Changes in neural activity during the combining affect labeling and reappraisal

Reappraisal, an emotion regulation strategy, is an effective way of controlling negative emotions. Conversely, it is known that affect labeling regulates negative emotions using a different process from reappraisal, and it is possible that the combined use of affect labeling and reappraisal might enhance the control of negative emotion. In this study, we compared the brain activity during combined use of affect labeling and reappraisal to negative emotion with the sole use of reappraisal by using fMRI. The participants performed a reappraisal after affect labeling to negative emotion which induced by negative image. In comparison to the sole use of reappraisal, increased activity was found in the bilateral inferior frontal gyrus and medial frontal gyrus, whereas decreased activity in the right amygdala. Furthermore, based on the results of a functional connectivity analysis using the seed region of the right amygdala, it was determined that coupling with the right amygdala increases due to the combined use of affect labeling and reappraisal. The results reveal that affect labeling of negative emotion potentially effects on reappraisal, which has implications for the more effective use of reappraisal.


Introduction
This function of changing the response to emotion is termed as emotion regulation (Gross, 2002). Emotion regulation is an automatic and regulated process that helps with the outbreak, intensity, maintenance, and adjustment of emotional states. It can also be referred to as the process by which an individual controls when, how, and in what way certain emotions are experienced and manifested. In a study of emotion regulation by Gross (Gross, 2002;Gross and John, 2003), the development of an emotion regulation process model was presented according to the timeline of the continuous processing of emotion experience from situation that evoke emotion to expression of emotion. In this process model, the strategy for cognitively changing evaluations of elicited emotion and situation is defined as "reappraisal (Gross and John, 2003), " which is a core element of Cognitive Behavioral Therapy (CBT; Ellard et al., 2010) for affective disorders, such as depression and anxiety disorders. Furthermore, reappraisal is one component of the CBT protocol aimed at correcting incorrect threat evaluation, in which there is a propensity to exaggerate the threat or overestimate the possible catastrophe (Smits et al., 2013). As reappraisal has been reported to have positive effects on psychological adjustment, such as reducing negative emotions (Gross and John, 2003;Ochsner et al., 2004;Sheppes and Meiran, 2007), it seems reasonable to include reappraisal in psychotherapy. Alternatively, in the meta-analysis by Webb et al. (2012), in cases in which emotional experiences are complex and unclear, the effects from the use of reappraisal may diminish. Considering the limited effects of reappraisal, investigating strategies to enhance the performance of reappraisal is of particular significance.
Research on reappraisal has yielded several findings on the factors that influence performance of reappraisal. Individual differences in working memory capacity (McRae, Jacobs et al., 2012), belief in the success of reappraisal (Gutentag et al., 2017), and the attitude toward one's own emotional experiences (Garland et al., 2015) have been correlated with the performance of reappraisal. In addition, individual differences in emotion recognition and their relationship to reappraisal have recently been studied, and it has been reported that changing everyday emotion recognition has a partial impact on the success of emotion regulation, including reappraisal (Kalokerinos et al., 2019). It has been revealed that a low level of emotion recognition affects the difficulty of reappraisal and the worsening of negative emotions (Boden and Thompson, 2015;Subic-Wrana et al., 2014). Further, it has been assumed that the connection between emotion recognition and reappraisal plays an important role in the pathology of affective disorders. In studies conducted by Suri et al. (2015) and Vine and Marroquín (2018), it was reported that raising the level of emotion recognition may achieve a positive effect in terms of strengthening reappraisal, thereby making it easier to perform reappraisal. Based on these findings, if it can be verified that emotion recognition enhances reappraisal and this mechanism, it may be useful for patients with affective disorders.
