Empathy is central to the human experience andthus has received much attention in the field of neuroscience. Empathy is generally studied in the context of vicarious rewardand punishment – one person processing someone else’s reward or punishment byunderstanding or feeling events from the other person’s perspective. There is currently a debate in neuroscience about whether humans’capability to learn by observation and experience vicarious rewards andpunishments relies on neural mechanisms similar to those employed when humanslearn based on personal experience.
The“extended common currency schema” school of thought would support this idea,contending that a neural circuit integrates information from all factors (i.e. social and non-social factors) relevant to achoice or experience. The“social-valuation-specific schema” provides an alternative explanation as tohow neural value is computed in social versus non-social contexts, proposingthat social and non-social information are processed via similarly computationprinciples but are implemented in distinct regions of the brain, specialized foreach type of information (Ruff & Fehr, 2014). Severalstudies support the idea that certain regions of the brain are involved in thevaluation of both non-social, personal experiences as well as socialexperiences concerning the observation of others.
In onesuch study, while being scanned in an fMRI, participants viewed contestantsplay a game in which they had to guess whether an unseen card was higher orlower than a second unseen card and would win money based on their performance. After viewing the contestants play the game, the participantsplayed the game for themselves. Resultsindicated that there were similar increases in ventral striatum activation forwhen participants observed others win rewards and for when participants wonrewards themselves. Additionally, perceived similarity between theparticipants and observed contestants modulated this value-related neuralactivity in the ventral striatum, with increased similarity elicitingheightened activity (Mobbs et al., 2009). Otherstudies have supported this finding that the ventral striatum integrates socialinformation and reward information.
Forexample, researchers conducted a study in which participants played a gamblingtask guessing the outcome of a coin flip. Theparticipants could either win or lose money for themselves, their best friend,or a disliked other – another “participant” who the researchers manipulated toplay an unfair strategy in an Ultimatum Game conducted prior to the gamblingtask. The only brain region that showed aninteraction between the recipient of the gambling task’s outcome and theoutcome itself was the ventral striatum. Therewas greater activation in the ventral striatum following gains than losses whenthe participants and their friends’ received the outcome. Conversely, the ventral striatum was more active following lossesthan gains when the disliked other received the outcome.
Notably, the dorsal medial prefrontal cortex (dmPFC) and thetemporal parietal junction (TPJ), two regions of the brain associated withsocial cognition, encoded information solely about the recipient of the outcome. In these regions, there wasgreater activity when the friend and the disliked other received the outcomethan when the participants received the outcome, regardless of the outcomeitself (Braams et al., 2014).
Similarto how the ventral striatum is associated with processing rewards to self and vicariousrewards, the anterior cingulate cortex (ACC) and the insula are associated withprocessing punishments to self and vicarious punishments. In one study, female participants in fMRI scanners sat next totheir partners and could see their partners’ right hand through a mirror system. Both partners in the couple had an electrode attached to theirhands that would deliver painful stimulations.
Randomcues on a screen indicated whether the female (self) or her partner (other)would receive a stimulation that was either painful or not painful. Experiencing the painful stimulation and observing their partnerexperience the painful stimulation both triggered increased neural activity inthe ACC and insula. Importantly, activity the brain regionsassociated with the sensory components of pain, such as in the sensorimotorcortex, only increased when the female participants themselves received thestimulation. This suggests that only the parts of the “painnetwork” associated with affective qualities – the ACC and insula – and notthose associated with sensory qualities mediate the experience of vicariouspunishment (Singer et al., 2004).
Thisresult regarding vicarious punishment also held up when “punishments” involvedthe smelling of disgusting odors. Whenparticipants passively viewed movies of people smelling the content of a glassand reacting with disgust and when participants inhaled disgusting odors, therewas increased activation in the anterior insula and ACC. This result indicates that at least for the feeling of disgust,the neural substrate of feeling the emotion and perceiving it in others issimilar (Wicker et al., 2003). Atfirst, these results for the ventral striatum and ACC and insula taken togethermay seem to support the idea that certain structures in the brain represent acommon currency that guides decision-making by integrating the motivationalrelevance of all possible stimuli or actions present in a situation.
This is because these parts of the brain encode bothreward-related and punishment-related information for both the self and others. In the context of non-social decisions, the common neural currencyin the brain is thought to be automatic and context-invariant. For example, in one study examining preference formation, during fMRIscanning, participants were asked to rate either the pleasantness (explicittask) or the age (distractive task) of images from various categories, such asfaces, houses, and paintings. Afterscanning, participants were presented with the same images in pairs and wereinstructed to choose which image they preferred. Resultsindicated that the brains regions in the limbic fronto-striatal circuit encodedparticipants’ preferences. Thiseffect was found across all three image types, suggesting context-invariance,and for both the explicit pleasantness rating task and the distractive agerating task, suggesting automaticity (Lebreton et al., 2009).However, in the context of social cognition, context-invariance was notobserved, as value representations of vicarious rewards in the ventral striatumvaried depending on the perceived characteristics of others.
