Abacus experts in Japan

In East Asian countries, it is quite customary, at least traditionally, for children to learn how to use the abacus. Still today, abacus is often taught in China, Taiwan and Japan as part of arithmetic classes in elementary schools. Abacus users learn to move an array of beads to represent numbers, perform arithmetic operations on them and thus to generate answers. Experts can perform complex operations with apparent ease. There is a small literature in cognitive and developmental psychology examining what might be happening in abacus experts as they solve arithmetic problems. Early on, a Japanese developmental psychologist, Giyoo Hatano, proposed that abacus experts acquire a mental representation of abacus and operate on the mental abacus (Hatano and Osawa, 1983; see also Stigler, 1994). One implication of this idea is that abacus experts represent numbers spatially in terms of locations of relevant beads on the mental abacus (Hatano et al., 1987).

Recently, Hanakawa et al. (2003) conducted an fMRI study, investigating some neural implications of the mental abacus hypothesis: if experts use a mental abacus in mental computation, they should engage parietal regions of the brain because they are linked to spatio-visual processing (Mellet et al., 1998). The researchers asked both abacus masters and novices (all Japanese) to solve a variety of arithmetic problems and observed patterns of brain activations. Their data suggest that novices tend to show activations in motor cortices as well as areas involved in linguistic processing (e.g. Broca’s area) during mental computation. However, for intermediate abacus users, a much more prominent activation was found in the left parietal lobe and, moreover, for abacus masters this parietal activation was found bi-laterally. One additional and crucial piece of data came from an analysis of how brain activations might increase as a function of difficulty of mental computation. The parietal activation was systematically greater as a function of the number of digits involved in the mental computation, thus indicating the crucial involvement of the parietal lobe in arithmetic processing.

Spatial navigation and the hippocampus

In all animals including humans, spatial memory and navigation is a skill that is essential for survival. One area of the brain that plays a fundamental role in this regard is the hippocampus (O'Keefe and Nadel, 1978). The hippocampus is a seahorse-shaped structure that forms part of the limbic system that surrounds the thalamus.

In humans, located inside of the medial temporal lobe, the hippocampus has long been known to serve a critical function in memory formation and retrieval (Squire and Schacter, 2002). People with extensive hippocampal damage are known to show substantial memory loss (Scoville and Milner, 1957). Recent work by Schacter and colleagues (Schacter et al., 2007; Schacter and Addis, 2008) has shown that the hippocampus is part of the system that forms and retrieves episodic memory representations as well as crafting representations of future episodic events. Furthermore, the hippocampus is known to retain the ability to produce new neurons throughout much of the life span (Erikssonet al., 1998). These neurons are likely to form new connections, which in turn could be culled and pruned out, to modify neural pathways.

Spatial navigation, especially, navigation of the self in a complex, everchanging environment, would require retrieving previous episodes, simulating possible future courses of action, and then deciding which routes to take to reach given destinations. It may not be a coincidence, then, that the hippocampus is also heavily implicated in spatial navigation. Numerous studies with rodents have identified neurons that respond specifically to particular places in space. These neurons, called place cells, are located primarily in the hippocampal areas (O'Keefe and Nadel, 1978). As may be expected, one neuroimaging study finds that while navigating the self in a virtual environment, subjects exhibit enhanced activations in the right hippocampus as well as in the right inferior parietal regions, again demonstrating the crucial role of the hippocampus in spatial navigation (Maguire et al., 1998).

Another study by Maguire and colleagues has demonstrated that extensive ‘training’ in spatial navigation can result in structural changes in the hippocampus (Maguire et al., 2000). The researchers tested London cab drivers who varied widely in the number of years of experience. Like in many other major European cities, London streets are highly complex. Moreover, cab driving, unlike, say, driving of a bus on regular routes, requires considerable ‘improvisations’ in finding short cuts, avoiding traffic jams, changing routes as a function of the day of the week, and the like. Driving as a cab driver in big, complex cities, then, would require continuously retrieving relevant episodic memories, simulating possible routes, and operating on the mental map of the place while making numerous driving decisions. In short, the cab drivers may constantly make heavy use of the navigating functions of their hippocampi.