Recent neuroimaging studies about reappraisal has been clarified that the use of reappraisal to negative emotion leads to increased activity in large parts of the prefrontal area. Based on the results of a metaanalysis, the implication is that the use of reappraisal shows increased activity in the ventrolateral prefrontal cortex (VLPFC), dorsolateral prefrontal cortex (DLPFC), dorsomedial prefrontal cortex (DMPFC), and, simultaneously, increased activity in the anterior cingulate cortex (ACC) and parietal lobule (Kohn et al., 2014;Messina et al., 2015). It has also been deduced that, in contrast to the above-mentioned increase in activity in the cortex, activity in the amygdala decreases. Hierarchical cluster analysis of these brain regions (Morawetz et al., 2020) clarified that the prefrontal region and ACC are categorized into clusters of neural circuits associated with executive functions, response inhibition, and language processing associated with the execution of reappraisal. On the other hand, it has also been suggested that decreased amygdala activity during reappraisal is associated with suppression of emotion generation and emotion appraisal processes. Moreover, a meta-analysis of changes in functional connectivity between the amygdala-prefrontal region following reappraisal of negative emotion (Berboth et al., 2021) revealed that functional connectivity of the DLPFC, VLPFC, and DMPFC with the amygdala consistently increased during reappraisal. Combining these findings, the DLPFC and VLPFC during reappraisal are thought to be involved in the selection of target of emotion regulation, and the linguistic process that change interpretations of emotions. Because executive functions such as switching attentional focus and maintaining cognitive processing are required to accompany these complex cognitive processes, it is possible that a vast amount of cognitive resources are used in the process. It may also be that the subcortical structures such as amygdala is suppressed by the top-down activity of the cortical circuits accompanying the performance of reappraisal, thereby resulting in a decline in emotional response.
Based on these neuroimaging study relating to reappraisal, to enhance the effects of reappraisal, it is necessary to stimulate the cognitive processing that mediates the execution of reappraisal and is responsible for its success. Conforming to the process model of emotion regulation advocated by Gross, prior to performing reappraisal, it is necessary to deploy attention on the emotion and the situation that incite emotion, that is the target of reappraisal, and to evaluate experiences of the own emotion. Based on process model of emotion regulation, Moyal et al. (2014) expanded on the emotion regulation process model and proposed that, prior to reappraisal, it is necessary to assess one's internal state in connection to the emotion and to specify the emotion to be regulated, and that this emotion recognition is essential to the success of reappraisal. A method by which individuals perform emotion recognition is to apply a linguistic label. Although it does not possess the goal of control emotion, affect labeling (AL; Lieberman et al., 2007), where a linguistic label is applied to emotional experiences and emotional stimulus. AL is known to reduce negative emotions that arise by means of the linguistic labeling of emotional stimulation and labeling of one's own emotional experiences (Torre and Lieberman, 2018). In addition, brain activity related to the execution of AL includes that activation of the prefrontal region, which is similar to reappraisal. Among these regions, the right VLPFC is strongly connected to AL execution, and it has been reported that the activation of the right VLPFC is particularly strong in comparison to reappraisal (Burklund et al., 2014). It has also been shown that, similar to reappraisal and distraction, amygdala activity decreases due to AL with regard to negative emotion (Lieberman et al., 2007;Tupak et al., 2014). Although no conclusion has been derived regarding the mechanism by which AL is able to regulate emotion, the assumption is that labeling one's emotions and emotion-provoking stimuli draws attention away from emotional experiences, and the verbalization of emotion may shift the emotional experiences into an abstract format, thereby creating psychological distance between oneself and one's emotions (Torre and Lieberman, 2018). Presently, although there is limited evidence that allows for a definitive explanation of the psychological mechanisms of AL, it is assumed to be a process that results in the automatic control of emotion even if there is no intention about the regulation of emotion. Alternatively, it has been recently reported that performing reappraisal after AL impairs the effect of reappraisal on reducing negative emotion. Nook et al. (2021) conducted a behavioral experiment with healthy participants where the participants engaged in naming of emotions they felt after they were presented with negative images, which was immediately followed by the performance of reappraisal to reduce the intensity of the negative emotion. In comparison to cases in which reappraisal is conducted normally without naming the emotion prior to reappraisal, when reappraisal was performed after naming, post-reappraisal negative emotion was reported to be higher. This finding contrasts with the existing theories of emotion regulation and experiential intuition from psychotherapy, which reveals that emotion recognition is adaptive and that it enhances emotion regulation. Therefore, it has not been demonstrated whether emotion recognition such as AL enhances reappraisal. An investigation of how emotion recognition impacts reappraisal from a neuroscience perspective may enable further multifaceted examinations.
