A Serbian version of the ANPS and its link to the five-factor model of personality

Abstract This short communication presents a Serbian version of the Affective Neuroscience Personality Scales (ANPS). The ANPS is a self-report measure assessing individual differences in primary emotional systems as derived from Jaak Panksepp’s Affective Neuroscience Theory. As a recent work by Montag & Panksepp (2017a) confirmed the original demonstration of strong associations between primary emotions and the Five-Factor Model of Personality (Davis et al., 2003) across different cultures (USA, Germany, China), we replicated these findings in a Serbian sample. Moreover, following the idea of a recent commentary of Di Domencio & Ryan (2017) on Montag & Panksepp’s (2017a), we present for the first time detailed associations between Five-Factor Model facets as assessed with the NEO-PI-R and primary emotions.


Introduction
The Affective Neuroscience Theory (ANT) by Jaak Panksepp represents one of the most important emotion theories to understand affective lives of mammalian beings Panksepp, 1998;Panksepp, 2011). By electrical stimulation of the mammalian brain, but also pharmacological challenge tests, Panksepp carved out seven primary emotional systems deeply rooted in subcortical areas of the mammalian brain (Panksepp, 2007). These systems have been coined SEEKING, CARE, LUST, and PLAY on the positive side of emotions and FEAR, RAGE, SADNESS on the negative side of emotions (for an overview see also )1. A better understanding of primary emotional systems on both molecular level, structural/functional brain level as well as on a psychological level has been shown to be of tremendous importance to also better understand affective disorders (e.g., Montag et al., 2017;Panksepp, 2010;Panksepp & Yovell, 2014).
Mounting evidence demonstrates the usefulness of ANT to also better understand individual differences in personality (e.g., Davis  2018). In a recent publication by Montag & Panksepp (2017a) it has been demonstrated with data from cross-cultural research that not only primary emotional systems as described above are robustly linked with the prominent Five Factors of Personality, but also that primary emotional systems could represent the phylogenetically oldest parts of human personality (for a detailed discussion see by Montag & Panksepp, 2017a and the follow-up work by Montag & Panksepp, 2017b). Together with an older work by Davis and his colleagues (2003), evidence from the USA, Germany and China point towards the idea that (high) SEEKING could reflect the emotional foundation of Openness to Experience, (high) PLAY for Extraversion, low ANGER/high CARE for Agreeableness and (high) FEAR, SADNESS and ANGER for Neuroticism (see also Barrett et al., 2013;. In all these works individual differences in primary emotional systems have been assessed with the Affective Neuroscience Personality Scales (ANPS), a wellknown tertiary measure of a person's endophenotype based on Panksepp's ANT (Davis & Panksepp, 2011). A commentary by Di Domenico & Ryan (2017) on the work by Montag & Panksepp (2017a) correctly pointed out the necessity to also understand how individual differences in primary emotion systems relate to facets of the Five Factor Model (FFM). Therefore, in the present work, we provide associations between individual differences in primary emotional systems as assessed with the ANPS 2.4 and the FFM as assessed with the NEO-PI-R. As the present study has been conducted in Serbia, we are able to investigate if a) the mentioned robust associations between primary emotional systems and the FFM can be extended to a Serbian sample and b) how individual differences in primary emotional systems are linked to individual differences in NEO-PI-R facets. Finally, and c), we present the Serbian version of the ANPS in this paper. The growing need to have a Serbian translation of the ANPS is also reflected in the many already existent translations and validations of the ANPS and its use around the globe. The scale is already available in the following languages other than English . We also know of several additional translations that have not been published including a Japanese version to be published soon (for a recent perspective on cross-cultural affective neuroscience see Özkarar-Gradwohl, 2019).

Declarations
We report how we determined our sample size, all data exclusions, all manipulations, and all measures in the study. The present work has been approved by the ethics committee of the Serbian Psychological Association at the Faculty of Philosophy, University of Belgrade. All participants signed an informed consent and all procedures adhered to the principles of the Declaration of Helsinki.

