in mu suppression is related to empathic abilities.

Introduction

Gastaut and Bert (1954) were among the first to report suppression of activity in the 8–13 Hz bandrange of the electroencephalogram (EEG) in response to cinematographic images. Since the discovery in macaque monkeys of Mirror Neurons (MN) by Rizzolatti and colleagues (di Pellegrino et al., 1992), some EEG researchers have theorized that so-called mu suppression, maximal over central scalp electrodes, may be modulated by MN input into sensorimotor cortices (Cochin et al., 1999, Holz et al., 2008, Keuken et al., 2011, Lepage and Theoret, 2006, Muthukumaraswamy and Johnson, 2004, Oberman et al., 2005, Perry et al., 2010, Perry and Bentin, 2009, Pfurtscheller et al., 1997, Pineda, 2005, Pineda and Hecht, 2009, Ullola and Pineda, 2007, Woodruff and Maaske, 2010). Activating both when one performs a specific action and when he watches another perform the same action, MN's appear to be a mechanism whereby visual representations of a conspecific's behaviors are efficiently translated into motor commands in the observer's motor system. This notion, that another person's intentions are efficiently transposed into similar intentions in the observer's motor system, clearly has potential relevance to social functions such as perspective-taking (PT). As Batson and others (Batson et al., 1991, Decety and Jackson, 2004, Lamm et al., 2007) point out however, simply experiencing, or simulating, another's intentions is likely insufficient to produce empathic concern. Rather, beyond simulation, one must recognize that the perspective he is taking is not his own but belongs to another—a process of self-other (SO) discrimination (Davis, 1983).

It follows from this that one could assess the relevance of given neural processes to PT by measuring the extent to which the characteristics of these processes reflect simulation, PT, or both. Much research has established that mu suppression reflects simulation (e.g. Cochin et al., 1999, Lepage and Theoret, 2006, Muthukumaraswamy and Johnson, 2004) in that mu suppression to observation of movements is similar to that during execution of movements. Whether mu suppression reflects SO discrimination is an open question however. By definition, mirroring processes are ones that respond both to self- and other-initiated actions. Mirroring therefore would appear to be a blurring of the distinction between self and other rather than discrimination. On the other hand, some argue that MNs play an important role in self-processes (e.g. Uddin et al., 2007), which entails a discrimination between self and other. We argue it is possible therefore that, in addition to being involved in simulation, mirroring processes reflected in mu suppression may be sensitive to SO discrimination in a way that is correlated with PT abilities. The purpose of this study was to test the hypothesis that mu suppression not only differs between self and other, but also the differentiation is correlated with PT abilities.

Some have pointed out that empathy involves various subcomponents, of which there are PT, empathic concern and personal distress to name a few (Davis, 1983, Decety and Jackson, 2004, Batson et al., 1991). Others have suggested that empathy additionally involves processes that lead to helping behavior (see Batson, 2009). While experiencing the intentions and affect of an observed person are necessary for empathy, they are not sufficient alone (Decety and Jackson, 2004, Decety and Lamm, 2007, Lamm et al., 2007). This is true because empathizing necessarily involves a target (Preston and De Waal, 2002). According to Batson's Empathy–Altruism hypothesis, simply observing another's situation, without successfully adopting the other person's perspective, leads to personal distress rather than to empathic concern (Batson et al., 1991). Empathy can occur once the observer becomes aware that the source of first-person thoughts and/or feelings is the other person (Decety and Lamm, 2009a, Decety and Lamm, 2009b; Batson, 1991).

Cognitive neuroimaging has contributed to the investigation of the neural processes underlying empathy. One example comes from Schülte-Ruther and colleagues (2007) in which they found the amount of hemodynamic signal change in premotor cortex was significantly correlated with empathic abilities. Using functional Magnetic Resonance Imaging (fMRI), they found that hemodynamic responses from inferior frontal cortex co-varied with results of the BEES (Balanced Emotional Empathy Scale; Mehrabian, 1997). This correlation was found only when the participant was required to process the emotional states of faces, in contrast to a control condition, in which participants simply reported age and gender.

