Self-forgiveness is associated with increased volumes of fusiform gyrus in healthy individuals

Self-forgiveness (SF) involves a process through which negative moral emotions directed at the self are replaced by benevolence and acceptance. Lower SF scores can be associated with less self-compassion, higher psychological distress, and lower life dissatisfaction. However, neural correlates of SF have not been investigated yet. We enrolled a total of 79 healthy individuals. The Self-Forgiveness Scale (SFS), Self-Compassion Scale (SCS), Connor-Davidson Resilience Scale (CD-RISC), Beck Depression Inventory-II (BDI-II), and Beck Anxiety Inventory (BAI) were evaluated. Voxel-wise correlational analyses showed a significant positive correlation between the total SFS scores and gray matter volumes (GMVs) in the fusiform gyrus (FG). In addition, the GMVs in the FG were significantly positively associated with the total SCS and CD-RISC scores and negatively correlated with the total BDI-II and BAI scores. These findings suggest that the FG related to the mirror neuron system might be a neural correlate of SF. Furthermore, its increased volumes of FG in healthy individuals can be associated with the capacity to overcome stressful life events.


Scientific Reports
| (2023) 13:5505 | https://doi.org/10.1038/s41598-023-32731-0 www.nature.com/scientificreports/ networks. Therefore, it may be inferred that the fronto-temporo-limbic regions play an important role in generating moral emotions. Moreover, SF was associated with high emotional well-being or self-esteem. Those who are good at SF are more friendly and sympathetic than those who are not 7,22 . Furthermore, individuals with a high SF have healthy psychological characteristics. Another study found that people good at SF feel less shame 23 . Those who genuinely forgive themselves have lower levels of neuroticism 16 and lower levels of psychological distress, such as anxiety and depression 7,22,24 . Therefore, it is essential to identify neural correlates associated with high SF that are correlated with changes in low levels of psychological distress, such as anxiety or depression. However, there are only a few studies on the direct structural neural correlations of SF among healthy individuals.
The present study investigated the structural neural correlates of SF in whole-brain regions of healthy individuals. It was hypothesized that (1) there would be a correlation between SF and gray matter volumes (GMVs) in moral emotions such as shame-or guilt-related brain regions, and (2) there might be a relationship between SF-related GMVs and psychological characteristics (self-compassion, psychological resilience, psychologicalemotional distress, and life satisfaction) in healthy individuals. Therefore, we will perform whole-brain voxelwise analyses to detect the neuroanatomical neural correlates of SF in healthy individuals and determine the relationship between psychological characteristics and SF-related brain regions.

Participants.
A total of 79 (36 men and 43 women) healthy right-handed individuals were recruited from the local Korean community through advertisements. All participants were between 23 and 61 years old (mean 36.68 years), underwent magnetic resonance imaging (MRI), and completed a psychological assessment such as SF.
Healthy adults were excluded based on the following criteria: history of major psychiatric disorders, including schizophrenia spectrum disorders, mood disorders, anxiety disorders, and substance use disorders; neurological disorders and traumatic brain injury; pregnancy; and any prohibition of brain MRI scanning. Additionally, none of the participants had an individual or primary-relative family history of psychiatric disorders according to the review of the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision (DSM-IV-TR) screening by interview.
All procedures were performed after review and approval by the Institutional Review Board of the CHA Bundang Medical Center (2019-05-030). After the participants had a detailed explanation of the study, written informed consent was obtained from the latest version of the Declaration of Helsinki. Further, the principles of Good Clinical Practice were acquired.
Dispositional self-forgiveness assessment. The Self-Forgiveness Scale (SFS) by Kim et al. measures dispositional SF as the process of responding by changing from a negative and punitive response to understanding and generosity about one's actions that are inappropriate for one's thoughts 25 . This scale consisted of 19 items used among healthy individuals, each rated on a 1-to 5-point scale, with the SFS score ranging from 19 to 95. The higher the total SFS scores, the greater the SF score in healthy individuals. Additionally, we used the three subscales of the SFS consisting of "acceptance and improvement, " "transferring of responsibility, " and "negative affect, thought, and behavior. " First, the "acceptance and improvement" subscale consists of accepting oneself to make mistakes and believing that one can do better in the future. Next, the "transferring of responsibility" subscale contains contents such as blaming others, disapproval, and avoidance and indicates a tendency to avoid responsibility for one's wrongdoing and attribute it to others or circumstances. Additionally, the "negative affect, thought, and behavior" subscale refers to the tendency to be unable to escape from negative affect, thought, and behavior regarding one's faults. Among the three subscales of SF, "transferring of responsibility" and "negative affect, thought, and behavior" have opposite meanings to SF. These two subscales were therefore counted as inverse scores when calculating the total SF scores, whereas they were counted as untransformed scores when performing the analysis for each subscale. Moreover, the SFS has a high Cronbach's α coefficient of 0.87. To obtain the total SFS scores, the inversely transformed values of "transferring responsibility" and "negative affect, thought, and behavior" were used.
Other psychological characteristics assessments. To evaluate the levels of self-compassion, we used the Korean version of the Self-Compassion Scale (SCS) by Neff and Kim et al., a self-reported test consisting of a 5-point scale with high internal consistency (Cronbach's α = 0.90) for the total scores. Further, it ranges between 0.74 and 0.81 for the six subscale scores 26,27 . It consists of 26 items, including five items for self-kindness, five items for self-judgment, and four items for common humanity, isolation, mindfulness, and over-identification. There were 13 reverse-scoring items out of the total items, and the total scores for self-compassion were calculated by adding up the individual subscale scores 27 . The scores for the negative subscales (i.e., self-judgment, isolation, and over-identification) were reverse-coded, and then the SCS total scores were calculated by averaging the six subscale means.
The Korean version of the Connor-Davidson Resilience Scale (CD-RISC), which contains 25 items, was used to evaluate resilience as the ability to cope with stress and adversity 28 . Connor and Davidson (2003) focused on the CD-RISC as a modifiable trait, and the CD-RISC contains items for evaluating resilience with a high internal consistency (Cronbach's α = 0.93).
Additionally, the Korean versions of the Beck Depression Inventory-II (BDI-II) and Beck Anxiety Inventory (BAI) were used to evaluate psychological distress, such as the severity of depression and anxiety. The BDI-II consists of 21 self-administered items, ranging from 0 to 63 29 . The BAI has 21 self-reported inventories for assessing clinical anxiety severity on a scale of 0 to 3 29  www.nature.com/scientificreports/ Image acquisition. A total of 79 participants underwent brain MRI on a GE Signa HDxt 3.0-Tesla scanner (GE Healthcare, Milwaukee, WI, USA) with an eight-channel phased-array head coil. Three dimensional T1-weighted fast spoiled gradient recalled echo sequence has the parameters that repetition time = 6.3 ms, echo time = 2.1 ms, flip angle = 12º, slice thickness = 1 mm, field of view = 256 × 256 mm 2 , matrix = 256 × 256, and isotropic voxel size = 1 × 1 × 1 mm 3 .

