An adaptive modelling approach to employee burnout in the context of the big five personality traits

Job burnout has been on the rise in the past decade, especially amongst the younger working generation. While work environmental aspects play an important role in predicting burnout, variations in personality traits are integral for understanding the syndrome ’ s risk factors, processes, and outcomes. This paper studies the complex interaction of personality factors on the one hand and work environment aspects on the other through the relatively novel adaptive causal network modelling paradigm. Due to the adaptive nature of the model, it can investigate the effects of changes in particular job demands and resources on the symptoms of burnout and their dependence on different personality traits. The model can also demonstrate how an individual ’ s personality traits, environmental perception, and burnout symptoms can adaptively be altered by individual therapy, in this case, mindfulness-based cognitive therapy. Using the dedicated software environment in MATLAB to simulate the designed adaptive causal network model, two main scenarios were explored, focusing on the neuroticism personality trait. The results demonstrate that neuroticism increases due to interpersonal conflict, indicating that neuroticism can be treated as an adaptive trait. Furthermore, when mindfulness-based cognitive therapy was introduced into the simulation, the likelihood of developing burnout decreased because the perception of the work environment was positively changed due to the therapy. This model contributes to the field of burnout modelling by repre- senting personality traits as adaptive factors that can be changed through individual interventions. More detailed research is needed to understand how organisational-level interventions can also impact burnout development through changes in environmental perception and personality.


Introduction
In today's work-centred culture, burnout has become a well-known phenomenon. Getting sick due to work seems to be becoming more and more common, with most people knowing an acquaintance or a loved one who suffers from the syndrome. In May 2019, the World Health Organisation (WHO) declared burnout a medical syndrome and occupational phenomenon. Recent figures from Statistics Netherlands show, for example, that work-related psychological fatigue is increasing, especially among people in their twenties and thirties (Atroszko, Demetrovics, & Griffiths, 2020;Centraal Bureau Voor de Statistiek, 2022). The current paper introduces a relatively new research paradigm to study the development of burnout, namely adaptive network modelling. Building upon empirical evidence for the Job-Demands Resources model and the Big Five Personality traits in relation to burnout, the adaptive model in this paper treats burnout as a dynamic and adaptive interplay between personality traits and the perception of job demands and resources.
Due to the adaptive nature of the model, the model can investigate the impact of changes in particular job demands and resources on burnout symptoms, considering the influence of different personality traits. Additionally, the model can simulate the effect of individual therapy on the perception of job demands and resources, taking into account specific personality traits. The adaptivity of the model allows a realistic representation of the complex dynamics of many internal and external factors that contribute to the development of burnout and helps to give insight into the potential effect of burnout-related therapy. For example, it can be used to explore how the adaptivity of one's perception of job demands and resources affects the possibility of developing burnout in the context of the Big Five personality traits.
In Section 2, relevant background literature is discussed, and in Section 3 the modelling approach used is briefly summarised. Section 4 introduces the adaptive network model, and in Section 5 the results of simulation experiments are described. Finally, Section 6 is a discussion.