Similarly, some EEG experiments have reported relationships between EEG mu suppression and sub-components of empathy. Mu rhythms oscillate within the alpha range of 8–13 Hz and are traditionally found to be highest in power in central electrodes (Cochin et al., 1999, Gastaut and Bert, 1954). Unlike traditional alpha rhythms, which are maximal with eyes closed and suppressed with eyes open, mu rhythms are maximal when an individual is motionless and suppressed when he makes a movement. Because mu suppression is elicited by conditions similar to those that elicit MN activity at the single-cell level in macaques, and because it is believed to result from premotor cortex input into sensorimotor regions, it is argued that they are a reliable measure of MN activity in humans (Pineda, 2005).

Mu suppression has been shown to have at least a limited relationship to empathy. Yang et al. (2009) found a positive correlation between event-related desynchronization of mu rhythms and the personal distress subscale of the Interpersonal Reactivity Index (IRI) for females, but not for males. Increased mu suppression was associated with increased personal distress in response to a video of a hand about to be cut with a pair of scissors. Interestingly, Yang et al. reported no correlations with any of the other three subscales of the IRI (perspective taking, empathic concern, fantasy).

Similar to Yang et al., 2009, Perry et al., 2010 found no correlations with other subscales of the IRI (or with the personal distress subscale for that matter). Perry et al. used animated point light displays of walking human figures and found significant mu suppression relative to a baseline that involved point lights making the same motions but randomly dispersed. While they found no significant correlations with the IRI subscales, they did report significant correlations between mu suppression and the Empathy Quotient (EQ; Baron-Cohen and Wheelright, 2004). This correlation was statistically positive. But, because log ratios were used, smaller (more negative) scores actually meant more mu suppression. Therefore, the relationship was such that more mu suppression was associated with smaller EQ scores. This negative relationship is somewhat surprising as one would expect that mu suppression reflecting mirroring processes would bear a positive relationship with empathic abilities. It might be considered surprising as well that no correlations obtained with any of the IRI subscales.

Many other EEG mu suppression studies have helped to further characterize the nature of this electrophysiological signal (e.g. Cochin et al., 1998, Cochin et al., 1999, Muthukumaraswamy and Johnson, 2004, Lepage et al., 2008 Perry and Bentin, 2009, Ullola and Pineda, 2007, Oberman et al., 2005, Oberman et al., 2008). The results of each of these studies have been broadly consistent with the claim that mu suppression reflects mirror neuron activity. One task strikingly simple, yet effective at eliciting mu suppression comes from Oberman et al. (2005). This task required participants to alternately execute an opening and closing action of the hand and to observe videos of another hand performing the same action. Woodruff and Maaske (2010) used the task to demonstrate reliably greater mu suppression for execution than for observation. This difference was taken to reflect a similarity to a difference found by Rizzolatti and colleagues in macaque monkeys (di Pellegrino et al., 1992). These data were taken to support the hypothesis that mu suppression reflects neural mirroring response characteristics similar to those seen in macaque (Fig. 1).

Differences in self- and other-induced desynchronization raise the possibility that mu suppression's relationship to empathy might be dependent on the extent to which that activity differs for execution and observation. Given claims that empathy requires the kind of SO discrimination that is involved in PT (Batson et al., 1991, Decety and Jackson, 2004, Decety and Lamm, 2009a, Decety and Lamm, 2009b, Lamm et al., 2007), the differential mu suppression observed by Woodruff and Maaske (2010) might serve as an indicator of the extent to which an individual is able to distinguish his actions from the actions of others. Many theorize that mirroring processes play an important role in empathy (Gallese, 2005, Goldman and de Vignemont, 2009, Iacoboni, 2008, Singer et al., 2006). Yet, Decety (2005) suggests that perception–action coupling mechanisms may not be sufficient for empathy as it may lack the necessary perspective-taking component. This assumes that mirroring processes do not differentiate between self- and other-related processing in a way that is relevant to empathy. If, however, mirroring processes do play a central role in generating empathy, then a perspective-taking-related difference in self- and other-induced mu suppression should be observable. The first part of this statement is critical. Woodruff and Maaske (2010) demonstrated that mu suppression differs for self and other. But it is possible that the variance associated with this difference is unrelated to empathy. An important question therefore is do differences in self- and other-induced mu suppression correlate with PT?

In the current study, we sought to address this question by adding an empathy self-report measure to the task employed by Woodruff and Maaske (2010). Our main hypothesis was that subscales of the IRI would correlate with execution minus observation mu suppression difference scores. More specifically, we expected that subcomponents of empathy that require SO differentiation (e.g. perspective-taking), to the extent these subcomponents are subserved by neural processes reflected in mu suppression, would correlate with the degree to which mu suppression differs between self and other.