In view of the above, the present study uses functional neuroimaging to investigate the impact of the recognition of negative emotional experiences using AL on subsequent reappraisals. On the behavioral level, we expect that AL on negative emotions will facilitate the regulation of negative emotions in subsequent reappraisals. Simultaneously, we expect an increase in activity in regions centering on the prefrontal cortex (DLPFC, VLPFC, DMPFC) that are related to emotion regulation, and in the right VLPFC activity that was found to be activated by AL. It should also be accompanied by a decrease in amygdala activity, which is related to generation of negative emotions. Further, an increase in functional connectivity between the amygdala and prefrontal cortex is expected when conducting reappraisal after AL. To investigate the above, stimuli arouse negative emotions are presented to healthy participants, who are instructed to perform AL and reappraisals, after which the degree of negative emotion is assessed. Brain activity was simultaneously measured during the task using fMRI to examine whether there were differences in prefrontal cortex and amygdala activity between the AL and reappraisal conditions and reappraisal-only condition. Furthermore, it was decided to analyze activity in the prefrontal cortex that couples with the amygdala as the seed by calculating functional connectivity. When calculating the difference in brain activity between conditions, Neurosynth was used to limit the analysis to brain regions that are consistently activated in relation to emotion regulation.

Participants
Participants were healthy individuals aged 20 or above with no current mental and neurological disorders (n = 21; 11 males, 10 females; mean age = 25.0, range 20-35). All participants were right-handed individuals whose native language is Japanese. They all had education of at least 12 years or more. First, the experimenter provided an explanation regarding the study's purpose to all participants, post which they signed a research participation consent form after their understanding of details had been confirmed. Fig. 1 Marchewka et al., 2014) and presented to participants. NAPS is standardized based on ratings of the emotional valence and arousal of the stimulus and basic emotional dimensions. The experiment task comprised the following four conditions. Affect Labeling and Reappraisal (AL_R) condition: Negative images are presented along with four affective labels describing the content of the images ("sad", "anger", "fear," and "disgust"). Participants are instructed to push a button to select the adjective that is closest to the negative emotion that the image incites (AL), and then to perform reappraisal to reduce the negative emotion. Procedure of AL was based on Burklund et al. (2014), but differed in that participants were forced to choose one of these affective labels. Reappraisal (R) condition: Affective labels were not presented, but four "xxx" were displayed below the negative images, and participants were instructed to press any button only once (to match the motor response with the AL_R condition). Apart from that, the process was similar to the AL_R condition. Control (C) condition: Negative images are presented and participants are asked to observe passively without performing any strategy for reducing the negative emotion, including reappraisal. Apart from that, the process was similar to the R condition. Neutral (N) condition: Only neutral images are shown; the remaining process was similar to the C condition.
The experimental task comprised a block design, with each condition instruction presented for 4000 ms at the beginning of each block. Then, one image stimulus was presented for 4000 ms per trial, followed by a blank of 2000 ms. In each block, five trials are presented consecutively. Immediately after completing the five trials, the participants were asked to rate the strength of negative emotion in the block on a four-point scale (from 1: none to 4: extremely high). The rating scale for each block was presented for 4000 ms. Each condition was performed in a pseudorandom order of four blocks each, for a total of 16 blocks.

