Stereotype threat (ST) refers to the feeling of threat which is induced when individuals are worried that their behavior might confirm a stereotype or stigmatize the social identity of their in-group (Schmader and Johns, 2003; Jamieson and Harkins, 2007; Wout et al., 2009). Based on this, it has been confirmed that the negative stereotypes about a social identity could add pressure or concern and lead to underperformance in domain tasks (Steele and Aronson, 1995; Steele, 1997). Although numerous studies show mechanisms of underperformance in domain tasks under ST (e.g., Martens et al., 2006; Wraga et al., 2007; Johns et al., 2008; Krendl et al., 2008), there are only two studies that attempted to uncover the neural mechanisms of ST by using functional magnetic resonance imaging (fMRI) (Wraga et al., 2007; Krendl et al., 2008). Thus, it is helpful to learn more about the neural basis of ST by using the resting-state fMRI (RS-fMRI) to investigate the effects of ST on the degree centrality in this study.

It has been suggested that the negative stereotype could be threatened when it is self-relevant (Steele, 1997). According to the integrated process model of ST, ST first induces an imbalance among the concepts of in-group, the ability of domain tasks, and self, which also implies that ST must be self-relevant (Schmader et al., 2008; Liu et al., 2021). Thus, when the ability of domain tasks is threatened by ST information, it can lead to the processes of self-doubt and self-validating the stereotype (Steele and Aronson, 1995), where negative thinking (Cadinu et al., 2005) and negative emotions (Johns et al., 2008) are further generated. To reduce the detrimental effects of negative thinking and negative emotions on performance, individuals attempt to suppress them, which reduces the working memory capacity required by the performance of domain tasks (Schmader and Johns, 2003; Johns et al., 2008).

Considering the effects of ST on self-relevant processes, negative emotions, and working memory capacity, the brain function related to them might be influenced by ST. For example, it has been found that the brain regions related to social emotion [e.g., angular gyrus (AG), left parietal cortex, and prefrontal cortex] were more active when a domain task was performed under an ST condition (Krendl et al., 2008). Wraga et al. (2007) also found that the brain regions associated with emotional load (e.g., medial prefrontal cortex, anterior cingulate cortex, and amygdala) were activated when a domain task was performed under an ST condition. Moreover, Forbes and Leitner (2014) found that negative feedback to the performance of domain tasks elicited a larger P100 component when individuals were under an ST condition, confirming that the working memory capacity is easily taxed by ST information. Mangels et al. (2012) found that the negative feedback related to event-related potential components is influenced by ST, implying that individuals are prone to disengage from domain tasks influenced by negative emotions.

Efficient brain function does not require the functioning of different brain regions during specific tasks, but it does require information being transported between them (Van Den Heuvel and Hulshoff Pol, 2010). Although some brain regions have been confirmed to be associated with ST, the brain network being influenced by ST is not clear. RS-fMRI can use the blood-oxygen-level-dependent (BOLD) signal of spontaneous activations during low-frequency fluctuations (0.01–0.1 Hz) to depict brain networks (Lowe et al., 2000; Van Den Heuvel and Hulshoff Pol, 2010). RS-fMRI degree centrality (RSDC) is a graph theory-based network analysis of the number of edges connecting to a node or the node strength for a given node (a voxel) in a whole-brain network (Zuo et al., 2012; Li et al., 2016). It has been found that RSDC has test-retest reliability and high sensitivity (Zuo and Xing, 2014) and that it can find the hubs of brain networks and provide functional connectivity of the entire brain (Zhang et al., 2019).

In previous studies (e.g., Schmader and Johns, 2003; Johns et al., 2008; Schmader et al., 2008), the female math ST was widely used to study the ST effect and had common mechanisms with other ST to some extent. Thus, the current study aimed to investigate the effects of ST on RSDC under math ST. According to the results of ST studies (Martens et al., 2006; Wraga et al., 2007; Johns et al., 2008; Krendl et al., 2008), we speculated that the RSDC of brain regions related to the regulation of social emotions (e.g., the medial prefrontal cortex and anterior cingulate cortex) might be increased. Considering the roles of the hippocampus in the generation of stressful responses (Buss et al., 2007; McEwen and Gianaros, 2010; Tottenham and Sheridan, 2010), we hypothesized that the RSDC of the hippocampus might be decreased under ST, which might make individuals more prone to experience stress-based arousal. Except in this region, due to the self-relevant processes being increased by ST (Steele and Aronson, 1995), we thought that the RSDC of the brain regions related to self-memory should be increased. Therefore, we further speculated that the RSDC of the brain regions, such as the right posterior parietal regions (PPC), related to the retrieval of episodic memory or autobiographical memory might be increased (Cabeza, 2008; Cabeza et al., 2011).

In the present study, the effects of ST on the RSDC of brain networks were investigated using RS-fMRI under math ST conditions. The results showed that the RSDC decreased in the left MOG and the left hippocampus, while the RSDC increased in the right MCG, right SPG, right AG, and left precuneus. Furthermore, the results also showed that the right ATL and right hippocampus were negatively correlated with the MC score in the ST group, while the right ATL and the right hippocampus were positively correlated with the MC score in the control group.

