Metamemory refers to the processes of monitoring and controlling memory activities (Nelson and Narens, 1990). One important form of metamemory is prospective metamemory judgment, which refers to prediction of future memory performance (Dunlosky and Metcalfe, 2009). The accuracy of prospective metamemory judgments is typically evaluated by examining the extent to which prospective metamemory judgments discriminate between remembered and forgotten items (Dunlosky and Metcalfe, 2009). If the subjective level of prospective metamemory judgments is higher for remembered items than forgotten items, then the accuracy of prospective metamemory judgments is high. In contrast, prospective metamemory accuracy is low if people give higher judgments to forgotten items. Prospective metamemory judgments can be made either at the stage of acquiring knowledge (called judgment of learning, or JOL) or at the time of retrieval (called feeling of knowing, or FOK) (Nelson and Narens, 1990). JOL is an important type of prospective metamemory judgment, and previous studies have indicated that JOL accuracy is essential for appropriate guidance of subsequent learning and memory processes (Dunlosky and Metcalfe, 2009; Bjork et al., 2013). People with more accurate JOLs can use strategies that are more appropriate in subsequent learning processes, such as allocating study time more appropriately during self-regulated learning or choosing more important and valuable items for restudy (Nelson and Narens, 1990; Dunlosky and Metcalfe, 2009). Given the importance of JOLs in learning processes, the biological basis of the JOL process is an important topic in metamemory research (Schwartz and Bacon, 2008; Chua et al., 2014).

Some lesion-based studies have investigated whether frontal lobe lesions significantly affect JOL accuracy (Vilkki et al., 1998, 1999). For example, Vilkki et al. (1998) asked healthy participants and patients with brain lesions to learn a list of words. Before the recall test, all participants had to make a JOL to predict how many words they could recall in the test. They found that patients with left or right frontal lobe lesions had significantly lower JOL accuracy than the control group. In Vilkki et al. (1999), patients and healthy control participants were required to learn the locations of different faces and make a JOL before the test. The results showed that patients with damaged right frontal lobes showed decreased JOL accuracy. These studies suggest that the frontal lobe may play an important role in the JOL process. However, Vilkki et al. (1998, 1999) did not identify which part of the frontal lobe is essential to JOL accuracy.

In addition to lesion-based studies, functional magnetic resonance imaging (MRI) studies have investigated the brain regions underlying the JOL process (Kao et al., 2005; Do Lam et al., 2012; Yang et al., 2015). An early study by Kao et al. (2005) found that the left lateral frontal cortex, left ventromedial prefrontal cortex and several temporal, parietal, occipital and limbic regions were associated with predicted encoding success (i.e., higher activation for high-JOL trials than for low-JOL trials). They also examined the relationship between individual differences in JOL accuracy and the activation of the regions related to predicted encoding success, and they found that greater JOL accuracy was correlated with higher activation in the ventromedial prefrontal cortex. Another study by Do Lam et al. (2012) separated JOL trials from initial encoding and indicated that the medial prefrontal cortex, medial orbitofrontal cortex and anterior cingulate cortex were more active during high-JOL trials. In addition, Yang et al. (2015) compared the neural activities for encoding trials subsequently given high and low JOLs, and found that the dorsolateral, rostrolateral and ventromedial prefrontal cortex, posterior cingulate cortex, middle temporal gyrus, superior lateral occipital cortex and angular gyrus showed subsequent JOL effects (i.e., higher activation for trials subsequently given high JOLs than those given low JOLs). They also found that greater JOL accuracy was correlated with smaller subsequent JOL effects in the ventromedial prefrontal cortex. In these previous studies, the only region consistently involved in the JOL process was the ventromedial prefrontal cortex, particularly the medial part of Brodmann area (BA) 11. Similarly, a study of FOK has indicated that the activation of medial BA 11 is also related to accurate FOK judgments (Schnyer et al., 2005), suggesting that medial BA 11 is essential to different types of prospective metamemory judgments.

Another important method for investigating the neural correlates of metamemory processes is to examine the relationship between individual differences in brain structure and metamemory accuracy. To our knowledge, no study has examined the relationship between JOL accuracy and brain structure. However, researchers have used structural MRI to investigate the relationship between the FOK process and the gray matter (GM) volume in the brain (Cosentino et al., 2015; Le Berre et al., 2016). For example, Le Berre et al. (2016) asked alcoholic patients to make FOK judgments in the recall phase about whether they could recognize the target word in a subsequent recognition test; they found that greater FOK accuracy was correlated with higher mean GM volume of the bilateral insula. Similarly, Cosentino et al. (2015) indicated that higher mean GM volume of the right insula was related to higher FOK accuracy in older adults. Both studies have shown that individual differences in FOK accuracy are related to the GM volume of the insula.