Burnout
Burnout is a "chronic, work-related mental condition in people, characterised by mental exhaustion that is accompanied by stress complaints, mental distance, and a sense of reduced competence" (Schaufeli & Enzmann, 1998, p. 36). In other words, people suffering from burnout lack the energy to face another "demanding" workday (Lloyd, King, & Chenoweth, 2002;Maslach, Schaufeli, & Leiter, 2001).
Burnout is usually defined according to three dimensions: emotional exhaustion, depersonalisation, and reduced personal accomplishment (Maslach & Jackson, 1981). Emotional exhaustion is the principal symptom of burnout and refers to a feeling of depletion of one's emotional and physical resources, including loss of energy, debilitation, and fatigue. Depersonalisation (also referred to as cynicism) is defined as the feeling of disconnectedness to one's job and having negative, cynical attitudes toward one's job. Depersonalisation can manifest itself in various ways, such as reduced involvement, low empathy, and negative or inappropriate attitudes towards colleagues (Roloff, Kirstges, Grund, & Klusmann, 2022). Reduced personal accomplishment, or inefficacy, encompasses feelings of incompetence, inability to cope, and low confidence in one's abilities in terms of one's work (Maslach et al., 2001). Our adaptive causal network model will incorporate these three dimensions as core elements.
Burnout can have a detrimental impact on people's physical and psychological health, which is often accompanied by reduced productivity, a decline in quality of work, and an increased likelihood of absenteeism and sickness leave (Maslach & Leiter, 2016;Salvagioni et al., 2017;Ogunfowora, Nguyen, Steel, & Hwang, 2022). This makes burnout not only an individual but also an organisational level issue.
Absenteeism and turnover due to burnout cause additional costs for organisations as employers must continue to pay employees their salary while they are not productive or absent. The association between burnout and financial loss has been investigated in some studies. For example, research by Dutch health insurer Zilveren Kruis indicates that a burnt-out employee costs employers on average 60,000 euros (Wester, 2017), and the U.S. Census Bureau and Bureau of Labor Statistics estimate the costs of absenteeism by $40 billion annually (Employer Assistance Resource Network on Disability Inclusion, 2019). The above individual and organisational consequences of burnout emphasise the need for early identification of this health condition in the work environment and preventive interventions. Tools such as Cognitive Behavioural Therapy (CBT) or more preventive methods such as stressreducing activities like mindfulness are commonly mentioned methods that may help to manage burnout and increase employees' productivity (Patel, Sekhri, Bhimanadham, Imran & Hossain, 2019). More specifically, Cognitive Behavioural Therapy might be effective for reducing emotional exhaustion, and stress-reducing activities such as mindfulness can positively influence all three dimensions of burnout and can significantly decrease stress (Korczak, Wastian, & Schneider, 2012;dos Santos et al., 2016).

Personality and burnout
Although burnout is a stress-related illness, its experience varies on an individual level. Individuals have varying stress tolerances, burnout thresholds, and symptoms. Research on burnout and personality plays an integral role in understanding the risk factors, processes, and outcomes of burnout. People have varying perceptions and reactions to stressful situations depending on their personality characteristics, which in turn affect factors like stress experience and threat perception (Maslach et al., 2001;Suls & Martin, 2005). This suggests that burnout is influenced not only by external factors in the workplace, such as job demands, but also by individual factors.
Much of the recent research on burnout and personality has been based on the Five Factor Personality Model, a model which looks at personality in terms of five broad dimensions, also known as the Big Five personality traits (Costa and McCrae, 1992): Within the Big-5 personality traits, emotional stability is defined as the ability to cope effectively with negative emotions; people with high emotional stability can better tolerate daily stressful situations. On the other hand, people with emotional instability, also referred to as neuroticism, have a tendency to easily experience negative emotions. Neuroticism is often considered the personality trait most closely related to burnout (Deary, Agius, & Sadler, 1996); Zellars, Perrewe, & Hochwarter, 2000). Individuals who are highly neurotic have cognitive traits such as low self-esteem and fearfulness and tend to have a higher risk of burnout across all three dimensions (McCrae & Costa, 1987). Individuals with high neuroticism have been found to be more likely to report feelings of emotional exhaustion and depersonalisation (Bühler & Land, 2003). Neuroticism has also been found to have a negative impact on personal accomplishment. For example, Bianchi (2018) conducted a relative weight analysis and found that neuroticism could account for 53.46 % of the variance in burnout, compared to only 31 % of variance explained by the work-contextualised factor of Effort-Reward imbalance. While the other four Big-5 personality traits will also be explored in this paper, neuroticism will be the primary personality trait examined in the designed adaptive causal network model simulations.
The personality trait extraversion is characterised by the ability to approach challenges in a positive manner as well as the ability to form social connections. A vast body of literature has examined the relationship between extraversion and burnout (González-Romá, Schaufeli, Bakker, & Lloret, 2006). The literature on the influence of extraversion on the development of burnout remains mixed, however, several studies have found a negative relationship between extraversion and depersonalisation (Francis, Louden, & Rutledge, 2004;Zellars et al., 2000). Additionally, a positive association was found between extraversion and personal accomplishment (Francis et al., 2004). Conscientiousness, characterised by orderliness, responsibility, and sometimes efficient problem solving, has also received mixed findings concerning its effect on burnout. Bakker et al. (2006) and Deary, Watson and Hogston (2003) found a positive relationship between conscientiousness and personal accomplishment, and LePine, LePine and Jackson (2004) found a negative relationship between conscientiousness and exhaustion.
The agreeableness personality trait has properties such as cooperativeness and trust; people high on agreeableness tend to appear cooperative and generous. Agreeableness has been found to relate negatively to emotional exhaustion and positively to personal accomplishment (Piedmont, 1993). It has also been outlined that agreeableness can reduce depersonalisation tendency (Zellars et al., 2000). The final personality trait, openness, is characterised by curiosity and originality and has been found to have the weakest relationship to burnout . Nonetheless, a positive relationship was found between openness and personal accomplishment and a negative relationship between the trait and depersonalisation (Zellars et al., 2000).