It has been found that brain regions in the posteromedial cortex, such as the precuneus and posterior cingulate cortex (PCC), are the hubs of the structural and functional networks of the brain (Hagmann et al., 2008), which means that these regions are highly connected to cortical and subcortical networks (Hagmann et al., 2008; Leech et al., 2012). These characteristics are involved in broad mental processing, such as episodic memory (autobiographical memory), mentalizing (Muscatell et al., 2012), self-referential processing, or self-reflection (Northoff and Bermpohl, 2004; van der Meer et al., 2010), and they allow the brain to acquire information from functionally distinct brain networks easily (Leech et al., 2012). Moreover, the left precuneus is a part of the default mode network (DMN) and plays an important role in self-generated thoughts (Andrews-Hanna et al., 2014; Axelrod et al., 2015). Therefore, we speculated that the increased RSDC in the left precuneus and MCG might reflect that some self-relevant thoughts are generated by retrieving autobiographical memory or episodic memory.

The self-relevant processes under ST could also be reflected by the increased RSDC in the PPC. Previous studies suggested that the PPC can be divided into the dorsal parietal cortex (DPC) and ventral parietal cortex (VPC), where the DPC includes the intraparietal sulcus and SPG, and the VPC includes the supramarginal and angular gyri (Cabeza, 2008; Cabeza et al., 2011). Although both VPC and DPC are involved in the retrieval of autobiographical memory or episodic memory, the DPC is mainly involved in the top-down attention of retrieving these memories, and the VPC is mainly related to the bottom-up attention of these memories that are activated automatically (Cabeza, 2008; Cabeza et al., 2011; Wu et al., 2015). Moreover, the PPC is also located in the DMN and plays an important role in self-generated thoughts (Andrews-Hanna et al., 2014; Axelrod et al., 2015). Therefore, the increased RSDC in PPC for the ST group might further reflect that the self-relevant processes were induced by ST.

It has been suggested that ST increases stress-based arousal, such as increasing the activation of the sympathetic nervous system (Murphy et al., 2007), blood pressure (Blascovich et al., 2001), and cardiovascular responses (Mendes et al., 2002). Results from both animals and humans confirm that the experience of stress has significant negative effects on the hippocampus (Buss et al., 2007; McEwen and Gianaros, 2010; Tottenham and Sheridan, 2010), where glucocorticoids and mineralocorticoids are released to regulate the hypothalamic-pituitary-adrenal (HPA) axis on which physiological processes depend (Henry et al., 1994; Kudielka et al., 2004). The hippocampus is an important brain region for the regulation of stress hormones (Kim and Diamond, 2002; Glover et al., 2010), such as inhibiting the HPA axis through glucocorticoid-mediated negative feedback to terminate the stress response (Sapolsky, 1992; McEwen and Sapolsky, 1995). According to the results, the RSDC of the left hippocampus was decreased, and the RSDC of the right hippocampus was negatively correlated with the MC score under the ST condition. These results might reflect that the importance of the hippocampus in the brain networks was decreased by the ST condition, which might make individuals prone to exhibit stress-based arousal mentioned earlier.

It has been found that the role of the ATL was related to linking person-specific memories to representations of faces at the perception level for face processing (Olson et al., 2007) and is associated with the processes of comprehending the mind of others based on personal memories (Moriguchi et al., 2006). In addition, the ATL has an important role in social concepts (Olson et al., 2013). For example, previous studies have shown that stereotypes (a social concept) are represented by the ATL, which might play an important role in linking certain types of people to specific personality or behavioral traits (Olson et al., 2013). According to the results, the RSDC of the right ATL was negatively correlated to the MC score in the ST group, while the RSDC of the right ATL was positively correlated to the MC score in the control group. These findings reflect that the relative importance of the brain regions related to social concepts is decreased under ST, especially for more threatened individuals. It is dangerous to increase the self-relevant processes related to ST and at the same time decrease the functions of social concepts. Doing so might make individuals prone to be influenced by their stigmatized social identity and discount other beneficial social identities. Consistent with this perspective, a previous study found that learning other beneficial social identities can decrease the effects of ST (Rydell et al., 2009).

We detected the effects of ST on brain network degree centrality by directly comparing the RSDC between the ST and control groups. These findings expand the knowledge of the neural basis of ST. Specifically, the results suggest that ST decreases the importance of brain regions related to social concepts and stress regulation in the whole brain networks and, at the same time, increases the importance of brain regions associated with self-relevant processes. We acknowledge several limitations of the study. First, the performance of the math problem was not measured in an RS-fMRI study, and we did not confirm whether the altered RSDCs were related to underperformance under ST. Second, Steele and Aronson (1995) found that ST could lead to some self-concerning processes, such as self-doubts, self-validating the stereotype, and stereotype avoidance. Although we speculate that the altered RSDCs might be related to these mental processes, we could not provide direct evidence for them because behavior related to these processes was not measured. Third, due to this evidence being exclusively from female math ST, future studies should verify whether these results generalize to other STs (e.g., racial ST or ST related to social status). Therefore, we suggest that a sophisticated experiment integrating behavioral, task, and RS-fMRI may be required to reveal the detailed mechanism of ST.

The datasets generated for this study are available on request to the corresponding author.

The studies involving human participants were reviewed and approved by Brain Imaging Center Institutional Review Board of Southwest China University. The patients/participants provided their written informed consent to participate in this study.

XW and YZ designed the experiment and wrote this manuscript. Both authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.