Procedure
After obtaining consent to participate in the study, the experimental task was explained outside the MRI scanner. Reappraisal strategies were explained in detail to perform reappraisal of negative emotions that occur due to the stimuli in the AL_R and R conditions. Although it was highlighted that there are some reappraisal substrategies, reinterpretation was used in this study (i.e., rethinking the meaning of a stimulus to change its emotional impact; McRae, Ochsner et al., 2012). Specifically, participants were presented with negative images not used in the experiment (e.g., a photograph of a crying child), and were instructed to attempt to change their interpretation of the image to reduce the negative emotion. Based on the images used by experimenter for the purpose of explanation, the participants were provided with examples of reappraisal, such as "crying is not always limited to bad things," and "people sometimes cry because they are happy." Next, as practice, one trials were conducted for each condition. After practicing, the participants were asked to verbally report how they attempted to change their interpretations through practice of the AL_R and R conditions, and if it was unclear whether it was a reappraisal, then the explanation and practice session were conducted again (all participants achieved reappraisal after one explanation and exercise). After the task explanation and practice, the participants entered the MRI scanner and T1 structural images were recorded. Then the fMRI scans were conducted while performing the task. After scanning the fMRI, the physical condition of the participants were checked, and the experiment was ended (none of the participants had any physical complaints). The time required from the explanation of the study until the completion of all procedures was approximately 40 min (20 min for study explanation and 20 min for MRI scanning). After completion of experiment, the participants were asked how they attempt to change interpretation during AL_R and R condition, and what difficult they feel during experiment. No participants reported performing reappraisal remarkably differently from experimenter's instruction. This study was approved by the Otomon Gakuin University Research Ethics Committee (number 2017-20).

fMRI acquisition
The fMRI procedure was performed using a Verio 3.0 tesla (Siemens, Munich, Germany). A total of 325 scans were obtained for each participant using T2 * -weighted, gradient echo, echo planar imaging (EPI) sequences. Each volume comprised 32 slices, with a slice thickness of 4 mm with no gap, and covered the entire cerebral and cerebellar cortices. The time interval between two successive acquisitions of the same image (TR) was 2000 ms, echo time (TE) was 30 ms, and flip angle was 80. The field of view (FOV) was 256 mm and the matrix size was 64 × 64, with voxel dimensions of 4 × 4 x 4 mm. Scan acquisition was synchronized to the onset of each trial. After functional scanning, structural scans were acquired using a T1-weighted gradient echo pulse sequence (TR = 2250 ms; TE = 3.06 ms; flip angle 9; FOV 256 mm; voxel dimensions of 1x1x1 mm), which facilitated localization.

Data analytic plan 2.5.1. Behavioral data
Negative emotion rating data from three participants were discarded due to technical problem (misconfiguration of the response pad used to obtain button-press responses). Data for the remaining 18 participants were included in analysis of behavioral data. Ratings for each condition were averaged across blocks, and the main effects across conditions were investigated using repeated-measures ANOVA. In all trials, although it was confirmed that all of the participants pushed the button each time, the button pressing response time was recorded.

fMRI data
Image processing and statistical analyses were conducted using statistical parametric mapping (SPM12). The first three volumes were discarded because the MR signals were unsteady. All EPI images were spatially normalized using the Montreal Neurological Institute (MNI) T1 template for group analysis. Imaging data were corrected for motion and smoothed with a 4 mm full-width half-maximum Gaussian filter. At the individual participant level, a whole-brain voxel-by-voxel multiple linear regression model was used to analyze imaging data.
The individual model comprised the covariate of no interest (realignment parameters). A general linear model analysis was then used to create contrast images for each participant, summarizing differences between conditions. We created the following three corresponding contrasts for the first-level analysis for each participant to subtract activity associated with N condition: AL_R comparing to N condition, R comparing to N condition, and C comparing to N condition. Next, these individual contrast images were used at the group-level paired t-test, random-effects analyses (AL_R vs. R condition, AL_R vs. C condition, R vs. C condition). These paired t-tests were restricted to a limited set of brain regions related to emotion regulation based on metaanalytic data to increase statistical power. The automated meta-analytic database Neurosynth (Yarkoni et al., 2011) was used to create forward inference statistical maps related to the terms "emotion regulation" (Multislice axial view and region labels of this statistical map are shown in supplemental material ( Fig. 2S and Table 1 S)). This statistical map was used as an including mask in above paired t-tests (significant level was uncorrected p < .001 as peak voxel level and p < .05 as cluster level). In addition, according to our a priori hypotheses, we reported amygdala significant at p < .05 family-wise error corrected at the cluster level after small volume correction (SVC) in SPM using the bilateral anatomical masks created with the WFU Pickatlas toolbox (www.fil.ion. ucl.ac.uk/spm/ext/). All coordinates are reported in MNI space and peak activations were labeled according to the Anatomical Automated Labeling.