Statistical analyses.
For analyzing GMVs, voxel-based analysis was performed using FreeSurfer (version 7.1.0; http:// surfer. nmr. mgh. harva rd. edu) with standard images processed using the standard form in Free-Surfer. In addition, the Destrieux atlas was used as a reference to label and calculate the mean value of GMVs, which was implemented in FreeSurfer 30,31 .
Multiple regression models with age, sex, year of education, and total intracranial volume (ICV) as covariates were performed in a healthy adult database. The results were corrected for multiple comparisons using a Monte Carlo simulation with a p-value of < 0.05.
And, the objectives of our study were to evaluate the correlation of GMVs in the neural correlates of SF with psychological characteristics and to determine the implications of these regions. Therefore, exploratory Pearson's correlation analyses were conducted between the GMVs of the neural correlates of SF and the scores of the psychological measurements (SCS, CD-RISC, BDI-II, and BAI). For analyses, possible outlier values were examined using adjusted standardized sores.
Except for voxel-wise analysis, statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS) Windows software version 27. A statistical p-value less than 0.05 was considered significant.

Results
Sociodemographic and clinical characteristics. Table 1 shows the sociodemographic and clinical characteristics of a total of 79 participants. Furthermore, there were no differences in SFS scores according to sociodemographic variables, such as age, sex, education level, presence of religion, marital status, and monthly income. Additionally, statistical non-significance was maintained when adjusting for covariates (age, sex, and ICV).
Voxel-wise correlation in whole-brain analyses between the self-forgiveness and the gray matter volume in healthy individuals. The total SFS scores were significantly positively correlated with the GMVs in the left FG (peak voxel of MNI: X = − 29.4, Y = − 70.9, Z = − 13.3, Monte Carlo simulation correction, cluster-wise p < 0.05) in healthy individuals (Fig. 1). Additionally, statistical significance was maintained when adjusting for covariates, such as age, sex, and ICV.

Exploratory correlation analyses between the gray matter volumes of the fusiform gyrus associated with the self-forgiveness scale's total scores and psychological characteristics in healthy individuals.
Our study found one potential outlier, which was subsequently excluded from our analyses. Thereafter, the exploratory correlation analyses showed more significant p-values than before in the CD-RISC, BDI-II, and BAI scales. There are significant positive correlations between the GMVs in the left FG and the total scores of the SCS and CD-RISC in healthy individuals. However, the higher the GMVs in the left FG, the lower the total BDI-II and BAI scores in healthy individuals (Fig. 2). Figure 3 showed that the GMVs in the left cuneus cortex, including "acceptance and improvement" subscale-related brain regions, were significantly positively associated with healthy individuals' total CD-RISC scores. The GMVs of the right PPC showed a significant positive correlation with the total CD-RISC and SCS scores in healthy individuals (all p-value < 0.05). Further, after excluding one potential outlier, the levels of significance were stronger than those before in the CD-RISC scale-related correlation analyses (Fig. 3).