To assess the validity of this analysis, the researchers obtained structural magnetic resonance images of the brains of the London cab drivers and analyzed the volume of the anterior, the middle, and the posterior parts of the hippocampus and observed that, relative to the hippocampi of matched control subjects, the hippocampi of the cab drivers showed a substantial enlargement in the posterior part and an equally noticeable, although unanticipated and currently somewhat puzzling, reduction in size in the anterior part. This structural difference, moreover, was attributable to experience because it was significantly correlated with the number of years of experience as a cab driver.

Arithmetic processing in China

Another important demonstration of malleability of brain functions as a function of experience comes from a recent study by Tang and colleagues on arithmetic processing (2006). It has long been known that when one mentally solves simple arithmetic problems, say, ‘12 + 32’, linguistic processing areas of the brain including the Broca’s and Wernicke’s areas are strongly engaged (Dehaene and Cohen, 1995; Dehaene, et al., 1999). It appears that individuals use linguistic codes in carrying out mental computation. In fact, the aforementioned study by Hanakawa and colleagues on abacus experts showed that novice Japanese do show a pattern that is consistent with this general conclusion.

Building on this previous work, Tang and colleagues investigated patterns of brain activation linked to mental computation among European Americans and Chinese and found strong evidence for the involvement of linguistic processing areas in mental computation. Interestingly, however, the evidence was at best very weak for the Chinese, who instead showed a strong activation in the pre-motor cortex during mental computation—another finding consistent with the data from Japanese abacus novices. Tang and colleagues attribute this surprising association of mental computation with the brain regions involved in motor responses to the fact that during training of Chinese language, a strong emphasis is placed on writing. The argument is that Chinese also use linguistic processing in mental computation, but the linguistic processing implicates motor movement much more prominently in Chinese than in English. This interpretation, however, seems somewhat inconsistent with the observation that Chinese participants did not show much activation in the traditional linguistic processing areas including the Broca’s and Wirnicke’s areas during mental computation.

At this point, then, it is not entirely clear why Chinese participants show strong activations in the pre-motor areas during mental computation. One conjecture comes from an observation that in traditional Confucian societies, particularly in China, Japan, and Korea, abacus training is very common. It would seem possible, then, that computation involved finger movements that are typical in the use of abacus. Alternatively, in these countries, arithmetic computation is typically taught in school with an emphasis placed on writing down intermediate steps on a sheet of paper and, importantly, extensive drills of this way of calculation are quite typical (as in the Kumon training that has become increasingly popular in the US in the recent years). For these reasons, Chinese participants in the Tang et al. experiment might have represented numbers spatio-visually by literally writing them down on the mental sheet of paper while solving arithmetic problems.

Summary

We presented a few examples that illustrate how repeated performance in the same routines can result in systematic differences in brain pathways that are engaged. This appears to be the case whether the routines at issue pertain to cultural tools such as abacus, cultural routines such as driving in a city that is complex to navigate, or cultural or educational practices such as arithmetic computation. The main point is that brain pathways can change as long as they are fired in certain scripted ways over an extended period of time. When fired together, the brain neurons begin to be wired together. The mind, then, becomes ‘retooled’ as a result (Wilson, in press).