We analyzed functional connectivity with right amygdala as seed region by using generalized psycho-physiological interaction (gPPI) analysis in Conn toolbox ver. 21a. Further, we inputted all condition to PPI modeling. Seed region was defined as radius sphere (6 mm) in the right amygdala based on result of group-level paired t-test between AL_R and R contrast. Task condition of interest included AL_R, R, and C conditions. AL_R vs. R, AL_R vs. C, and R vs. C contrasts were examined. We set the statistical threshold level was uncorrected p < .001 at peak level, and FWE corrected p < .05 at cluster level. Button pressing response times of each condition are shown in Table 1 S (supplemental material). Repeated-measures ANOVA for response time revealed significant effect between conditions (F(3, 51) = 45.60, p < .001 η 2 p = .73). Response time in AL_R condition was slower than in the other three conditions (p < .001, Holms). Response time in R condition was slower than C (p < .05) and N (p < .01) conditions. Response time in C condition was slower than N condition (p < .01).

fMRI data
Paired t-test results for each contrast are summarized in Table 1 and Fig. 2. In case of contrast AL_R comparing to C or R condition, we identified significant brain activity in the bilateral inferior frontal gyri (Brodmann Area [BA] 44 and BA 45 on the left, BA 47 on the left and right, BA 9 and BA 46 on the left) and bilateral medial frontal gyri (BA 8 and BA 9). Figs. 3a, 3b shows activation map from AL_R comparing to R and C condition. The AL_R comparing to C contrast also exhibited activation of the left middle frontal gyrus (BA 10), left thalamus, left parietal cortex, left temporal gyrus, left precentral gyrus, left lingual gyrus, bilateral caudate, and right fusiform gyrus. There were many overlaps the contrasts AL_R comparing to R and C condition, we represented these overlaps to Fig. 2S (supplemental material).
There was no significant increased activation in R condition comparing to AL_R condition. Alternatively, contrast of R comparing to C condition showed increased activation of the left inferior frontal gyrus (BA44, BA45 and BA47), left medial frontal gyrus (BA6, BA8, BA9, BA32), left precentral gyrus, left temporal gyrus, left angular gyrus, and right thalamus (Fig. 1c). There was no significant increased activation in C condition comparing to both AL_R and R conditions. Because there was a significant difference in response time, the increased brain activity in AL_R vs. R condition may have reflected differences in the mental effort required for the response during choice, rather than affect labeling. To examine this possibility, we conducted the correlation analysis between response time in AL_R and each brain activity in the AL_R > R contrast. However, there were no significant correlation between response time and these brain activities. Correlation matrix and statistic values are presented in Fig. 3S (supplemental material).
ROI analysis revealed more decreased activity of the right amygdala in AL_R comparing to R condition (x = 32, y = 0, z = − 22, t = 4.52, cluster size = 9 voxel, SVC) and C condition (x = 32, y = − 2, z = − 24, t = 4.75, cluster size = 5 voxel, SVC). Based on results of the ROI analysis, we performed gPPI analysis for contrasts of three conditions (AL_R, R, and C) in the present study. Seed region was defined as radius sphere (6 mm) at peak voxel coordinates of the right amygdala in AL_R comparing to. R condition. gPPI analysis revealed significant increasing connectivity between seed region and right parietal cortex (BA 40), right middle frontal gyrus (BA 8, BA 9), and left parietal cortex (BA 40) in AL_R comparing to C condition. Some voxels (six voxels) in the right middle frontal gyus were overlapped with the forward inference statistical maps related to the terms "emotion regulation" made by Neurosynth. Additionally, decreasing connectivity was observed between seed region and bilateral fusiform gyrus (BA 19, BA 37) and left middle temporal gyrus (BA 37). No significant functional connectivity was observed in other contrasts. Table 2 and Fig. 4 shows coupled regions which showed increased or decreased connectivity to seed region.