Although previous studies have investigated the neural correlates of the JOL process from different perspectives, two important issues require further examination. First, although the activation of medial BA 11 has been shown to be consistently associated with the JOL process (Kao et al., 2005; Do Lam et al., 2012; Yang et al., 2015), no study has examined whether individual differences in JOL accuracy are related to the GM volume of medial BA 11. Second, previous structural MRI studies have not investigated whether JOL is also related to the GM volume of the insula, as found in structural MRI studies concerning FOK (Cosentino et al., 2015; Le Berre et al., 2016). Researchers have suggested that in FOK processes, the insula may be related to the monitoring of task performance (Cosentino et al., 2015) or self-inferential processes to generate accurate future estimations (Le Berre et al., 2016). These cognitive processes are also involved in the JOL process (Koriat, 1997; Dunlosky and Metcalfe, 2009). Thus, it is possible that the GM volume of the insula may also be related to JOL accuracy. Investigating these issues could further reveal the neural correlates of the JOL process. To address these issues, in the present study, we first investigated the relationship between JOL accuracy and GM volume in the brain.

In addition, previous neuroimaging studies have mainly focused on the particular brain regions related to JOL process and have not examined whether JOLs are correlated with the functional connectivity between different regions. According to the widely accepted cue-utilization theory, JOLs are inferential in nature and are related to the processing and integration of various cues (e.g., study time, processing fluency) relevant to study materials. When making JOLs, people first monitor and process the cues that may contribute to the memory process and then integrate these cues to generate estimations for future memory performance (Koriat, 1997; Dunlosky and Metcalfe, 2009). Thus, it is possible that JOLs are associated with the coupling between different regions that are related to the processing and integration of different cues. In addition, a previous study suggests that retrospective confidence judgment (RCJ), another type of metamemory judgment made after a memory test, is related to functional connectivity in the brain (Baird et al., 2013). For example, Baird et al. (2013) indicated that individual differences in the accuracy of retrospective confidence in a memory test were significantly correlated with the functional connectivity between the anterior prefrontal cortex, in which the brain structure is related to the accuracy of RCJs (Fleming et al., 2010), and several parietal regions during the resting state. Based on the results concerning RCJs, we speculated that JOLs might also be related to functional connectivity between different regions. Furthermore, making JOLs is similar to the decision-making process (Schwarz, 2004), which is also related to the integration of different information (Bettman et al., 1998; Lim et al., 2013). A previous study has suggested that the evaluation process in decision making is associated with the functional connectivity between several regions of the temporal cortex (which are related to the processing of different information) and the prefrontal cortex (which is related to the integration of information) (Lim et al., 2013). Thus, JOLs may also be associated with the functional connectivity between various brain regions. The second purpose of the present study was to investigate the relationship between JOL accuracy and functional connectivity in the brain.