Neural representations of the self

Recent work by Zhu and colleagues (2007) has shown that the structure of the self varies systematically across cultures at the level of brain representations. Previous work provides abundant evidence that while the self is thought about and elaborated on, the medial prefrontal cortex (mPFC) is engaged (Craik et al., 1999; Kelley et al., 2002; Lieberman et al., 2004; Gutchess, et al., 2006). Zhu and colleagues hypothesized that the mPFC should also be engaged by close others among people in China—an interdependent culture. One reason is that the self and close others are supposedly tightly connected for Chinese. Another possible reason is that interdependent selves are defined, at least within the Chinese context, by relational attributes, which are shared among close others. Closely connected selves are likely to have highly overlapping representations. Yet another reason is that in interdependent contexts close others are just as relevant to the self as the self itself is. In contrast, for those with independent self, the self might still be quite distinct from others no matter how close the two might be, as the relationship is based on the assumption of mutual separation. Furthermore, these selves tend to be defined by unique, idiosyncratic features. The findings were consistent with this reasoning, with mPFC strongly engaged when one’s mother was thought about and elaborated on among Chinese, but not among people of Western cultural origin (Zhu et al., 2007).¹

It is interesting to note that when people think about the self from others’ point of view (e.g. ‘Does my teacher think that I am lazy?’), the mPFC activation disappears. Instead, a more dorsal and posterior part of the PFC receives prominent activations (D’Argembeau et al., 2007). A recent, intriguing series of imaging studies by Han and colleagues have shown that when Chinese with strong religious commitments to Christianity and Buddhism think about the self, they show activations in the dorsal regions of the PFC rather than in the rostal and ventral mPFC that is typically linked to the self (Han et al., 2008; Han et al., in press). One interpretation is that religious people take the perspective of the God or Buddha and draw a judgment on the self. Perspective taking, however, might be more common as a cultural practice in interdependent cultures including Chinese culture (Leung and Cohen, 2007; Wu and Keysar, 2007). It is thus important to find out whether the hypothesized linkage between religiosity and perspective taking would generalize to other cultures, such as the USA, that value independence more.

Two decades ago, Kitayama et al. (1990, described in Markus and Kitayama, 1991) hypothesized that the self would be more salient, elaborated, and/or accessible than others including close others for independent selves, but this effect might be weaker or even reversed for interdependent selves. Drawing on work by Tversky and colleagues (see Tversky, 1977 for a review) on asymmetry of similarity judgment between elaborate vs impoverished concepts (i.e. North Korea is more similar to Russia than Russia is similar to North Korea because typically people know much more on Russia than on North Korea), Kitayama and colleagues tested the hypothesis and found, as predicted, that for North Americans one’s friends are judged to be more similar to the self than the self is to the friends. Importantly, however, the pattern was non-significantly reversed for Indian subjects.

Recent work with ERPs has validated this initial observation, showing that when shown a facial photo and asked to make a simple judgment of the face orientation (i.e. looking to the right or the left), people exhibit an ERP component that is sensitive to covert orienting processes to target stimuli (negative peak approximately 200 ms after the presentation of the photo, called N200 or N2). As may be expected, this N2 response has been shown to be reliably stronger for the self-face than for the faces of one’s colleagues among the British participants. Importantly, this self-face advantage in N2 response was attenuated among Chinese participants (Sui et al., 2009). Another relevant neuroimaging study shows that this effect can be influenced by priming of independence, demonstrating the causal significance of independence and interdependence in modulating the salience of the personal self. Specifically, when independence is primed with a manipulation that highlights the personal self (Gardner et al., 1999), the salience of the self-face becomes quite prominent even among Chinese (Sui and Han, 2007). Along with a recent imaging study by Chiao and colleagues (2009), which shows a remarkable degree of within-culture variation in the self, this work highlights the significance of examining individual differences and situation-dependencies of the self within a cross-cultural framework.

Cognitive biases linked to schemas of independence vs interdependence

One notable consequence of independence (vs interdependence) can be found in person perception. When people apply a cultural schema of independence, they tend to focus on dispositional attributes of another person in lieu of available situational constraints. The resulting cognitive bias of highlighting dispositional factors in person perception (called the dispositional bias) has proven to be highly robust. In fact, Lee Ross (1977) coined the term, the fundamental attribution error, to refer to this bias. However, what if an alternative schema of interdependence were more salient or more accessible as should be the case in East Asia? Under such conditions, the emphasis on disposition (vis-à-vis situation) in person perception may be greatly attenuated. Miller (1984) provided initial evidence for this prediction in her comparative work involving North Americans and Indians. The basic finding has since been replicated and extended by subsequent researchers (Morris and Peng, 1994; Chua et al., 2005b; Kitayama et al., 2006; Kitayama et al., 2009; see Choi et al., 1999; Mason and Morris, in press, for reviews).