Participants
A sample of 340 participants (57 males and 283 females, age = 20.94, SD = 2.76; age-range: 19-46) filled in ANPS 2.4 and NEO-PI-R. Please note that as part of a cross-cultural project, also HEXACO data were collected, which have been published in Knežević et al. (2019), recently. We did not aim at presenting ANPS, HEXACO and NEO-PI-R data in one work, because Knežević et al. (2019) investigated HEXACO and ANPS in a German and Serbian sample as a separate project, whereas in Germany no NEO-PI-R data was collected. All participants had a university/student background. Participants stated to be healthy upon participation. Students volunteered to participate in the study and signed informed consent forms before completing the instruments. Students received feedback regarding their assessment results, but no extra course credit was provided. Instruments were administered via an online Moodle platform, not allowing participants to skip the answers. There were no missing data, and we did not apply any additional manipulations to the data. The participants filled-in personality inventories during regular lessons. All data, materials, syntax of questionnaires, and other supplementary information are available at https://osf.io/gp2z8/. The sample size was calculated having in mind the goal to detect correlations which size is at least 0.15, with the power of 0.80 at p = 0.05 level (the calculation was done by using G*Power free software).

Questionnaires
The ANPS 2.4 by Davis & Panksepp (2011) consists of 112 items and has been translated from English into the Serbian language (and back-translated) by bilingual psychologists. The complete Serbian ANPS questionnaire is provided in Appendix 1 and is available at https://osf.io/gp2z8/. With each ANPS scale having 14 items and a Likert scale ranging from strongly disagree (0) to strongly agree (3) individual differences in SEEKING, CARE, PLAY (positive affects) and FEAR, ANGER, SADNESS (negative affects) can be assessed. The seventh primary emotion of LUST is not assessed in the ANPS questionnaire, because of possible negative carry-over-effects on answers with respect to all other items of the ANPS. Twelve additional items assess individual differences in spirituality. Spirituality represents not a primary emotion but is of potential interest in the treatment of patients suffering from addictive disorders. Spirituality will be not further investigated in the present work.
The NEO-PI-R consists of 240 items with a five Likert scale ranging from strongly disagree (0) to strongly agree (4). The NEO-PI-R assesses the FFM and uses the following factor labels: Neuroticism, Extraversion, Openness to Experience, Agreeableness, and Conscientiousness. Moreover, it segments each factor into six facets for a total of 30 facet scores. These facets are presented in the results section. The Serbian version of the NEO-PI-R was used in this study (Đuričić-Jočić, Džamonja Ignjatović, & Knežević, 2004).

Statistical Analysis
Several correlational analyses using Pearson's coefficient were computed searching for associations between the dimensions of the ANPS and the NEO-PI-R. The influence of gender on the personality dimensions will not be investigated, because the recruited sample largely consists of females. Correlations with age and the dimensions will not be further considered in the results section, because no significant associations with age and the NEO-PI-R or ANPS were observed (after correction for multiple testing). All correlation probabilities in the paper are two-tailed. Finally, a factor analysis (Principal Axis Factoring) yielded two factors out of the ANPS dimensions, namely positive and negative affect.

Descriptive statistics of the present sample
Descriptive statistics are presented in Table 1. The skewness values presented in this table suggest that some variables might slightly deviate from a normal distribution: Openness, Extraversion, Agreeableness, CARE and SEEKING have the largest negative asymmetry (grouping of the scores around the higher values), while ANGER and SADNESS have the largest positive asymmetry (grouping around the lower values). Nevertheless, visual inspection of the variables shows a rather normal distribution of variables supporting the validness of parametric testing.
Internal consistencies of ANPS scales obtained in this study are very good, and in case of NEO PI-R domain scales, excellent (again see Table 1). Cronbach's alphas for Serbian NEO PI-R facets range from .67 (Activity and Excitement Seeking) to .89 (Anxiety), except for (Openness to) Values and Tender-Mindedness, which had reliability coefficients below .60. The median Cronbach's alpha of the Serbian NEO facets is .77.