Discussion
In this study, fMRI was used to examine whether brain activity during the combined use of AL and reappraisal differs from that of reappraisal alone. The fMRI data revealed that the combination of AL and reappraisal shows an increase in activity in brain regions, which, according to previous studies, are known to contribute to the implementation of emotion regulation. Specifically, it was deduced that the activity in the lateral and medial prefrontal cortices, including the bilateral DLPFC and VLPFC, increases while the activity in amygdala decreases. Additionally, a gPPI analysis revealed that the right amygdala activity is linked to the right middle frontal gyrus when the use of AL and reappraisal is combined.

Function of the brain activation while preforming reappraisal with affect labeling
We found increasing activity of the right inferior frontal gyrus and a decrease in activity in the amygdala during combination use of AL and reappraisal in comparison to reappraisal alone. The right inferior gyrus contributes to control processes relating to information regulation, including motor response (Berkman and Lieberman, 2009), emotion distraction (Dolcos and McCarthy, 2006). Additionally, Hallam et al. (2015) reported that, when participants are told to simply perform the process of reappraisal rather than using it with the objective of controlling emotion, there is an increase in right inferior frontal gyrus activity and a decrease in amygdala activity. Considering this based on the study findings, the right inferior gyrus has an important role in the regulate of emotion, suggesting that the increased activity of the right inferior gyrus promotes amygdala inhibition. Part of the right inferior frontal gyrus overlaps with the VLPFC, which is partially consistent with the results of prior studies reporting increased activity in the right VLPFC due to AL (Burklund et al., 2014;Lieberman et al., 2007;Tupak et al., 2014). Apart from language processing, the VLPFC has functional characteristics relating to emotional processing, social cognition, and behavior control. Anatomically, the VLPFC is directly connected to the amygdala, and afferent projections from the amygdala reach the VLPFC via the anterior insula (Ray and Zald, 2012). Additionally, it has been highlighted that the VLPFC has close neurological connections to the medial prefrontal gyrus and superior temporal gyrus, and it functions as a hub that regulates different brain network activities by relaying information between the subcortical regions and prefrontal regions that do not have a direct connection (Kohn et al., 2014). The right VLPFC is assumed to play a role in controlling the amygdala by relaying the results of emotion regulation to subcortical regions rather than being involved in the active processes required for regulation of emotions. Although neither reappraisal with AL nor reappraisal alone were effective in reducing subjective negative emotion rating, the fact that reappraisal with AL resulted in an increase in right inferior frontal gyrus activity, including the VLPFC, and a decrease in amygdala activity suggests that the activation of the right inferior frontal gyrus due to AL may have a promotional effect on reappraisal. However, the results of the gPPI analysis revealed no significant functional connectivity between the right amygdala and right VLPFC. Therefore, there is no clear evidence that the activity of the right VLPFC caused by the combined use of AL and reappraisal in this study directly inhibits the amygdala.
The combined use of AL and reappraisal also produced a significant increase in activity in regions, including the medial frontal gyrus and left inferior frontal gyrus. In comparison to the sole use of reappraisal, the major activation of these brain regions may reflect differences in effort, task difficulty, attention, and participation. The left inferior frontal gyrus engages in linguistic processes during emotion regulation, supports the active reinterpretation of the meaning of emotional stimulation via AL, and may even promote the selection and implementation of a target-oriented reappraisal (Buhle et al., 2014;Morawetz et al., 2020).   The medial frontal gyrus is robustly recruited during tasks that require reasoning with respect to mental states (Dixon et al., 2017), which is why participants focused their attention toward their own emotional condition when attempting to regulate their emotions, and the process of assessing the emotional condition was reinforced by performing AL in advance. Because activity in this region is nonspecific to the direction of emotional change, it appears to be specifically involved in cognitive control processes that regulate the current emotional state regardless of emotional valence (Frank et al., 2014). Additionally, the medial frontal gyrus, which undergoes increased activity due to the combined use of AL and reappraisal, spatially overlaps with the DMPFC. The DMPFC may refine the emotional significance of the stimulus and also support meaning-related and self-introspective processes associated with the recognition of one's own emotional state (Buhle et al., 2014). In other words, AL may have enhanced the process of reappraisal, which is the process of evaluating one's own emotional experiences and reinterpreting their meaning.
