The Five-Pattern Personality Inventory (FPPI), a Chinese localized personality scale, relates to topological properties of resting-state brain networks

Background Most studies assessing personality–brain associations have used Western personality questionnaires, which have been adapted for individuals in Western cultures. However, personality is inuenced by culture and customs, and Western personality questionnaires may not objectively reect the personality characteristics of individuals in Eastern cultures such as China. Hence, we adopted the graph analysis approach and the Chinese localized scale, the Five-Pattern Personality Inventory (FPPI), to explore the personality–brain associations of individuals in China. Methods The functional connectivity among 90 specic regions of the brain was mapped using functional magnetic resonance imaging (fMRI) data obtained from 109 healthy adult Chinese subjects. This allowed for the construction of brain networks and the analysis of personality–brain associations. Results In the present study, the Chinese personality–brain associations were traced to specic regions in the brain, including the frontal, temporal, and occipital cortices. Shao Yin scores and nodal metrics were associated with the brain regions of the executive control network and the salience network (SN), primarily in the left hemisphere. Characteristics of the Shao Yin and Shao Yang traits were consistent with the function of their corresponding correlated brain regions, which may supply evidence for the neurobiological basis of personality traits in Chinese culture. Conclusion In summary, these results provide the brain mechanism foundations of personality traits in the FPPI. Furthermore, these ndings provide a new perspective to help researchers understand personality–brain associations in the Chinese culture.


Background
The study of personality is at the core of clinical psychology. Personality refers to individual tendencies to express stable and speci c patterns of cognition, emotion, and behavior [1]. Researchers have long believed that personality is a neurobiological product [2], yet there have been few advances in our understanding of the neural mechanisms of human personality during the past decade. In recent years, a growing number of studies examining brain-personality mechanisms have employed the Western personality-encompassing Eysenck Personality Questionnaire (EPQ) and NEO Personality Inventory and Tridimensional Personality Questionnaire (TPQ) for guidance [3][4][5]. However, personality is shaped and formed by an intricate interplay of biological, environmental, and experiential (i.e., cultural) factors and in uences [6].
Whether the Western personality scales can objectively re ect the personality characteristics of people in East Asia, especially in China, is questionable. Many cross-cultural studies have revealed differences in common personality traits between Western and Eastern cultures, suggesting that a single questionnaire may not be feasible for assessing personality traits between two different populations [7,8]. Therefore, geographical region-speci c personality scales should be considered when conducting brain-personality studies across different cultures.

Page 4/19
The research titled "The yin and yang of neurotrophin action" by Lu has successfully introduced the Yin and Yang model of traditional Chinese culture to neuroscience [9]. Recently, the Five-Pattern Personality Inventory (FPPI) has become more popular for studying personality in East Asia. The FPPI originates from the Chinese philosophy of Yin and Yang and its ve principles of Tai Yang, Shao Yang, Ying Yang, Shao Yin, and Tai Yin. These principles are used to characterize various aspects of social behavior and emotional responsiveness [10]. Our previous study have revealed the mechanism differences of typical yin and typical yang trait [11]. In addition, the FPPI is referred to as the Eastern version of the EPQ [12].
And our group have also found the Chinese ve-pattern personality traits have a close relationship with Eysenck's personality traits and that both may be engaged in similar neurobiological mechanisms to some extent [13]. These two scales, which are classi ed as biological perspective scales, both assume that personality is a neurobiological product and concentrate on the relationships between personality traits and neurobiological mechanisms.
In the current study, we aimed to show that inter-individual differences in the personality traits of Chinese individuals assessed by FPPI could be associated with the topological characteristics of functional brain networks. To accomplish this goal, we utilized the graph theory to identify the intrinsic functional connectivity mechanisms associated with personality traits using the resting-state fMRI data of 109 healthy Chinese adult subjects.

Participants
All participants were recruited by bulletin advertisements in the Shanxi province of China. The exclusion criteria for this study included subjects having neurological or psychiatric diseases, cognitive disabilities, contraindications to imaging studies, history of substance abuse (i.e., illicit drugs or alcohol), or any other major medical illnesses. After applying the exclusion criteria, the remaining 109 subjects were included in this study, consisting of 24 males and 85 females with an average age of 24.8 ± 2.4 years (range: 21-33 years). The subjects provided written informed consent before the study and were reimbursed for their time. Approval for this study was obtained from the Shanxi Medical University Ethics Committee.

