先前的研究表明人们在加工和自我相关的信息时会更加敏感，形成自我偏见，然而这种敏感性可能会因为语境的不同
而不同。本研究运用功能磁共振成像技术，考察单语语境（全局控制）和混合语境（局部控制）对自我偏见的影响。结果发现，
单语第二语言（L2）语境下，自我- 消极比他人- 消极引发左侧额下回和左侧顶下小叶更大的激活；双语第一语言（L1）语境下，
自我- 积极比他人- 积极引发右侧楔前叶更大的激活。这表明语境引发的语言控制影响双语者解决身份与情绪关联之间的冲
突。

It is well known that different languages have different impacts on human cognition, emotion, and physiological states. In the context of
native language, individuals tend to associate positive emotions with themselves, and negative emotions with others. This leads to the self-positive
and other-negative bias. However, using a foreign language may reduce emotional involvement. We hypothesized that using a foreign language may
affect the connection between identity (self vs. other) and emotion (positive vs. negative) through reducing emotional involvement. In addition, we
hypothesized that the ability of language control when they switch between two languages may also affect bilinguals’ connection between identity and
emotion. Language control refers to the choice of using the target language in a specific situation and inhibiting the interference of non-target language.
Participants do not need to switch between languages in the monolingual context, which requires less inhibitory processing (global language control
occurred). In contrast, participants have to switch from one language to another in bilingual context frequently by inhibiting the interference from nontarget
language (local language control occurred). We hypothesized that connection between identity and emotion occurred automatically under global
language control, while it required more cognitive resources under local language control. Thus, the automatic process may be weakened in the context
of mixed language.
In the current study, bilingual context and monolingual context were designed. The stimuli in each scanning run was either presented in twolanguage
or in one-language context, corresponding to local and global language control. The variables of language (Chinese vs. English), identity (Self
vs. Others) and emotional valence (Positive vs. Negative) were manipulated. Identity cues were represented by " 我", " 他", "I" and "He". The data
of 29 unbalanced bilinguals whose native language was Chinese were collected by fMRI scanner. In each trial, the identity cues were presented first,
followed by emotional words. The participants were required to determine whether the emotional words were true or pseudoword words (false words
were generated by replacing radicals or letters of real words), and their reactions were recorded. The DPABI and SPM were then used to preprocess
the data and do F test. For the brain regions with significant interactions of identity and emotional valence, we extracted the beta value of brain region
and performed repeated measures ANOVA to compare the activation of different conditions. At the same time, repeated measures ANOVA was also
performed for the reaction times.
The fMRI results showed that in the monolingual L2 context, the left inferior frontal gyrus and the left inferior parietal lobule showed greater
activation in the other-positive condition and the self-negative condition than that in the self-positive condition and the other-negative condition. In
bilingual L1 context, the right precuneus exhibited significantly greater activation in self-positive condition than other-positive condition. The response
times results showed that the reaction of the self-positive condition was faster than that of the other-positive condition in the monolingual L1 and L2
context, respectively. In the bilingual L2 context, the reaction of the other-negative condition was faster than that of the self-negative condition.
Based on the findings, we infer that: (1) In the monolingual L2 context, the self-negative and the other-positive bias were in conflict with each
other because they were not consistent with the self-positive bias. The frontoparietal network was employed to process the conflicts. (2) The precuneus
activity in the bilingual L1 context may reflect the attribution processes that distinguishing others’ emotions and self-emotion, leading to one of them
is more sensitive to self. (3) The differences between global and local language control may indicate that bilinguals have more cognitive resources
to solve the conflicts between identity and emotion in foreign language context (global language control occurred), while in the context of mixed
language, more cognitive resources are required when processing L2 than L1 (local language control occurred). It seems that participants have extra
cognitive resources to process the association between identity and emotion when using their native language. Overall, these findings suggest that
language context induces bilinguals’ language control, which affects their resolution of the conflicts between identity and emotional valence.
Key words language context, language control, self-bias, emotion, left inferior frontal gyrus, left inferior parietal lobule, right precuneus
Participants do not need to switch between languages in the monolingual context, which requires less inhibitory processing (global language control occurred). In contrast, participants have to switch from one language to another in bilingual context frequently by inhibiting the interference from nontarget language (local language control occurred). We hypothesized that connection between identity and emotion occurred automatically under global language control, while it required more cognitive resources under local language control. Thus, the automatic process may be weakened in the context of mixed language.
In the current study, bilingual context and monolingual context were designed. The stimuli in each scanning run was either presented in twolanguage or in one-language context, corresponding to local and global language control. The variables of language (Chinese vs. English), identity (Self vs. Others) and emotional valence (Positive vs. Negative) were manipulated. Identity cues were represented by " 我", " 他", "I" and "He". The data of 29 unbalanced bilinguals whose native language was Chinese were collected by fMRI scanner. In each trial, the identity cues were presented first, followed by emotional words. The participants were required to determine whether the emotional words were true or pseudoword words (false words were generated by replacing radicals or letters of real words), and their reactions were recorded. The DPABI and SPM were then used to preprocess the data and do F test. For the brain regions with significant interactions of identity and emotional valence, we extracted the beta value of brain region and performed repeated measures ANOVA to compare the activation of different conditions. At the same time, repeated measures ANOVA was also performed for the reaction times.
The fMRI results showed that in the monolingual L2 context, the left inferior frontal gyrus and the left inferior parietal lobule showed greater activation in the other-positive condition and the self-negative condition than that in the self-positive condition and the other-negative condition. In bilingual L1 context, the right precuneus exhibited significantly greater activation in self-positive condition than other-positive condition. The response times results showed that the reaction of the self-positive condition was faster than that of the other-positive condition in the monolingual L1 and L2 context, respectively. In the bilingual L2 context, the reaction of the other-negative condition was faster than that of the self-negative condition.
Based on the findings, we infer that: (1) In the monolingual L2 context, the self-negative and the other-positive bias were in conflict with each other because they were not consistent with the self-positive bias. The frontoparietal network was employed to process the conflicts. (2) The precuneus activity in the bilingual L1 context may reflect the attribution processes that distinguishing others’ emotions and self-emotion, leading to one of them is more sensitive to self. (3) The differences between global and local language control may indicate that bilinguals have more cognitive resources to solve the conflicts between identity and emotion in foreign language context (global language control occurred), while in the context of mixed language, more cognitive resources are required when processing L2 than L1 (local language control occurred). It seems that participants have extra cognitive resources to process the association between identity and emotion when using their native language. Overall, these findings suggest that language context induces bilinguals’ language control, which affects their resolution of the conflicts between identity and emotional valence.