The cross-cultural variation calls into question the degree to which the dispositional bias is ubiquitous across cultures or truly fundamental in that sense. It has been argued, however, that the cultural variation can be explained without challenging the key assumption that dispositional bias is still universal and ‘fundamental’. The argument is based on the hypothesis that there are two stages in person perception, with a first stage of automatic dispositional inference followed by a second stage of deliberate situational adjustment (Gilbert and Malone, 1995). Whereas the dispositional inference is automatic, quick, and obligatory, the situation adjustment is deliberate, slow, and optional. This theoretical framework can accommodate the cultural variation in dispositional bias by assuming that cultures vary in terms of the second optional process of situational adjustment: Relative to European Americans, Asians are more likely to attend and cognitively elaborate on situational constraints, thus showing an attenuated dispositional bias (Gilbert et al., 1988). The validity of this analysis, however, rests on the claim that the initial dispositional inference is equally automatic and obligatory across cultures.

In a series of studies, Na and Kitayama (2009) tested if cultural variation exists not only in situational adjustment, but also in spontaneous dispositional inference. They hypothesized that given the interdependent model of person, dispositions are less important in accounting for another person’s behavior. As a consequence, interdependent individuals are less likely to routinely engage in dispositional inferences, and hence less likely automatically producing dispositional inferences. To address this issue, the researchers presented European Americans and Asian Americans with many pairs of a facial photo and a simple behavior. Participants were asked to merely memorize the pairs. Subsequently, participants performed a lexical judgment task. Immediately before a stimulus word (or non-word) was presented, they were briefly shown each of the facial photos used in the first phase of the study. On some trials, the stimulus word referred to the trait that was associated with the face (the congruent trials). On some other trials, the word had no semantic relation with the trait associated with the face (the neutral trials). On the remaining trials, a word-like sequence of alphabetical letters was shown. To the extent that the trait was inferred and then assigned to the face, the lexical judgment should be more efficient, as revealed in shorter response time, on the congruent than on neutral trials. This facilitation effect served as an index of dispositional inference.

Na and Kitayama found a reliable cultural difference: Whereas European Americans showed a facilitation effect, thereby demonstrating spontaneous trait inference, the effect completely vanished for Asian Americans. The implication is both clear and important: cultures vary not only in the salience of situational information (as shown in previous studies), but also in the degree to which dispositions are automatically inferred and ascribed to the target person. Na and Kitayama conceptually replicated the initial behavioral evidence with an ERP measure, by showing that an ERP marker for the detection of semantic incongruity (N400) is reliably greater for incongruous traits than for congruous traits in the lexical judgment task for European Americans. But, this effect was vanished for Asian Americans.

Recent neuroimaging studies have demonstrated that when dispositions of another person are inferred, a network of the brain defined by the mPFC, temporal parietal junction (TPJ), and temporal pole is reliably activated (Harris et al., 2005; see also Mason and Morris, in press). Extrapolating from the Na and Kitayama finding, we may expect that this network will be less likely to be spontaneously recruited by Asians. In another recent study, Kobayashi and colleagues (2007) scanned the brains of Japanese and North American children as they read either stories that encourage mentalizing (i.e. mind-reading) of protagonists or stories that do not have any mentalizing elements. The participants were not explicitly instructed to infer the mental states of the protagonists. The researchers found a reliable activation of the mentalizing network (mPFC and TPJ) for both Japanese and Americans. Importantly, however, the activation was reliably more pronounced for European Americans than for Japanese, thereby corroborating the thesis of Na and Kitayama.