Personality assessment
The FPPI consists of 103 "yes" (score = 1) or "no" (score = 0) items. These questions assess personality in terms of ve different traits: Tai Yang, Shao Yang, Yin Yang, Shao Yin, and Tai Yin. In addition, "lie," which includes 8 items, is another personality dimension used to measure the truthfulness of subject responses.
If the score on this dimension was > 4, the questionnaire was regarded as invalid and excluded from the study. The FPPI has shown good reliability, construct validity, and convergent validity with other personality scales [14].

Data acquisition 2.4 Image pre-processing
The fMRI data were preprocessed with the SPM12 toolset (http://www. l.ion.ucl.ac.uk/spm). Due to signal instability and subject acclimation, the initial 10 volumes were dismissed. For the 202 images that remained, corrections were implemented to account for delays in slices and head movement. The translational parameters were greater than allowable (2.5 mm) in ve datasets. After excluding these ve datasets, the remaining images were normalized with respect to the traditional SPM8 echo-planar image template and resampled to 3 mm cubic voxels. Next, spatial smoothing was applied (4 mm full width at half maximum [FWHM] Gaussian kernel), and the linear trends were removed from the nal images. Lastly, we regressed the cerebrospinal uid (CSF), white matter, and six head movement parameters.

Computation of functional connectivity networks
Anatomical parcellation was accomplished with automated anatomical labeling (AAL) by separating the images into 45 regions of interest (ROI) for each hemisphere of the brain, which resulted in 90 total ROI.
Edges of the network were de ned by generating 90 × 90 correlation matrixes and calculating the regionwise Pearson's correlation coe cients. The threshold was set as the sparsity (S), or the ratio of actual existing edges divided by the maximum possible edges in the network. The construction of the brain networks is shown in Fig. 1

Graph theoretical analysis: Network metrics
The topological features of the brain functional networks were assessed by the nodal and global network measures. First, the global metrics, consisting of the normalized clustering coe cient (γ) and characteristic path length (λ), were computed. In comparison to the random networks with low clustering coe cients and shorter path lengths, small-world networks display higher clustering coe cients and similar path lengths (i.e., γ = Cp/Crand > 1, λ = Lp/Lrand ≈ 1) [15]. Small-worldness is the quantitative measurement that includes the combination of these two conditions (i.e., σ = γ/λ > 1) [16]. Meanwhile, the nodal measures encompassed Freeman's betweenness centrality (BCi) and degree (Ki). More details on the mathematical de nitions and interpretations of the network metrics may be found in the Supplementary Material.

Associations between the network metrics and personality traits
All of the nodal measures reached the threshold in the range of 0.05 ≤ T ≤ 0.40 (interval = 0.01). The area under the curve (AUC) was calculated for each network metric in order to obtain a summarized scalar for topological characterization that was independent of the single-threshold selection [17]. A partial correlation was determined between each of the ve trait scores and the AUC of each network metric, with age, gender, and education (years) as covariates. The false discovery rate (FDR) was applied to the data for correction [18] using the code developed by Groppe in 2010 with Matlab (MathWorks, Natick, MA, USA).

Statistical analysis
P-values < 0.05 were considered statistically signi cant.

Descriptive statistics of the personality traits
The demographic and personality data of the participants are shown in Table 1 (n = 109; males = 24). The corresponding correlation analysis is shown in Table 2. There were both negative and positive correlations found among the ve personality traits. The next goal was to determine the effects that were uniquely driven by each personality trait. Thus, the four remaining personality traits were included as covariates for computing the partial correlation between the AUC of the network metrics and each of the personality traits.

Associations between the personality traits and nodal metrics
Those personality traits and nodal metrics that showed strong positive correlations are shown in Table 3.
We found that Tai

Discussion
We have shown that personality traits and brain topological properties are associated with both global and regional metrics. Our ndings showed obvious small-world properties of the functional brain networks, which is in accordance with several previous studies [19]. Meanwhile, the personality-brain associations in the Chinese participants were found in speci c regions of the frontal, temporal, and occipital cortex. Compared with previous Western personality-brain correlation studies, we found that the personality traits of Chinese and Western individuals show both similarities and differences in brain mechanisms. Furthermore, these ndings provide a new perspective for researchers to better understand the personality-brain associations in Chinese culture.