The job demands -resources model
The Job Demands-Resources (JD-R) Model consists of two main A. Bashkirova et al. processes related to a person's wellbeing (Demerouti, Bakker, Nachreiner, & Schaufeli, 2001). First, job demands are defined as "physical, social, or organisational aspects of the job that require sustained physical or mental effort and are therefore associated with certain physiological and psychological costs" (Demerouti et al., 2001, p. 501). The JD-R model proposes that high job demands lead to strain and health impairment (i.e., the health impairment process; Schaufeli & Taris, 2014). The second process within the JD-R model is defined by job resources, which are "physical, social, or organisational aspects of the job that may do any of the following: (a) be functional in achieving work goals (b) reduce job demands and the associated physiological and psychological costs (c) stimulate personal growth and development" (Demerouti et al., 2001, p. 501) The JD-R model proposes that high job resources lead to increased motivation and higher productivity (i.e., the motivational process; Schaufeli & Taris, 2014). The theory of the JD-R model implies that psychological strain or burnout among employees results from an imbalance between demands and resources experienced by the employee. Specifically, high job demands and low resources could lead to employee burnout. In contrast, high resources can act as protective factors for the job demands people face (Van den Broeck, De Cuyper, De Witte, & Vansteenkiste, 2010; Schaufeli & Bakker, 2004).
The job demands and resources that seem to be most closely related to the three burnout dimensions were selected for the adaptive network model introduced in Section 4. As such, based on ample literature concerning burnout and the JD-R model, the job demands workload, interpersonal conflict, and job insecurity were selected to include in the model, as well as the job resources social support and job autonomy (Jacobs & Dodd, 2003;Charoensukmongkol, Moqbel, & Gutierrez-Wirsching, 2016;Fujiwara et al., 2003;De Witte et al., 2010;Bakker, Demerouti, & Euwema, 2005).