Taken together, the main goals of the present study were (1) to examine the relationship between individual differences in JOL accuracy and the GM volume in the brain and (2) to investigate the functional connectivity associated with individual differences in JOL accuracy. To achieve the first goal, we analyzed the correlation between individual differences in JOL accuracy and the GM volume. In particular, we examined the correlation between JOL accuracy and the GM volume of medial BA 11 and the insula. Medial BA 11 is consistently involved in the JOL process (Kao et al., 2005; Do Lam et al., 2012; Yang et al., 2015) and has been correlated with JOL accuracy in previous task-related fMRI studies (although the direction of the correlation is controversial; see Kao et al., 2005; Yang et al., 2015). Thus, the GM volume of medial BA 11 may also be correlated with JOL accuracy. In addition, FOK studies have consistently revealed positive correlations between FOK accuracy and the GM volume of the insula (Cosentino et al., 2015; Le Berre et al., 2016). According to the discussion above, some cognitive processes that underlie FOK and may be related to the insula are also involved in the JOL process (Koriat, 1997; Dunlosky and Metcalfe, 2009), and it is possible that the GM volume of the insula might also be related to JOL accuracy. We hypothesized that JOL accuracy might show positive correlation with the GM volume of the insula (similar to FOK studies) and might also show significant correlation with the GM volume of medial BA 11 (although the direction of the correlation is unknown). There are two commonly used methods to analyze the relationship between GM volume and individuals’ performance: the traditional region-of-interest (ROI) approach used in previous FOK studies (Cosentino et al., 2015; Le Berre et al., 2016) and voxel-based morphometry (VBM) analysis. In the ROI-based approach, the correlations between individuals’ performance and the mean GM volume within the ROIs are analyzed. By contrast, VBM analysis permits the calculation of the correlation between individuals’ performance and GM volume voxel by voxel (Luders et al., 2004; Bhaskar et al., 2013). One pitfall of the ROI-based approach is that it is used primarily on large areas, and differences in GM volume in small parts of the ROIs may therefore be overlooked. By contrast, VBM analysis can examine focal differences in brain anatomy and reveal specific clusters related to individuals’ performance (Bhaskar et al., 2013). Thus, in this study, we chose VBM analysis to evaluate the relationship between GM volume and JOL accuracy. We conducted VBM analysis in a mask containing the bilateral insula and medial BA 11 to examine whether JOL accuracy was significantly correlated with the GM volume in these regions.

To achieve the second goal, we used resting-state fMRI to examine whether individual differences in JOL accuracy are related to the resting-state functional connectivity between different brain regions. Resting-state functional connectivity is indexed by correlations in low-frequency fluctuations of the resting-state fMRI signal (Lee et al., 2013). In contrast to task-related brain activity, which is evoked by specific task stimuli, resting-state functional connectivity may characterize the intrinsic functional organization of the brain (Fox and Raichle, 2007) and has high reliability and reproducibility (Lee et al., 2013). Although resting-state functional connectivity is not directly related to any cognitive task, a previous study suggests that individual differences in resting-state functional connectivity can predict task-induced brain activity (Mennes et al., 2010). In addition, resting-state functional connectivity has been shown to be closely related to individual differences in various cognitive functions, such as language, decision making and intelligence (Song et al., 2008; Li et al., 2013; Xiao et al., 2016). In the resting-state analysis, we used the significant cluster found in VBM analysis as a seed region and conducted whole-brain seed-based functional connectivity analysis. Using the cluster found in VBM analysis as the seed region is a widely accepted method in resting-state analysis (Loh and Kanai, 2014; Evans et al., 2015). We hypothesized that if the VBM analysis revealed that the clusters in medial BA 11 or insula were significantly correlated with JOL accuracy, the functional connectivity between these clusters and other regions might also support the process of making accurate JOLs. In addition, to further explore the relationship between JOL accuracy and functional connectivity in the brain, we conducted an exploratory ROI-wise analysis to examine whether JOL accuracy was correlated with the functional connectivity between regions connected to the seed region in the seed-based analysis. We also performed exploratory factor analysis on the functional connectivity found in the analyses above. Factor analysis is a statistical technique that can determine which variables form coherent subsets (factors) that are somewhat independent. Factors are hypothesized to reflect latent constructs or underlying processes that result in correlations between variables (Fabrigar et al., 1999). Thus, factor analysis may provide a better understanding of the functions of the resting-state network related to JOL accuracy by revealing the relationships between functional connectivity outcomes (Jacobs et al., 2014).

The present study investigated the neural correlates of individual differences in JOL accuracy. Participants completed a memory task in which they made JOLs about their recognition performance, and we used structural MRI and resting-state fMRI to examine whether individual differences in JOL accuracy were correlated with the GM volume and the resting-state functional connectivity in the brain. We found that the GM volume of the left mid-insula could predict individual differences in the participants’ JOL accuracy. The participants’ JOL accuracy was also correlated with the resting-state functional connectivity between the left mid-insula and various other brain regions, including the left SPL/Pcu, bilateral IPL/IPS, right frontal pole and left parahippocampal gyrus/fusiform gyrus/cerebellum, and the functional connectivity related to JOL accuracy could be divided into two subsets. Our findings provide novel evidence for the involvement of the left mid-insula and other distributed brain regions and their functional connectivity in the processes related to accurate JOLs and suggest that these regions may play an important role in the JOL process.