Associations between ANPS 2.4 and the NEO-PI-R
To replicate earlier associations (see Montag & Panksepp, 2017a) Table 2. The main discrepancy in Table 2

The associations between ANPS and the NEO-PI-R facets by personality dimension
When it comes to the Neuroticism facets, FEAR and SADNESS were both most strongly linked to Anxiety (N1), which likely further reflects the close association between the ANPS FEAR and SADNESS scales as shown in Table 7. Not surprisingly, ANGER was most strongly linked to Angry Hostility (N2), which likely is a reflection that six of the N2 items deal closely with anger and frustration with the remaining two items dealing with disgust and resentment. Correlations between FEAR, SADNESS, and ANGER with Neuroticism facets are displayed in Table 3. Note: All correlations significant at p < .001.
The two highest associations between PLAY and facets of Extraversion can be observed with Excitementseeking (E5) and Gregariousness (E2), while two highest associations between SEEKING and facets of Openness to Experience can be observed with Ideas (O5) and Actions (O4; see Table 4 for all associations). As already mentioned with respect to Agreeableness both high CARE and low ANGER were demonstrated to be of relevance. On Agreeableness facet level the strongest association could be observed between CARE and Tender-Mindedness (A6). Interestingly, this facet had several items similar to ANPS CARE items such as dealing with helping the poor and elderly, sympathy for panhandlers, human needs, and sympathy for the less fortunate. The other strong correlation for the ANPS CARE scale was with Altruism (A3), which again had items similar to those on the CARE scale such as helping others, being charitable, and being known for generosity (see Table 5 for all facet associations). With respect to ANGER, the Agreeableness facets of Compliance (A4) was most robustly associated (here inverse); which is not surprising given the number of items dealing with anger such as being sarcastic and cutting, expressing anger, fighting back, and getting into arguments (see Table 5 for all associations). The full correlation matrix is available in Online Supplementary Materials (https://osf.io/6rjhm/).

Results of the Principal Axis Factoring (PAF) using Promax rotation (ANPS 2.4 data)
The Factor Analysis (Principal Axis Factoring) yielded two factors, namely the factor positive affect and the factor negative affect. These factors were visible using a Kaiser-Gutman-criterion with factor negative affect explaining 35.69% of the variance (the first eigenvalue 2.14) and factor positive affect explaining 26.97% of the variance (the second eigenvalue 1.62). The rotated factor matrix is depicted in Table 6. Correlation between the factors was -.20.
For reasons of comparability with older works by Davis et al. (2003) and Davis & Panksepp (2011), we also report intercorrelations between primary emotions as depicted in Table 7. In addition to presenting associations between ANPS dimensions and the facets of the NEO-PI-R, the present work also introduces a Serbian version of the ANPS to the literature. Given the external validation of the ANPS with the data from the NEO-PI-R together with good psychometric properties of the Serbian ANPS (internal consistencies) and a replication of the factorial structure as presented in Davis et al. (2003) and Davis & Panksepp (2011), we hope that the Serbian ANPS will be a valuable addition to the toolbox of psychologists in the Serbian/Croatian research community.
Finally, we address some limitations. The sample size of the present work is rather small and unfortunately only a very limited number of male participants could be recruited for our research endeavor. This is noteworthy as gender effects have been observed on some of the ANPS scales, perhaps with the most robust linking higher scores in the CARE dimension to females compared to males. Such associations could not be investigated in the present work. Moreover, future studies clearly need to replicate the present associations with the ANPS and NEO-PI-R facets to get insights on the robustness of the findings. Aside from this, we stress that research endeavors testing principles of Affective Neuroscience Theory in the realm of Personality Neuroscience will strongly rely on the implementation of brain science methods going beyond self-report, something which we could not achieve with the present work. Finally, future works might also want to rely stronger on other more elaborate statistical methods such as structural equation modeling.
Contributions: CM and GK designed the present study. GK and LBL collected data in Serbia. GK and LBL translated the ANPS questionnaire. CM drafted the first version of the manuscript. CM and GK conducted the statistical analysis. KLD critically revised the present manuscript and improved the language. All authors approved the final version of this work.

Ethics statement:
The present work has been approved by the ethics committee of the Serbian Psychological Association at the Faculty of Philosophy, University of Belgrade. All participants signed an informed consent and all procedures adhered to the principles of the Declaration of Helsinki.