Job demands, resources, work environment, and personality
Most recent literature on Big-5 personality traits suggests that personality traits are not static throughout the lifespan. Rather, personality traits can be malleable, with change continuing into adulthood (Wu, 2016;Roberts, Walton, & Viechtbauer, 2006). Since work generally requires the most time and commitment of adults, it is vital to investigate the impact of work environment on personality.
Firstly, it is important to examine how the work environment, including job demands and resources, may influence one's personality. A study by Hudson, Roberts and Lodi-Smith (2012) found, for example, that investment and commitment to a job were connected to extraversion, agreeableness, conscientiousness, and emotional stability. Positive longitudinal change and investment positively affected these personality traits, especially conscientiousness (Hudson et al., 2012). Work can have a powerful influence on personality as it continually shapes one's values and activities, which, in turn, influence individual behaviour and cognition, thus solidifying personality change via a bottom-up process (Li, Fay, Frese, Harms, & Gao, 2014;Wu, 2016).
Extending the concept of bottom-up personality change, Wu (2016) built on the job demand-control model to further explain how the work environment can change Big-5 personality traits (Karasek, 1979). He found that changes in job demands and control can shape stress perceptions and experience in the workplace, which can, in turn, affect personality. Job demands, specifically time demands, led to increased job stress, making individuals more neurotic and less extroverted. In this study, job control was also found to predict changes in the Big-five but did not predict changes in neuroticism and extraversion. Wu (2016) found results consistent with Gray's biopsychological theory of personality (1981; 1990), where job stress was critical in driving changes in extraversion and neuroticism.
A reverse relationship was also found where personality changes have been found to be predictive of the changes in one's perception of the work environment and experience (Langelaan, Bakker, van Doornen, & Schaufeli, 2006;Sutin & Costa, 2010). For instance, Sutin and Costa (2010) found that disagreeable individuals have often found themselves in jobs with higher physical demands and more hazardous working conditions. Emotionally stable and conscientious employees were found to hold jobs with higher autonomy. A study by Langelaan et al. (2006) found that neurotic employees have a generally more pessimistic view of their work environment, predisposing them to burnout. Furthermore, neuroticism has been found to be particularly linked to the job demand for interpersonal conflict (Zahlquist et al., 2022).
Following the person-environment fit model, Zahlquist and colleagues (2016) found that daily interpersonal conflicts strongly impact individuals with high neuroticism. Furthermore, a study conducted in Hong Kong found that neurotic individuals were predisposed to having higher levels of stress (Stoeva, Chiu, & Greenhaus, 2002). In contrast, in a sample of Zimbabwe teachers, Tshababa et al. (2021) found that high extraversion was a potentially protective factor against stress when there were high demands and low resources in the workplace, hence decreasing the risk chances of burnout development. The study confirms previous findings that extraverts have more positive situational perceptions and experience lower stress levels (Diener & Lucas, 1999). Conscientiousness as a personality trait has been found to reduce workplace conflict due to effective time management (Wayne, Musisca, & Fleeson, 2004).
The current adaptive model builds on the connection between the Big-5 personality traits and job demands and resources and considers the strength of those relationships and their subsequent influences on employee burnout.

The adaptive network modelling approach
An adaptive network model was designed to illustrate how the Big Five personality traits combined with job demands and resources influence the development of burnout, and to exemplify how burnout symptoms can be altered by therapy. This kind of network provides a type of dynamic and adaptive modelling that simulates the adaptive causal relations and pathways between burnout, personality traits and therapy in relation to the individual perception of the external and internal factors that contribute to the development of burnout.
To design the network, an adaptive causal network modelling approach has been adopted (Treur, 2016(Treur, , 2020a. This approach models networks of causal relations and pathways in terms of network characteristics, such as connections between nodes. These network characteristics are often adaptive, because they can change over time. Central to these so-called adaptive temporal-causal networks are the dynamics of the nodes (also called states) in the network (dynamics within the network) versus dynamics of the characteristics of the network (dynamics of the network). Both types of processes are specified mathematically by declarative mathematical functions and relations. The states Y in such a network have activation levels Y(t) over time t. The network characteristics for this modelling approach are described as follows: Connectivity characteristics: a connection weight ω X,Y represents the connection strength from state X to state Y.
Aggregation characteristics: for each state Y a combination function c Y is used to aggregate multiple incoming impacts on a state.
Timing characteristics: Y represents the speed factor η Y of state Y which determines the timing of the impact on Y.
The difference (or differential) equations that are used for simulation purposes and also for analysis of network dynamics incorporate these network characteristics ω X,Y , c Y, η Y : where X 1 , …, X k are the states from which state Y gets incoming base connections. An example of an often-used combination function is the advanced logistic sum function, defined by: (1 This function is obtained from the classical simple logistic sum function by subtracting its value for sum 0 from it and rescaling the result for the [0, 1] interval. Here, the simple logistic sum function is: The adaptivity of the network is modelled according to the principle of self-modelling or reification, which means that for each adaptive network characteristic, a state is added to the network (called self-model state or reification state) which represents this (adaptive) network characteristic (Treur, 2020a). Such nodes are depicted at the next level (selfmodel level), where the original network is at a base level. The types of characteristics with their self-model states and roles are shown in Table 1.
The concepts presented above enable the design of the adaptive network model and their dynamics in a declarative manner based on mathematically defined functions and relations.
In the neuroscientific literature, a distinction is made between synaptic and nonsynaptic (intrinsic) plasticity or adaptation. The classical notion of synaptic plasticity has been extensively described in the literature since long ago, for example, for stimulus-response conditioning or Hebbian learning, e.g. (Hebb, 1949;Shatz, 1992). This type of adaptation addresses how the strength of a connection between different states is adapted over time due to the simultaneous activation of the connected states. By contrast, the non-synaptic adaptation of intrinsic excitability of (neural) states has been addressed in more detail more recently, e.g. (Chandra & Barkai, 2018;Debanne, Inglebert, & Russier, 2019). The latter form of adaptation has been related, for example, to homeostatic regulation (O'Leary et al., 2013) and to how deviant dopamine levels during sleep make that in dreams, more associations are used due to more easily excitable neurons, e.g., (Boot et al., 2017).
Moreover, both (synaptic and nonsynaptic) forms of adaptation can easily work together, e.g. (Lisman et al., 2018). Using the self-modelling principle, both types of plasticity can be modelled easily. Synaptic plasticity can be modelled by introducing self-model states of the form W X, Y to represent the adaptive weight of the connection from state X to Y; for shortness, such states are also called W-states. Non-synaptic plasticity for excitability thresholds can be modelled by introducing self-model states of the form T Y to represent the adaptive excitability threshold of state Y; such states are also called T-states. Note that a lower excitability threshold means higher excitability (more sensitive), and a higher excitability threshold means lower excitability (less sensitive). Both W-states for adaptive connection weights and T-states for adaptive excitability have been applied in the model introduced below.