Our study found that greater JOL accuracy was related to higher GM volume of the left mid-insula. Our results are consistent with previous FOK studies indicating that greater FOK accuracy is correlated with higher GM volume of the insula (Cosentino et al., 2015; Le Berre et al., 2016). These studies suggest that in FOK processes, the insula may be related to performance monitoring (Cosentino et al., 2015) or to self-inferential processes for generating accurate future estimations (Le Berre et al., 2016). JOL and FOK are both prospective metamemory judgments and involve similar cognitive processes (Dunlosky and Metcalfe, 2009). Thus, the insula may play similar roles in FOK and JOL. In the JOL process, the insula may be associated with the monitoring of the performance in the memory-encoding phase and the generation of the predictions for future memory test. In contrast to previous FOK studies that used ROI-based approach to analyze the GM volume of the insula (Cosentino et al., 2015; Le Berre et al., 2016), this study used VBM analysis and found that the GM volume of the left mid-insula was specifically correlated with JOL accuracy. Previous studies have indicated that the mid-insula plays an important role in the evaluation of bodily state (i.e., interoception) (Craig, 2011). In the interoceptive process, the mid-insula re-represents the primary neural representations of the state of the body and integrates the representations of the body with other neural inputs to form a combined representation of the individual’s internal and external environment (Craig, 2009, 2011). Thus, it is possible that in the JOL process, the left mid-insula may monitor and process the neural representations of different information or cues related to the performance in the memory-encoding phase and then integrate the cues to generate estimations for future memory test (Koriat, 1997; Kao et al., 2005). In addition, Chua et al. (2009) showed that metamemory process was associated with less activity in brain regions related to the processing of external visual stimuli, and with greater activity in regions responsible for internally directed cognition. Chua et al. (2009) suggest that metamemory is characterized by both a shift toward internally directed cognition and away from externally directed cognition. One possibility is that the left mid-insula may serve as a hub between external representation and internal representation and may be related to the shift toward internally directed representation. Moreover, Critchley et al. (2004) suggest that in the interoceptive process, the insula is related to the subjective feeling states that may underlie the conscious representation of our internal bodily processes. Perhaps the left mid-insula is also associated with the conscious representation of our experience in the memory-encoding phase. Future research should further examine the role of the left mid-insula in the JOL process.

One difference between this study and previous FOK studies is that previous studies have indicated that FOK accuracy is related to the right or bilateral insula (Cosentino et al., 2015; Le Berre et al., 2016), whereas this study revealed a significant relationship between JOL accuracy and the left mid-insula. One possible explanation is that the insula in different hemispheres might be related to prospective metamemory judgments at different stages: JOL is made in the encoding phase, while FOK is made during the memory test (Dunlosky and Metcalfe, 2009). In addition, we also note that results from previous FOK studies are not identical (Cosentino et al., 2015; Le Berre et al., 2016), and it is possible that the different tasks used in this study and previous studies may lead to differences in lateralization. For example, Cosentino et al. (2015) asked participants to learn information about different people and make FOK judgments. They found that FOK accuracy was related to the right insular volume. By contrast, Le Berre et al. (2016) required participants to learn unrelated word pairs and found that the FOK accuracy was related to the GM volume of the bilateral insula. In contrast to previous studies, this study required participants to learn a list of words and perform a recognition test. Future research should use VBM analysis to examine the relationship between the GM volume in the insula and both types of prospective metamemory judgments in a single experiment.

Contrary to our hypothesis, individual differences in JOL accuracy were not related to the GM volume of medial BA 11, which is activated in the JOL process according to previous task-related fMRI studies (Kao et al., 2005; Do Lam et al., 2012; Yang et al., 2015). In addition, our results from resting-state fMRI analysis showed that the functional connectivity of medial BA 11 was not correlated with JOL accuracy. Structural MRI and resting-state fMRI analyses aim to reveal the variability in brain volume or resting-state activity in relevant neural structures that reflect individual differences in performance. It does not strictly follow that the identified region is more active during the relevant tasks, because the relationship between task-related fMRI findings and the results of structural MRI and resting-state fMRI analyses is complex (Mennes et al., 2010; Kanai and Rees, 2011; McCurdy et al., 2013). Future research should combine structural MRI, resting-state fMRI and task-related fMRI to further investigate the neural correlates of the JOL process. In addition, even in previous task-related fMRI studies concerning JOL, the direction of correlation between JOL accuracy and the activation of medial BA 11 has been controversial (Kao et al., 2005; Yang et al., 2015). For example, Kao et al. (2005) required participants to learn pictures of scenes and make JOLs for all pictures. They found that JOL accuracy showed a positive correlation with activation in medial BA 11. In contrast, Yang et al. (2015) asked participants to learn a list of words and make JOLs for only half of the words. Their results revealed that greater JOL accuracy was correlated with a smaller subsequent JOL effect (i.e., difference in activation between trials given high and low JOLs) in medial BA 11. One possible explanation for these conflicting results is that the different tasks used in different studies may affect the relationship between JOL accuracy and medial BA 11. Future studies should examine the effect of the task type on the relationship between JOL accuracy and medial BA 11.

