Neural representation of perceived race mediates the opposite relationship between subcomponents of self-construals and racial outgroup punishment

Abstract

Outgroup aggression characterizes intergroup conflicts in human societies. Previous research on relationships between cultural traits and outgroup aggression behavior showed inconsistent results, leaving open questions regarding whether cultural traits predict individual differences in outgroup aggression and related neural underpinnings. We conducted 2 studies to address this issue by collecting self-construal scores, EEG signals in response to Asian and White faces with painful or neutral expressions, and decisions to apply electric shocks to other-race individuals in a context of interracial conflict. We found that interdependent self-construals were well explained by 2 subcomponents, including esteem for group (EG) and relational interdependence (RI), which are related to focus on group collectives and harmonious relationships, respectively. Moreover, EG was positively associated with the decisions to punish racial outgroup targets, whereas RI was negatively related to the decisions. These opposite relationships were mediated by neural representations of perceived race at 120-160 ms after face onset. Our findings highlight the multifaceted nature of interdependent self-construal and the key role of neural representations of race in mediating the relationships of different subcomponents of cultural traits with racial outgroup punishment decisions in a context of interracial conflict.

Keywords: EEG; interdependence; intergroup conflicts; punishment decisions; race.

Fig. 1

Results of the EFA. A) Items in self-construal scale and their corresponding factors in the 6-factor model. B) The 6-factor model of self-construals. The unidirectional arrows indicate the factor loadings of each item. The bidirectional arrows indicate intercorrelations between factors. EG (i.e., Esteem for Group) and RI (i.e., Relational interdependence) are from the original interdependence scale, whereas assertiveness, pursue of uniqueness, individualism, and behavioral consistency are from the original independence scale.

Fig. 2

Associations between EG/RI and race-related punishment decisions. A) SEM models relating EG and RI to outgroup punishment decisions. Significant regression parameters are shown in solid lines. All parameter estimates shown are fully standardized. B) Bar plots illustrating the outgroup punishment decisions for individuals endorsed with low and high EG/RI or interdependence scores.

Fig. 3

Results of the RSAs. A) Illustration of the RSA model. At each time point, the neural dissimilarity matrix between conditions was constructed by calculating the MD for all pairs of conditions, which captures the multivariate distance between topographies. This neural dissimilarity matrix was then modeled as a linear combination of models based on race, pain, and gender dissimilarity (1 corresponding to between and 0 corresponding to within category, respectively). This analysis produced one beta estimate time course for each dimension at each time point. B) RSA results using MD. The time courses of regression weights show the contributions of race, pain, and gender dissimilarity to neural dissimilarity for the large sample EEG participants (n = 676). Lines below plots indicated significant times using a cluster-based permutation test (cluster-defining threshold P < 0.05). C) SEM models relating EG and RI to early neural representations of race and pain. Significant regression parameters are shown in solid lines. Insignificant regression parameters are shown in dashed lines.

Fig. 4

Illustration of the results of mediation analyses. The beta weight for race RDM at the early time window (120–160 ms) mediates A) the positive relationship between EG and racial outgroup punishment behavior and B) the negative relationship between RI and racial outgroup punishment behavior.

Fig. 5

Illustration of sampling variability of the associations of EG/RI scores with outgroup punishment decisions and neural representations of race. The correlation coefficients between EG/RI scores and outgroup punishment decision (left panels) and the correlation coefficients between EG/RI scores and the race-related beta weight (right panels) are shown. Solid lines show the mean across 1,000 resamples for the corresponding associations. Shading represents the minimum to maximum correlation range across subsamples for a given sample size. Dashed lines represent 95% and 99% confidence intervals.