The adaptive causal network model
To better illustrate and analyse the network of multiple adaptive causal pathways, the model can be interpreted in the following way. Central to the model is the adaptivity of perception of various factors, designed in accordance with the JD-R model. Since either external factors or therapy influences perception in a continuous state of change, the perception of these various factors is constantly modulated.
Changes in perception are modelled by the dynamics of the selfmodel W-states for the (perception) connection weights (placed on the first level of reification), thus having a direct downstream modulating effect on the perception of the external and internal factors. The five personality traits directly influence the perception connection weight self-model W-states. This illustrates one of the two types of first-order adaptivity of the network model.
Similarly, the personality traits also directly affect the excitability threshold self-model T-states that have a downstream effect on the three main burnout factors, making them less or more excitable or sensitive.
That is yet another type of first-order adaptivity modelled at the firstorder reification level. The model's primary focus is to observe the behaviour of the three burnout factors; these three factors interact with each other in a certain way, thereby receiving direct or indirect modulating impact from the perception connection weight and excitability threshold self-model W-states and T-states of the first-order reification level.
In addition, the three main burnout factors also receive indirect impact from the second-order level of reification, which is designed to modulate the excitability thresholds for changes in personality traits through time by using second-order self-model T-states for these excitability thresholds. So, although personality traits are often intuitively seen as constant factors, in this adaptive network model they can also change, following the literature discussed in Section 2. Such changes will, in turn, also have a second-order effect on the three main burnout factors. Although these changes are slower than the other changes, they still affect the rest of the states in the model. This represents yet another form of adaptivity, which is actually second-order adaptivity and is modelled by a second-order level of reification.

The base level
The picture in Fig. 1 shows the base level model of our adaptive causal network model. The base level of the adaptive model includes the job demands and resources, building upon empirical evidence for the Job Demands-Resources model in relation to burnout (Demerouti et al., 2001). The base level also includes the three dimensions of burnout: emotional exhaustion, depersonalisation, and reduced personal accomplishment. Based on previous scientific evidence on job demands and resources in relation to the three burnout dimensions, we included links between specific job demands and resources and the three burnout dimensions. For instance, research shows that the perception of high workload (e.g., subjective workload) is positively related to emotional exhaustion and depersonalisation, which is why workload is linked to emotional exhaustion and depersonalisation (Jacobs & Dodd, 2003). Additionally, in line with previous literature, we connected job insecurity to all three burnout dimensions, and the job demand interpersonal conflict was connected to emotional exhaustion and depersonalisation Table 1 Different network characteristics and self-model states for them. Fujiwara et al., 2003). Interestingly, research also indicates that depersonalisation is likely to increase interpersonal conflict, which is why there is also a link from depersonalisation towards interpersonal conflict (Madigan & Kim, 2021;Leiter & Maslach, 1988). As can be seen in Fig. 1, a perception factor is added to simulate a person's perception of job demands and resources. Additionally, we have added a therapy factor in the base level model, as there is reason to believe that certain types of individual therapy, for example, Mindfulness Cognitive Based Therapy (MCBT), can reduce burnout symptoms by adjusting people's perception of reality to perception without emotional or intellectual distortions (dos Santos et al., 2016). Lastly, we included an external state factor in the base level model to introduce a particular external change, for example, the appointment of a new manager in a person's team.