Furthermore, our resting-state fMRI analysis showed that JOL accuracy was associated with a resting-state network that included the functional connectivity between the left mid-insula and different regions. These results are consistent with a previous study indicating that the mid-insula is functionally connected with various brain regions (Cauda et al., 2011). For example, Cauda et al. (2011) indicated that the anterior and posterior insula are involved in two different brain networks: the anterior insula is functionally connected to the frontal cortex, anterior cingulate cortex, parietal cortex and superior temporal gyrus, whereas the posterior insula is functionally connected to the sensorimotor cortex, the supplementary motor cortex and several regions in the temporal cortex and limbic system. In addition, Cauda et al. (2011) suggested that the mid-insula is a transitional area between the anterior and posterior insula and that the two networks overlap in the mid-insula. Thus, the mid-insula may be connected to regions in both networks. Furthermore, our results revealed that the functional connectivity found in this study could be divided into two subsets or factors (Factors 1 and 2; see Table 2 and Figure 4). The functional connectivity within one factor was more closely correlated with each other than with the functional connectivity in the other factor, and the functional connectivity within a single factor may share similar functions (Jacobs et al., 2014). In addition, the two factors could independently predict individual differences in JOL accuracy, and a higher score of one factor was significantly correlated with a higher score of the other.

According to our factor analysis, Factor 1 contains the functional connectivity between the left mid-insula and various regions (including the left parahippocampal gyrus/fusiform gyrus/cerebellum, left SPL/Pcu, bilateral IPL/IPS and right frontal pole) and between the right frontal pole and the left SPL/Pcu. The parahippocampal gyrus is an important part of the medial temporal lobe and plays a key role in the memory encoding of various stimuli (Kirchhoff et al., 2000; Strange et al., 2002). The fusiform gyrus is related to the encoding of phonological or lexical features into episodic memory (Kirchhoff et al., 2000). In addition, the functional connectivity between the left mid-insula and left parahippocampal gyrus/fusiform gyrus is likely related to the posterior insula network, which may play an important role in sensorimotor integration (Cauda et al., 2011). It is possible that in the JOL process, the functional connectivity between the left mid-insula and the parahippocampal gyrus/fusiform gyrus may also be related to the integration of information about the encoding phase. In the JOL process, the left mid-insula may integrate the information about the phonological or lexical attributes of words (from the fusiform gyrus) and other information in the encoding process (from the parahippocampal gyrus), and higher efficacy of this information integration process may predict higher JOL accuracy. In addition, the parietal regions found in the present study (including the SPL, Pcu, IPL, and IPS) are closely associated with episodic memory (Wagner et al., 2005; Baird et al., 2013), and previous studies have revealed that regions in the dorsal parietal cortex, such as the SPL, Pcu, and IPS, are related to top-down attention in memory encoding (Cabeza et al., 2008; Uncapher et al., 2011). Moreover, the functional connectivity between the left mid-insula and left SPL/Pcu, bilateral IPL/IPS and right frontal pole may be related to the anterior insula network, which is associated with attention control (Cauda et al., 2011). The relationship between JOL accuracy and functional connectivity between these regions suggests that the control of our attention in the JOL process may support accurate JOLs. In addition, given the possible role of the mid-insula in the information integration process (Craig, 2011), another explanation for the role of the functional connectivity between the left mid-insula and these regions in the JOL process is that when JOLs are made, the left mid-insula may receive and integrate the information about attention control from these regions.