Functional brain correlates of the Shao Yin trait
The Shao Yin is a trait re ecting restrained behavior, toughness, and alertness to the surroundings. While topological correlates were observed for Tai Yang and Shao Yang, the Shao Yin trait showed the strongest connection with brain function, including positive correlations with the normalized clustering coe cient on the global metrics and regional function of brain areas in the salience network (SN) and executive control network (one of the sub-networks of attention). These two networks are involved in the processing and integration of a variety of external cognitive and emotional information [20,21]. The executive control network, which encompasses the anterior cingulate gyrus, supplementary motor area, and caudate, is thought to be associated with con ict resolution abilities and the inhibition of automatic responses [22]. The SN, which comprises the anterior cingulate cortex and insula, is involved in the allocation of attention and segregation of internal and external cognitive data [23]. Meanwhile, the SN also functions as the biological switch between the task-related and self-monitoring networks [24].
Herein, the function of these two networks was consistent with the Shao Yin trait feature of restrained behavior and alertness to one's surroundings, which suggests a rapid shift from awareness to consciousness. This indicates that the Shao Yin trait, which is the central characteristic of Chinese personality (Li et al. 2014), is related to most functional variables of the brain. In addition, several studies have shown positive associations between neuroticism in the Eysenck Personality Questionnaire (EPQ) and three brain regions (ACG, INS, and CAU) via resting-state fMRI [25][26][27], task fMRI [28,29], and positron emission tomography (PET) [30]. Therefore, we believe that Shao Yin and neuroticism have similar brain mechanisms.
The left hemisphere of the brain showed the strongest correlations in our study, indicating the lateralization of regions within the left hemisphere with the Shao Yin trait. While the left hemisphere is primarily involved in social approach, the right hemisphere is preferentially associated with social withdrawal in personality [31]. A number of studies have revealed interventions that increase intrinsic alertness, resulting in leftward-moving spatial biases [32,33]. Thus, the alertness of Shao Yin could be the intrinsic motivation of "approach." It would be over-simplistic to associate every aspect of the Shao Yin trait with the left hemisphere, so further studies are needed to explore the restrained characteristics of Shao Yin and lateralization.

Functional brain correlates of the Tai Yang and Shao Yang trait
Tai Yang, which is a trait re ecting impulsive, aggressive, and adventurous behaviors, is positively correlated with nodal metrics in the left cuneus (CUN.L). CUN is believed to play an essential role in cognitive processing [34] and risk-taking actions. For instance, previous fMRI studies have shown that risk-taking actions elicit signi cantly stronger activation of the CUN [35,36]. Moreover, CUN activation was found to be higher in aggressive veterans compared with the general population, which suggests a close connection between CUN and aggression [37]. We infer that CUN appears to subserve aggressive and risk-taking actions, aspects commonly associated with Tai Yang. In addition, the link between CUN and extroversion or extroversion-related traits has been well established [38,39]. The Shao yang trait, which re ects the tendency toward more movement of the body and physical activity, is positively associated with nodal metrics in the bilateral paracentral lobule (PCL). PCL is regarded as the continuation of the precentral and postcentral gyrus and is known to regulate the motor and somatosensory representation of the foot and leg [40,41]. PCL may subserve the movement of the body, a trait commonly related to Shao Yang.

Limitations
This study has two limitations that should be addressed in future work. First, the neurobiological mechanisms of only three of the ve personality traits were discovered. Unfortunately, we were unable to elucidate the mechanisms of the Yin Yang and Tai Yin personality traits. This could have been due to our small sample size, and future studies using larger populations may provide more insight into this topic. Second, the molecular mechanism involved in the connection between Chinese personality traits and structural brain networks remains unknown, yet future studies using diffusion tensor imaging (DTI) may provide new insights.

Conclusions
In summary, the present study uncovered strong associations between Chinese personality traits and brain topological properties based on the commonly utilized FPPI. This was accomplished by applying graph theoretical analysis to the data from the resting-state fMRI. These results offer a new perspective for our understanding of personality-brain associations in the Chinese culture. The subjects provided written informed consent before the study and were reimbursed for their time.

Abbreviations
Approval for this study was obtained from the Shanxi Medical University Ethics Committee.

Consent for publication
Not applicable.

Availability of data and materials
The datasets analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.