In comparison with the combined use of AL and reappraisal, when reappraisal alone is conducted, no specific brain activities were observed in the results of this study. Burklund et al. (2014) made a direct comparison of AL and reappraisal, and reported that the performance of reappraisal leads to greater activity in the VMPFC in comparison with AL. This implies that AL and reappraisal drive partially different neurological circuits, suggesting that they are qualitatively different emotion regulation strategies. If the combined use of AL and reappraisal produced brain activity related solely to the performance of AL, then while making a direct comparison with the sole use of reappraisal, we would expect to observe brain regions that are activated by the sole use of reappraisal. However, because no brain regions showed a significant increase in activity with reappraisal alone compared to AL, it follows that the combination of AL and reappraisal drives both reappraisal-related brain activity and AL-related brain activity, and it can be interpreted as producing more emotion-regulation related brain activity.

Functional coupling between the amygdala and prefrontal cortex in reappraisal with affect labeling
In the gPPI analysis, the right amygdala was used as a seed region to examine the regions where functional connectivity was significantly enhanced during the combined use of AL and reappraisal. As a result, in comparison with the C condition, there was an increase in functional connectivity with the right middle frontal gyrus that partially overlaps the DLPFC. The increase in amygdala and DLPFC functional connectivity due to the performance of reappraisal is consistent with prior studies (Banks et al., 2007;Berboth and Morawetz, 2021;Kanske et al., 2011). The combined use of AL and reappraisal has also been reported to show increased activity in the wider lateral prefrontal region, including the DLPFC, which often shows engagement in working memory and response selection (Menon and D'Esposito, 2022;Miller and Cohen, 2001). Additionally, it has been proposed that the DLPFC is the central regulatory brain region for regulating emotion that is produced by the amygdala, insula, and VLPFC (Phillips et al., 2003), thereby suggesting that it plays an important role in emotion regulation (Ochsner and Gross, 2005). As VLPFC and DLFPC are assumed to play a key role in the interaction of attention and emotion regulation (Dolcos et al., 2020), it is possible that the increased allocation of attention to emotional information due to AL is fixated on information regarding the emotion to be regulated, and there is a corresponding increase in resources required for the implementation of reappraisal.
When focusing on the findings of meta-analytic connectivity modeling, Kohn et al. (2014) reported that the DLPFC shows wide area coactivation caused by performing of various emotion regulation along with the ACC, angular gyrus, insula, and middle frontal gyrus. However, the DLPFC has no strong relationship to emotional processes; rather it is active in various cognitive processing, such as working memory, social cognition, and general cognition, suggesting that the DLPFC plays a general role in cognitive control regardless of emotional or nonemotional content. Anatomically, the DLPFC is positioned in a location that regulates a wide range of behavioral responses from various motor actions (Cieslik et al., 2013) to approach and avoidance, via its connections to the ventral striatum (Haber and Knutson, 2010). As the DLPFC does not have direct anatomical neural connection to the amygdala (Ray and Zald, 2012), it may perform indirect regulation of subcortical regions that are involved in emotional generation (Wager et al., 2008). Alternatively, based on the contrast between the combined use of AL_R and C condition, functional connectivity decoupling was observed between the amygdala and bilateral fusiform gyrus. This implies that in comparison with AL_R condition, no emotion regulation is conducted in C condition; however, the coupling between the amygdala and fusiform gyrus is strengthened when naturally viewing negative images (during C condition). Ferri et al. (2016) reported that when comparing the intentional attentive regulation of nonemotional information in negative image stimuli with the natural observation of negative images, there was an increase in functional connectivity in the amygdala (seed) and fusiform gyrus (coupled region) during natural observation. Based on the neuroanatomical connections of the amygdala, the amygdala operates to increase neurological activity in the fusiform gyrus, which, in turn, increases the potential for visual information containing emotional value to reach the conscious level Duncan and Barrett, 2007). However, there are also reports stating that the functional connectivity of the amygdala and fusiform gyrus increases during reappraisal (Stephanou et al., 2016). It remains uncertain as to how meaningful is the coupling and decoupling of the amygdala and fusiform gyrus during emotion regulation or reappraisal.