The first-order reification level and the types of adaptation modelled by it: Perception modulation and sensitivity modulation of the burnout dimensions
The model with the first-order reification level added is shown in Fig. 2. The links between the perception (X 1 ) and the job demands (X 2 , X 3 , X 4 ) and resources (X 5 , X 6 ) all have adaptive connection weights since the aim of this model is to investigate how the perception of job demands and resources changes in response to external changes (X 7 ) or therapy (X 11 ). Furthermore, the first-order reification level of our model includes the Big Five Personality traits. By positioning the Big Five Personality states on the first-order reification level of the model, they can be related directly to the self-model states of the network characteristics of the base level. These traits can then play their modulating role for the base-level processes.
Many studies show that personality traits are related to burnout symptoms (Costa and McCrae, 1992). That is why our model's Big Five personality characteristics are linked to the three burnout dimensions, thereby modulating them. For example, as written in Section 2, agreeableness is seen in the literature as a protective factor against burnout, having a negative relationship with all three burnout dimensions (Piedmont, 1993;Zellars et al., 2000). Therefore, in our model, agreeableness is related to all three burnout dimensions. The links of the personality traits to the burnout dimensions go through excitability threshold self-model T-states, which refer to the excitability thresholds of the burnout dimensions (i.e., the degree to which a certain burnout dimension becomes activated). For example, the personality trait neuroticism is positively linked to all three burnout dimensions (Bühler & Land, 2003). Our model reflects a low excitability threshold for all three burnout dimensions related to neuroticism. In other words, the more neurotic a person is, the lower the excitability threshold of developing emotional exhaustion, depersonalisation, or reduced personal accomplishment.
Furthermore, as explained in Section 2, personality changes have been found to be predictive of changes in one's perception of the work environment and experience, which gave us reason to include links in our model from the personality traits to the W-states representing the weights of the connections between perception and the job demands and resources (Langelaan et al., 2006;Sutin & Costa, 2010). Lastly, a context state (X 25 ) is included in the model.

. The second reification level and the adaptation of personality characteristics
As discussed in section 2, there is indication to believe that  personality traits can be adaptive, albeit to a small extent. Several studies show that personality changes are possible, for example, through ageing or clinical interventions such as psychotherapy (Roberts & Mroczek, 2008;Roberts et al., 2017). Therefore, to account for the potential changes in personality traits, a second-order reification level was added. Here, five excitability threshold T-states were added, each linking to one personality trait. Additionally, as seen in the overall secondorder adaptive network model in Fig. 3, links from some of the base-level job demands and resources to the T-states of the personality traits were added. For example, research shows that neuroticism can be strengthened by ineffective responses to stress (Hamilton et al., 2017). Interpersonal conflict is a stress factor that strongly impacts individuals with high neuroticism (Zahlquist et al., 2022). In other words, if a person experiences stress because of interpersonal conflict, they may become more neurotic, which makes them even more sensitive to stress, making this person even more neurotic. This is one of the cyclic effects in the whole process. A link was also made between interpersonal conflict and the T-state for agreeableness since employees that are impacted by a common chronic stressor (e.g., an abusive supervisor) may start to experience a disconnect between motivations related to personality (e. g., agreeableness) and their workplace environment (Barrick, Mount, & Li, 2013), thereby causing a personality change (e.g., a decrease in agreeableness). Lastly, research by Wu and colleagues (2020) showed that chronic job insecurity slightly decreased conscientiousness and agreeableness and increased neuroticism, therefore links were also drawn between those factors. Altogether, we designed the following adaptive causal network model, as shown in Fig. 3. Table 2 gives an overview of the states in the network model, together with the level of the states.

Results of simulation experiments
Several simulation experiments have been conducted. Some of them are discussed below.