Discussion
Our study is the first to demonstrate that a high SF was significantly correlated with greater GMVs in the FG, which was positively correlated with self-compassion and resilience and negatively associated with psychological distress such as depression and anxiety among healthy individuals. Additionally, concerning subscale analyses, the changes in GMVs in the cuneus cortex, PPC, and FG related to SF may be associated with self-compassion and resilience among healthy individuals. This study showed that the higher the level of SF, the higher the GMVs in the FG among healthy individuals. According to the SF model, SF could relate to empathy and critical emotions, such as shame and guilt 32 . A previous study on multidimensional empathy revealed that individuals who could take others' perspectives were prone to have higher SF 33 . Another study showed that lower empathy could be associated with reduced activation of the FG in autism spectrum disorder 34 .  www.nature.com/scientificreports/ There are significant positive correlations between the level of guilt and facial recognition 35 . Among the vital neural regions related to facial emotional recognition of themselves or others, the FG is one of the core neural regions that is activated for the forgiveness of others or guilt states 17,18,21,[36][37][38][39] . Guilt is initially thought to be positively correlated with SF 40,41 . Therefore, our finding-that SF may be associated with FG-is consistent with previous findings.
Our research indicated that greater GMVs in the FG related to high SF levels were associated with higher self-compassion and resilience and less severe psychological distress such as depression or anxiety. Generally, both SF and self-compassion appear to have similarities in that they relate to emotional regulation, use adaptive means of self-regulation, and affect broad psychological functions 20 . Furthermore, previous studies show that SF associated with self-compassion support this finding 5,42 .
Regarding the relationship between SF and resilience, one study 6 showed that SF could increase resilience and reduce suicidal ideation in older adults. Further, a structural neuroimaging study indicated that increased cortical thickness in the FG is observed in highly resilient individuals 43 .
The SF has been presented to be negatively associated with the levels of depression and anxiety 44 . According to a systematic review, self-blame, which can be considered the opposite of SF, may predict depression, anxiety, and posttraumatic symptomology in bereaved parents. Our finding of this negative correlation could result from the potential role of low self-blame.
Our study showed that the higher the "acceptance and improvement" subscale scores, the higher the GMVs in the cuneus cortex and PPC, which are related to emotional attention and emotional regulation 45,46 . The subscale "negative affect, thought, and behavior" compromise inverse questions and is negatively correlated with the GMVs in the FG. That is, higher the negative affect, thought, and behavior can be related to lower GMVs of FG among healthy individuals. Taken together, it can be assumed that the GMVs of these brain regions are related to emotional regulation or recognition associated with SF in healthy individuals.
Our findings showed that healthy individuals with less transfer of responsibility had smaller GMVs in the precuneus and parahippocampal gyri. In other words, healthy individuals who blame themselves may have a smaller GMVs in the precuneus and parahippocampal gyrus. The precuneus is part of the DMN and is involved in self-referential processing [47][48][49] . The parahippocampal gyrus is involved in the retrieval of negative contextual memory 50 . Therefore, these findings suggest that individuals who blame themselves are more likely to have a smaller GMVs in the precuneus and parahippocampal gyrus.
Meanwhile, a mirror neuron system (MNS) is a group of neurons mirroring the behavior of others, and plays an essential role in imitation and empathy 51 . The three major components of the human classical MNS include the superior temporal sulcus (STS), inferior parietal lobule, and complex inferior frontal gyrus and premotor cortex [52][53][54] . Empathy is the tendency of observers to project themselves into their observations, a process referred to as an inner imitation 55 . Mirroring of emotional facial expression appears to be involved in not only the abovementioned classical MNS regions but also extended MNS regions, including the face processing networks of the fusiform gyrus 56,57 . Moreover, previous research has presented that SF and empathy are likely related to each other 58,59 , which suggests that SF may be related to MNS. Structural changes in MNS-related brain regions may be associated with the capacity to empathize, self-compassion, and low psychological distress among healthy individuals.
Our study had several limitations. First, we only investigated the association between brain structure and SF using a relatively small sample size (n = 79). Therefore, it may be necessary to examine the neurostructural correlates of SF in larger participant numbers. Second, the present study focused on the exploratory correlation between brain structures and psychological characteristics. Thus, further validation is required to establish causal relationships.
In conclusion, our study suggested that healthy individuals with high SF showed increased GMVs in the FG, presumably mediated by guilt or shame. Their correlations were associated with other psychological characteristics (high resilience, high self-compassion, high life satisfaction, and low psychological distress). Moreover, in the subscale analyses, the GMVs in some brain regions (cuneus cortex, PPC, precuneus cortex, parahippocampal gyrus, and FG) showed a significant positive correlation with self-compassion and psychological resilience. Further, they showed life satisfaction and a negative association with psychological distress such as depression and anxiety in healthy individuals. This result implies that SF is related to FG, which is responsible for facial recognition and emotional regulation and can be included in the mirror neuron system. These findings may provide a neuroscientific basis for understanding SF.

Data availability
All data used or analyzed during this study were included in this article. And further request can be directed with the detailed description to the corresponding authors.