Limitations
Several issues with the method of this study limit the interpretation of the findings obtained. First, there was no AL only condition in this study. For this reason, the difference in brain activity between the AL_R condition and the R condition may occur when using AL alone rather being a synergistic effect of AL and reappraisal. Future study should be focus on comparison between the sole use of AL and the combined use of AL and reappraisal. Second, there was no effects of reappraisal to the subjective negative emotion rating when participants performing AL_R and R conditions, the participants may not have fully understood the method of performing reappraisal. However, explanations and practice of task instruction were conducted prior to the fMRI scan until the participants understood the instructions. Although there is no quantitative data, participant's impression as reported by post-experimental retrospective evaluations, the overall strength of the negative image arousing negative emotion was very high, and there were several reports stating that it was difficult to regulate negative emotions despite their intention to perform reappraisal. Because of the excessive strength of the stimulation from images selected from the NAPS, it is hypothesized that the participants were unable to alleviate their own subjective negative emotional response. Third, the sample size in this study was relatively small. We used Neuropower(http://neuropowertools.org/neuropower/) (Durnez et al., 2016) tool to calculate a posteriori the statistical powers of the AL_R>C, AL_R>R, and R>C contrast maps analyzed in this study. As a result, the powers of these contrasts were less than 0.8 for these contrasts (AL_R > C: power =0.75; A_R > R: power =0.65; R>C: power =0.70). Despite the fact that this study included 21 participants, it was not possible to recruit a large number due to time constraints. For this reason, the robustness of the present results has not been completely verified. It is necessary to investigate reproducibility based on data from a larger sample size to clarify what brain activity is induced by reappraisal in combination with AL. Fourth, this study only examined frontal regions that showed functional connectivity with the amygdala. Therefore, causal relationships between these regions are unknown. Considering that previous studies (Morawetz et al., 2017;Nicholson et al., 2017) have reported on directional interaction between the amygdala and frontal regions using Dynamic Causal Modeling (DCM), we have also performed DCM in present study according to (Zeidman et al., 2019). However, only 17 participants were able to extract VOI data from individual data of effects of interest (uncorrected p < .05). Goulden et al. (2012) simulated that it is preferable to obtain at least 20 participants when conducting DCM. reported that at least 20 participants are required when performing DCM. Because the sample size for this study was below this recommendation, we discontinued subsequent analysis for DCM.

Conclusion
By hypothesizing that emotion recognition processing is a factor promoting reappraisal, this study examined brain activity and functional connectivity between brain regions when AL was performed on negative emotional experiences, followed by reappraisal. As a result, in comparison to the sole use of reappraisal, the combined use of AL and reappraisal produced increased activity in the lateral and medial frontal regions, and decreased activity of the amygdala. Moreover, amygdala increased the connectivity with the right lateral frontal cortex, including DLPFC. These results suggest that performing AL on the emotion to be regulated produces brain activity relating to subsequent reappraisal, and increases top-down regulation of subcortical regions, especially the amygdala. While a more multifaceted study will be required, these insights may be effective in improving the efficacy of psychotherapy, such as CBT, which comprises learning regulatory strategies for negative emotions.

Funding source
This research was supported by JSPS KAKENHI Grant number 19K03354. The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Ethics approval and consent to participate
Research protocol of this study was approved by ethics committee of Otemon Gakuin University(reference number:2017-20). Participation is voluntary and all respondents will provide written informed consent before inclusion. Otemon Gakuin University is former affiliation of the corresponding author.