Scenario 1 -simulating the impact of an external change on the development of burnout within a person high in neuroticism
Research found that neurotic employees are predisposed to burnout because they generally have a more pessimistic view of their work environment (Langelaan et al., 2006). Furthermore, research has shown that neuroticism is more malleable than previously thought (Barlow et al., 2014). Hence, it was interesting to simulate a scenario focused on neuroticism as the dominant personality trait. To illustrate the adaptivity of neuroticism, a simulation was run displaying how a person high in neuroticism would react to a change in the external circumstances, in this case, the appointment of a new manager. The scenario was set up in such a way that the new manager changed the person's perception of autonomy from high to lower. This perception of low autonomy impacts depersonalisation, which, in turn, increases interpersonal conflict. In Section 4.3 it is indicated that interpersonal conflict has a strong impact on individuals high in neuroticism. Therefore, it was simulated that higher interpersonal conflict makes a person slightly more neurotic, which, in turn, makes the person even more sensitive to the perception of low autonomy.
Given that the person in this scenario already has a predisposition to neuroticism in the first place, the possibility of developing burnout becomes higher. Fig. 4 depicts the model for this scenario. Because the simulation focuses on neuroticism as the dominant personality trait, the initial value of neuroticism was set to a high value (0.8). Similarly, since this scenario simulates a change in external circumstances, the value of the external state was also set to a high initial value (0.9), the effect of autonomy to depersonalisation was set to a negative factor (0.1). For all the factors designed in this simulation Scenario 1, the role matrices and overall simulation graph can be found in Appendix A.
In this scenario, the external state X 7 causes autonomy to be lower. The graph in Fig. 5 shows that the perception of autonomy X 15 (which is W X 1 ,X 5 ) indeed goes down as a result of the external state. The graph further shows that when the perception of autonomy goes down, depersonalisation becomes higher, which increases the perception of interpersonal conflict (X 14 ). Furthermore, in line with what was expected, the perception of interpersonal conflict slightly increases neuroticism. Because neuroticism becomes higher, the chance of Fig. 3. Overall second-order adaptive network model for burnout and Big Five personality traits.
developing burnout increases across all burnout dimensions. As such, the simulation of this scenario in our adaptive network model shows that the response of a person with high neuroticism to a change in external circumstances causing low perception of autonomy, can increase the likelihood of developing burnout.

Background literature on neuroticism and the impact of therapy
Individuals high on neuroticism are more prone to maladaptive emotional regulation (Eysenck & Eysenck, 1985). This makes high neuroticism a strong predictor of burnout. By contrast, mindfulness can be described as being aware of and accepting ongoing emotional experiences and is a protective factor against the negative aspects of neuroticism (Bishop et al., 2004). The neuroticism personality trait has been found to be more malleable than previously believed (Barlow et al., 2014). While interventions explicitly aimed at reducing neuroticism are limited (Armstrong & Rimes, 2016), altering a neurotic state can often be done using similar interventions applied to depression and anxiety (Tang et al., 2009).
Mindfulness-based interventions have been proposed as an alternative to one-to-one cognitive behavioural interventions, as they have been found to be more targeted to neuroticism as well as being more Table 2 Overview of the states in the network model.  (Segal et al., 2002(Segal et al., , 2013Armstrong & Rimes, 2016). Armstrong and Rimes (2016) ran a randomised control trial (RCT) examining the effectiveness of MCBT on neurotic patients. Individuals who received MCBT had lower levels of neuroticism four weeks after the intervention. The active condition neurotic patients also underwent more profound process changes, displaying lower levels of rumination and higher levels of decentring and compassion (Armstrong & Rimes, 2016).
Within the workplace environment, according to a systematic review by Virgili (2015), general mindfulness interventions were found to be effective in reducing the negative psychological effects of the working environment. In an RCT study by Wolever and colleagues (2013), mind-body interventions have also been found to effectively decrease the perception of negative work demands and overall stress levels.
MCBT aims to primarily allow individuals to see themselves as separate from their moods and thoughts. This can enable them to break out of the cycle of negative and often ruminative thinking. Through practices such as meditation and awareness, the individual can, over time, learn to perceive work stressors, such as, for instance, interpersonal conflict mentioned in Scenario 1, in a more non-judgemental and non-reactionary way.