One basic pattern that can be gleaned from the cross-cultural studies on person perception is that person as a figural object (vis-à-vis situational information) stands out in social perception of North Americans. For Asians and Asian Americans, however, the figural object seems more embedded because it is tied more closely to situational constraints or affordances. Extrapolating from this literature, it may be anticipated that nonsocial attention mechanisms can also be influenced by culture. Specifically, attention may be more focused for North Americans and it is more diffused or more holistic for Asians and Asian Americans. Recent studies by Masuda and Nisbett (2001), Kitayama and colleagues (2003), Chua and colleagues (2005a), and many others have provided converging evidence for this anticipation with a variety of behavioral measures including response time, task performance, and eye-tracking.

Some recent imaging studies provide additional evidence. For example, Hedden and colleagues (2008) have shown in their fMRI study that European Americans recruit an attention network of the brain (defined by prefrontal cortex and parietal cortex) when they attend holistically to both an object and its context than when they focus exclusively on the object. However, Asians show a reversed pattern, recruiting the same attention network more in the focused attention task than in the holistic attention task. It appears that people use intentional and effortful attention to compensate for their ‘cultural cognitive deficits’.

In another relevant study, Gutchess and colleagues (2006) found that when processing focal objects, areas involved in object-processing (the bilateral middle temporal gyrus, the left superior parietal and the right superior temporal) were recruited more for Americans than for Chinese, particularly in elderly populations. In this study, however, the two cultural groups did not differ in the brain activity involved in background processing. More recent fMRI studies that tested a similar idea with somewhat different procedures have obtained consistent findings (Goh et al., 2007; Jenkins et al., in press).

One important ambiguity in interpretation of the two lines of studies noted above stems from the fact that whereas Hedden and colleagues used the observed brain activation to suggest that people compensate for the culturally conditioned lack of fluency of processing, Gutchess and colleagues use the observed brain activation to suggest the cultural fluency of processing (see also Gutchess et al., in press, for a related issue). The findings therefore must be carefully interpreted in light of other available data, both behavioral and neural. In future work, some behavioral or self-report data may be quite useful to test the Hedden et al. interpretation that the brain activation observed by their study is related to conscious effort. It will also be important to use additional measures to show that the activation observed by Gutchess and colleagues is related exclusively to fluency in automatic processing.

The hypothesized sensitivity or attentional attunement to contextual information for Asians has also been investigated with ERPs. Goto et al. (in press) demonstrated that an ERP marker corresponding to the detection of semantic incongruity (N400; Ganis and Kutas, 2003) is enhanced when a background scene in a picture is incongruent with an object featured in the picture. As may be expected, this incongruity response was significantly more pronounced for Asians than for European Americans. Importantly, this response was observed to become weaker for both cultural groups as a function of independent self-construal as assessed by Singelis’ (1994) measure. Ishii and colleagues (in press) reported a compatible finding in an ERP study that looked at the processing of word meaning as a function of background vocal tone.

The context sensitivity in attention may also involve a temporal dimension. When observing an object (say, a car), individuals may vary in their propensity to holistically attend to other related objects that are observed in the past. These objects constitute a ‘mnemonic context’ for the perception of the impinging object. Extrapolating from the existent work on spatial attention, Duffy and Kitayama (2007) hypothesized that the mnemonic context effect would also be greater for Asians than for European Americans. The researchers presented subjects with a series of circles that varied in size. Subjects were to judge the diameter of each circle. Within this task, a context effect is typically found such that the size judgment of a target circle is a weighted average of the objective size of the target and the average size of all the circles that have been presented in the series. Duffy and Kitayama (2007) replicated this context effect. As predicted, however, this context effect was greater (i.e. a greater weight given to the previous circles) for Japanese than for European Americans. This behavioral finding has been conceptually replicated with an ERP measure (Lewis et al., 2008). They used an oddball paradigm, in which a target stimulus and an unanticipated, distracting stimulus is presented amongst more frequent standard non-target stimulus, and found that a novelty P3 is significantly greater for Asians than for European Americans. The novelty P3 is an index of a ‘surprise response’ to an unanticipated stimulus. It therefore is likely to increase as a function of weight given to a default expectation that is formed by repeatedly presented, anticipated stimuli. Importantly, this brain response (novelty P3 within an oddball paradigm) was reliably predicted by interdependence as assessed by a questionnaire measure.