Activity in the ventral striatum was modulated by perceivedsimilarity to oneself (Mobbs et al., 2009), and friendship and priorfair behavior (Braams et al., 2014). Thisfinding seems to call into question the validity of a common neural currencyincorporating social factors into reward representation. Eventhough the ventral striatum, ACC, and insula are active similarly for personalexperiences and vicarious experiences, there are regions of the brain that areactive during vicarious experiences that do not process rewards or punishments.
This suggests that observational learning may not solely rely onneural mechanisms that process the valuation of personal experiences. As previously mentioned, when participants played a coin flipgambling task and won or lost money for themselves, their friend, or a dislikedother, the dmPFC and TPJ did not encode information about the outcome of task. Instead, there was only increased activity in these two regionswhen the recipient of the outcome was the friend or the disliked other,suggesting that these regions only process the social aspects of vicariousreward (Braams et al., 2014). The TPJhas also been implicated to have a unique role in the processing ofagent-specific information when predicting socially guided decisions. In one study, participants played a simplified poker game againsteither human or computer opponents. Foreach trial of the game, participants received either a high or low card andcould choose to either bet or fold. If theychose to bet and their opponent called, the participants would win if they hadthe high card and lose if they had the low card.
Participantscould also win money if they bet while having the low card and their opponentfolded. The researchers examined the uniquecombinatorial performance (UCP) of 55 regions of the brain to determine whethera particular region of the brain contributed to participants’ future behaviormore than all other regions of the brain. Theiranalysis revealed that the TPJ distinctly carried social information comparedto all other regions, as its UCP against the human opponent was much greaterthan its UCP against the computer opponent.
Thus,they concluded that the TPJ uniquely preferentially processed socialinformation, further supporting the idea of social-valuation-specific regionsof brain (Carter et al., 2012). A recentmeta-analysis of 25 functional neuroimaging studies further supports the ideathat certain brain regions are engaged in the processing of vicarious rewardsdue to their involvement in social cognition, but not in the processing ofpersonal rewards.
This study found that regions related tomentalizing – the act of inferring others’ mental states – are preferentiallyassociated with vicarious rewards. Theseregions include the TPJ and dmPFC. Theprimary explanation for this effect is that the vicarious sharing of others’rewards typically requires someone to be able to understand “the extent towhich others value a particular outcome”. Inorder for this understanding to occur, especially when an observer’spreferences differ from a social target’s, one must project his or herselfoutside their present circumstances and infer the preferences of others (Morelli, Sacchet, & Zaki, 2015).
Finally,recent single-unit neural recoding studies in non-human primates have begun toidentify neurons in the ventral striatum that selectively encode either social ornon-social aspects of rewards. Onestudy employed a two-alternative forced-choice while recording the firing ratesof individual neurons in macaques’ striatum. Themonkeys indicated their choices with saccades to one of two visual targets. One target yielded a juice reward followed by a picture of amonkey or a gray square while the other target only yielded a juice reward. Results indicated that certain striatal neurons only fired forsocial information – the image of another monkey – while other neurons onlyfired for reward information – receiving the juice (Klein & Platt, 2013). Whileit is not feasible to conduct similar single-unit recording studies in humans,it is possible that the ventral striatum in humans has neurons that are alsoselective for either reward or social information.
Thisdifferential selectivity would not be revealed in fMRI imaging studies due totheir lack of adequate spatial resolution but may explain why the ventralstriatum is active for both social and non-social reward processing. On abroader level, this debate about the existence of a common neural mechanism tounderstand others’ experience of reward and punishment is fundamentally aquestion of what it means when one brain region similarly activates fordifferent actions. For example, when the same brain activates forboth personal and vicarious experiences, does it mean the ventral striatum isactually representing the vicarious experience as if it was a personal experienceor is it merely representing what is happening in the given social scenario? Inthe context of vicarious punishment, a brain imaging study is hardly needed toconclude that observing a painful shock is different from actually experiencinga painful shock, even though both actions may activate the ACC and insula.
This is because experiencing pain is an affective experienceinvolving sensorimotor context. However, for vicarious reward, the answer tothis question is not immediately clear and warrants future research beforewider conclusions can be made about what similar activation patterns in theventral striatum represent. Aworthwhile direction for future research would be examining whether the twotypes of reward processing actually influence choice behavior and learning insimilar manners.
This is because currently, “similaritiesbetween social and non-social reward processing and valuation are oftenproposed based exclusively on reverse inference” (Ruff & Fehr, 2014). Thereare few insights into how the brain treats rewards involving non-social andsocial aspects in the classic biological sense. Basedon this lack of clarity presently, I think it would be misguided to assert thata common neural currency exists that integrates non-social and social rewardinformation, especially the given the findings that implicate other brainregions encoding social information but not reward information.