Moreover, the frontal pole has also been shown to be involved in the JOL process. For example, Ryals et al. (2016) found that transcranial magnetic stimulation (TMS) on the frontal pole could significantly increase JOL accuracy. The frontal pole is the apex of the rostro-caudal hierarchically organized prefrontal cortex (Badre, 2008; Badre and D’Esposito, 2009), and Ryals et al. (2016) suggest that the frontal pole may interact with other memory-processing regions in the JOL process and integrate information from these posterior regions. Thus, the functional connectivity between the right frontal pole and left SPL/Pcu may be associated with the integration of information in the JOL process. Taken together, the functional connectivity in Factor 1 may reflect the process of information integration, which is likely to support accurate JOLs.

In our factor analysis, Factor 2 contains the functional connectivity between the left SPL/Pcu and bilateral IPL/IPS. These regions mainly belong to the dorsal attention network (DAN) and the frontoparietal network (FPN) (Yeo et al., 2011). The DAN and FPN support the top-down control of attention and goal-directed processing (Dosenbach et al., 2007; Cabeza et al., 2008). Previous research has shown that selective attention toward the to-be-remembered stimuli can enhance the processing of the stimuli (Uncapher et al., 2011), which may have promoted JOL accuracy in the present study. Our results indicate that JOL accuracy may be related to the functional connectivity within the DAN and FPN, suggesting a possible role of selective attention in making accurate JOLs. Interestingly, the functional connectivity between the bilateral IPL/IPS in Factor 2 also had a factor loading higher than 0.3 on Factor 1, suggesting that this functional connectivity may be related to the integration of information in the JOL process. Future studies should further investigate the role of the bilateral IPL/IPS in the process of information integration when people make JOLs.

Our behavioral results indicated that individual differences in JOL accuracy did not correlate with recognition accuracy. In addition, we found that the GM volume and functional connectivity of the left mid-insula, which showed significant correlation with JOL accuracy, was not related to recognition accuracy, suggesting that the cognitive processes underlying JOL accuracy and memory performance might be distinct. Recent studies have obtained similar results with brain stimulation techniques (Chua and Ahmed, 2016; Ryals et al., 2016). For example, Ryals et al. (2016) found that TMS on the frontal pole could increase JOL accuracy but showed no effect on recognition performance. These results are also consistent with previous functional MRI studies showing that different brain regions are activated during memory process and JOL, indicating a dissociation between the two processes (Kao et al., 2005; Do Lam et al., 2012). According to the cue-utilization theory, participants do not base their JOLs on direct assessments of memory strength. Instead, their JOLs are based on various cues obtained in the encoding phase, and JOL accuracy depends on whether these cues can accurately predict memory performance (Koriat, 1997; Dunlosky and Metcalfe, 2009). Thus, JOL accuracy reflects the correspondence between JOL and memory performance (rather than memory performance itself), and the brain regions that are essential for JOL accuracy may not be involved in the memory process. Future studies should further disentangle the neural correlates of memory performance and JOL accuracy.

Our study has several limitations. First, the sample size (35 participants) in the present study was not large enough. Second, similar to all VBM and resting-state functional connectivity studies, our results are correlational in nature and do not imply any direct causality between individual differences in JOL accuracy and the GM volume of the left mid-insula or resting-state functional connectivity. Although individual differences in brain structure and resting-state activity are closely linked to differences in brain function (Kanai and Rees, 2011; Lee et al., 2013), more evidence is needed to further examine the results of this study and our explanation concerning the neural mechanisms related to individual differences in JOL accuracy. Third, our conclusions are not exhaustive, as they do not exclude the possibility that other regions may also be related to JOL accuracy without any difference in the GM volume or resting-state activity. Finally, we did not investigate whether individual differences in JOL accuracy are related to other cognitive functions. For example, a previous study suggested that executive function may be related to metamemory processes (Destan and Roebers, 2015). Future studies should use larger sample sizes and combine event-related fMRI, structural MRI and resting-state fMRI techniques to further examine the neural mechanisms concerning individual differences in JOL accuracy. In addition, future studies should also investigate the similarities and differences between the neural mechanisms of individual differences in JOL accuracy and other cognitive functions, such as executive function.

This study was carried out in accordance with the recommendations of the Institutional Review Board of the State Key Laboratory of Cognitive Neuroscience and Learning, Beijing Normal University with written informed consent from all subjects. All subjects gave written informed consent in accordance with the Declaration of Helsinki.

XH and LL designed the experiment. XH, WC, JZ, NS, WW, and LL performed the experiment. XH, ZL, and LL analyzed the data. XH, ZL, CL, and LL wrote the manuscript.

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.