In sum, the emerging body of neuroscience studies on cultural variation in cognition has begun to demonstrate that (i) as compared to interdependent Asians independent European Americans tend to engage in processing of focal social objects in lieu of background contextual information and, moreover, (ii) analogous cultural differences can be found even in non-social domains. Both fMRI and ERP methods have been quite instrumental in showing specific brain mechanisms that are involved.

Emotions and emotion regulation

There are very few systematic cross-cultural neuroscience investigations on emotion. Thus, there is very little to report in this section. Nevertheless, a few promising areas of research may be suggested (see also Seligman and Brown, in press for a related discussion). In the current cultural psychology literature, one highly consistent behavioral finding involves emotions that are derived from independence such as pride in the self, feelings of superiority, anger, and frustration (called socially disengaging emotions) and emotions that are derived from interdependence such as friendly feelings, respect, guilt and shame (called socially engaging emotions) (Kitayama et al., 2006; Kitayama and Park, 2007). An empirical generalization is that the socially engaging emotions are experienced more for Asians and socially disengaging emotions are experiencec more for European Americans. Systematic cross-cultural variations might be expected on brain areas recruited to yield experiences of social engagement and social disengagement (see Schaefer et al., in press, for a similar approach applied to the evaluation vs potency dimensions of the semantic differential).

Another extension would test cross-culturally divergent emotional responses to conditions of adversity. For example, in a recent study, Kimel et al. (2009) adopted a cyber-ball paradigm (Williams and Jarvis, 2006) to manipulate social exclusion among European Americans and Japanese. In a cyber-ball paradigm, participants are either ostracized or included during an online ball tossing game by two or three other players who are, in fact, controlled by an experimenter. As may be predicted by the supposition that Japanese are relatively more interdependent, Japanese reportedly experienced more negative emotions after social exclusion than Americans did. Importantly, however, this cross-cultural difference was found in sadness, but not in anger. Sadness is typically linked to interpersonal loss and a resulting motivation to restore the original relationship; whereas anger is unequivocally self-oriented, expressing frustrations associated with interfered goals of the self.

It is possible that pain areas (e.g. the ACC and the anterior insula) are implicated in social exclusion across cultures (Eisenberger et al.,2003); yet given the Kimel et al., finding, the somatic sensation of pain might be construed in different fashion to produce divergent emotional experience across cultures. This possibility is consistent with Damasio’s (2003) somatic marker hypothesis, which holds that evolutionarily grounded somatic arousals are differentially perceived and interpreted to produce subjective feelings. Also consistent is an observation by Levenson et al. (1992). These researchers commented that relational contexts might be required for Asians to experience emotions, but probably not for North Americans. Indeed, a recent study by Uchida and Kitayama (2009) has shown that unlike Americans, Japanese experience interpersonal harmony as an integral constituent of happiness. It is of interest to see if this cross-culturally divergent emotion construction might entail activation of different areas of the brain.

Yet another important area of fruitful cross-cultural exploration involves emotion regulation. A recent review of self-report measures of emotion regulation shows that Asians are far more likely to suppress their emotions than European Americans (Matsumoto et al., 2008). This finding goes hand-in-hand with a Confucian idea that emotions often hinder ever-important social relations. If Asians routinely try to suppress their emotions and, moreover, if Asian culture provides them with some effective means to do so, they may become more capable in emotion regulation. One possibility, suggested by a recent study by Grossmann and Kross (in press), is that interdependent people might be more capable than independent people to take a third person perspective even to their own emotional distress, thereby accomplishing a greater degree of self-detachment, which could allow them to regulate their emotions.

A recent study by Goldin et al. (2008) shows that European Americans exhibit an even greater amygdala activation when they try to suppress their emotions to pictures that evoke negative emotions. This paradoxical effect is likely due to a need to pay close attention to an emotion even when the goal of doing so is to suppress the emotion (Wegner et al., 1993). In contrast, another recent imaging study by Ohira and colleagues (2006) tested Japanese and found that their Japanese subjects were entirely capable of eliminating the amydgala activation when asked to do so (although they showed some evidence of autonomic arousal).²

Yet another domain for fruitful exploration relates to emotional responses to ingroup vs outgroup members. Recent work by Xu and colleagues (2009) suggests that empathetic responses to another person’s distress are much stronger if the person is from an ingroup than if he/she is from an outgroup. Adams and colleagues (in press) showed that the amygdala activation to a fear face depends on both gaze and group membership of the stimulus face. In particular, the amygdala responds especially strongly to an ingroup fear face of averted gaze (indicative of a common threat in the direction of the gaze) and an outgroup fear face of direct gaze (indicative of a potential threat to the self) (see also Chiaoet al., 2008; Moriguchi et al., 2005). Interestingly, in these studies, there is no cross-cultural difference in the responses to the ingroup vsoutgroup stimuli (see also Rule et al., in press, for a similar point). Ethnocentrism and an assortment of psychological and neural processes involved in it might prove to be widespread and relatively uniform across cultures although part of this is likely to be due to the relative familiarity of ingroup (as opposed to outgroup) members (Adams et al., 2009).

Personal versus social motivation

Given the general significance of personal or private attributes for independent selves and social or public attributes for interdependent selves (Kim and Drolet, 2009), the nature of social motivations may also vary cross-culturally. A general hypothesis has been implied by Taiwanese psychologists, Yu and Yang (1994), who observed that achievement motivation in Taiwan is ‘socially oriented’. In their view, what appears to be a purely personal striving of achievement is in fact anchored in expectations of significant others and the sense of social obligations and duties to the others in Taiwan. This observation is also echoed in an earlier investigation by DeVos on achievement motivation in Japan (DeVos, 1965).

Recent behavioral studies have provided abundant evidence for this prediction. Iyengar and Lepper (1999), for example, tested intrinsic motivation and showed that European American children are motivated to perform a task they have chosen. In contrast, Asian American children did not show this effect at least to the same degree. Instead, these children were motivated more by a choice made by their mothers, underscoring the hypothesis that interdependent selves are more strongly motivated when one’s motivation is anchored in others’ expectations. A more recent series of studies on cognitive dissonance by Kitayama and colleagues (2004) are consistent with this hypothesis. Replicating a large number of previous Western studies, these researchers have found a choice justification effect among Asians and Asian Americans. However, the conditions in which the effects are obtained vary systematically across cultures. Whereas European Americans (or Canadians) show a strong justification effect when the choice is perceived as private and personal, Asians and Asian Americans show the choice justification effect only when the choice is perceived as public. From this pattern of data it has been proposed that European Americans justify their choice when the choice is private and thus it threatens the personal sense of the self (e.g. ‘I might be irrational and stupid’), whereas Asians and Asian Americans justify their choice when the choice is public and, thus, it threatens the public sense of the self (e.g. ‘They might think that I am irrational and stupid’).

Notably, the ‘public’ choice in some of the studies cited above is no more than a choice that subjects make while being exposed to a set of schematic faces. Evidently, when one is unobtrusively exposed to a schematic face while making a decision, the person spontaneously categorizes the decision to be public, as being made under others’ scrutiny, and treats it as such. It is also possible that if the situation is categorized as public, pertinent social norms are activated (Haley and Fessler, 2005; Rigdon et al., 2009).

Extrapolating from this body of evidence, we may hypothesize that when European Americans are exposed to watching faces, they are reminded of a strong norm for independence. At the same time, the watching faces give rise to the impression of their decision as public. Decisions made under public scrutiny are perceived as less than independent and, as a consequence, they may be seen as counter-normative and, thus, as irrelevant to the self. Likewise, when Asians are exposed to watching faces, they may be reminded of a strong norm for interdependence. At the same time, the watching faces give rise to the impression of their decision as public. Decisions made under public scrutiny are perceived as quite relevant to their interdependent status and, as a consequence, they are seen as highly relevant to the self. The pattern observed in the foregoing set of studies is consistent with the general hypothesis that private decisions are more significant for European Americans, but public decisions are more significant for Asians and Asian Americans.

The foregoing hypothesis implies that whenever Asians (or European Americans) make decisions in public (or in private), they tend to treat the decisions as more important. Because this contingency is likely to be repeated and thus overlearned, we may expect that cues indicating the public (or private) nature of the decisions will automatically alert the brain of the individuals that the decisions are significant to the self. In a recent ERP study, Park, Gehring, and Kitayama (2009) tested this possibility by examining the error-related negativity (ERN), an event-related brain potential component that is associated with error commission in speeded choice reaction time tasks (Falkenstein et al., 1991; Gehringet al., 1993). Previous work showed that the ERN magnitude is sensitive to the motivational significance of the errors. For example, ERNs are stronger as a function of monetary incentives associated with correct decisions (Hajcak, et al., 2005; Pailing and Segalowitz, 2004).

In the Park et al., study, participants were asked to perform a flanker task where they were asked to judge the direction of the center arrow that was surrounded by two arrows on each side (e.g. >><>> or <<><<). As a cue indicating the publicity of a decision, Park et al. briefly exposed the participants to a facial photo or a control image (i.e. a scrambled face or a house) right before the presentation of the stimulus arrows. If the face cue were sufficient to alert Asians to potential significance of the decision, it would enhance the ERN for Asians. For European Americans, however, the publicity of a decision can diminish the significance of the decision. If so, their ERNs might be weaker in the face priming trials relative to the control trials. The results were consistent with these predictions. Asians showed a stronger ERN in the face priming trials than in the control trials. In contrast, European Americans showed a pattern that is completely opposite, exhibiting a weaker ERN in the face priming trials than in the control trials. Importantly, individuals’ self-beliefs of interdependence (vs independence), as assessed with Singelis scale (1994), reliably predicted the ERN magnitude in the face priming trials. In fact, the cultural difference in the face priming effect was completely mediated by the cross-culturally differential level of interdependence (vs independence).

Summary

When Chiao and Ambady (2007) published a chapter on cultural neuroscience, there was not much to be reviewed in the literature. Now only in a matter of a few years, this new field of research has produced a fair number of empirical findings. Although some of the empirical claims may have to be qualified by future studies, the rapid rate at which the new body of knowledge on culture and the brain has grown is impressive. It seems quite clear that the brain and various neural processing pathways it contains are influenced, sometimes quite substantially, by culture and, as a consequence, when two or more vastly different cultures are compared, highly systematic differences in brain responses can be observed. While the basic anatomic structure of the brain will never be influenced by culturally organized experience, patterns of connectivities among various structural elements or areas of the brain as well as parameters associated with such connectivities appear to be influenced substantially.

We believe that these cross-cultural findings may best be understood and interpreted in light of the evidence for brain plasticity as a function of cultural tools and practices reviewed in the preceding section of this article. This evidence strongly suggests that repeated engagement in conventionalized behavioral patterns (as in abacus use or in cab driving) results in highly systematic changes in pertinent brain responses. By extrapolating from this observation, it would be safe to hypothesize that culture’s influences on the brain result from repeated participation and engagement in culture’s conventions, routines, and socially shared scripts for action. This point leads us to the next section, in which we present a theoretical model designed to bridge culture and the brain by emphasizing the critical role of ‘cultural tasks’ as a middle range layer connecting culture to the brain and vice versa.