Simulating the scenario for the impact of individual therapy on the development of burnout within a person high in neuroticism
The graph of this simulation is shown in Fig. 6. As in the previous scenario, the initial value of neuroticism is still high, making the perception of the job demands workload, job insecurity, and interpersonal conflict negative (i.e., graph is increasing). However, job autonomy and social support are increasing here. Since the aim was to simulate the response of a person high with neuroticism to adding therapy, the value of state X 11 is initially high. The role matrices and overall simulation graph can be found in Appendix B for all the factors designed for this simulation.
The graph in Fig. 6 shows that the perception of autonomy decreases at first, then increases continually, while the emotional exhaustion decreases until 0 rather fast. Similarly, reduced personal accomplishment is seeing an increase at first, yet it decreases dramatically after the first 25 iterations. As such, the simulation of this scenario in our second-order adaptive model shows that the introduction of therapy to a person high in neuroticism changes their perception to work environmental aspects positively, causing a decrease in emotional exhaustion and, therefore, a decrease in the chance of developing burnout.

Discussion
This paper introduced an adaptive modelling approach to employee burnout, investigating the interactions between the Big Five Personality Traits, the perception of job demands and resources, and the effect of therapy on the development of burnout. The specific focus of this paper was on the personality trait neuroticism, as this personality trait is most closely related to burnout, and research has indicated that this personality trait was more malleable than previously believed (Deary et al., 1996;Zellars et al., 2000;Barlow et al., 2014).

Results obtained
The results of the simulations are promising. The first simulation that explored the impact of an external change (i.e., change of manager leading to lower autonomy) on the development of burnout within a person high in neuroticism showed that, as expected, neuroticism slightly increased as a result of increased interpersonal conflict due to the perceived lower autonomy. Because neuroticism slightly increased, the likelihood of developing burnout increased. The simulation of Scenario 1 thus showed that it is possible to simulate a scenario where neuroticism is adaptive. For the simulation of the second scenario, the focus was on the effect of individual therapy (Mindfulness cognitivebased therapy) on persons high in neuroticism. The simulation of this scenario showed that the response to the therapy of a person high in neuroticism indeed causes a decrease in emotional exhaustion and, therefore, a decrease in the chance of developing burnout. Although individual therapy has proven to be very important in preventing and mitigating the likelihood of developing burnout, more and more research is starting to investigate the role of the organisation in combating burnout (Gabriel & Aguinis, 2022).

Contribution to the current literature
This research contributes to the current literature by addressing the topic of burnout development by using the relatively unexplored method of adaptive causal network modelling. Extending from previous research that modelled burnout development, the current study extends the earlier findings (Weyland, Jelsma & Treur, 2021) by introducing personality traits and work environment in the form of job resources and demands as adaptive, interconnected nodes that could change over time and predict burnout development.
Furthermore, this paper also contributed to the research on potential burnout interventions. The current adaptive model has analysed the effectiveness of MCBT therapy in reducing burnout through its impact on neuroticism. Previous research has found that the MCBT was a potentially effective intervention to reduce burnout, outlining the critical connection between mindfulness and burnout (Askey-Jones, 2018). However, while the MCBT has been established to reduce neuroticism in clinical and subclinical settings, this paper demonstrates its potential to reduce neuroticism in organisational settings (Armstrong & Rimes, 2016). Due to the success of previous mindfulness techniques in the workplace, their combination with the empirically tested CBT could positively impact highly neurotic employees at risk of burnout and allow for more personalised support.

Future research
For future research that focuses on adaptive modelling of burnout, it would be interesting to experiment further with the concept of adaptive personality traits. For example, research indicates that people exposed to chronic job insecurity also have a higher risk of neuroticism and decreased agreeableness and conscientiousness. Therefore, it could be interesting to simulate a similar scenario for the effect of increased job insecurity on neuroticism, agreeableness, and conscientiousness (Wu, Wang, Parker & Griffin, 2020). Both the work environment and personality traits have been found to be very influential on burnout development. However, the connection between them remains a topic that has led to much disagreement and requires more research (Sutin & Costa, 2010). Using an adaptive network modelling approach can allow future researchers to simulate how real-life external influences can impact one's perceptions of the work environment and the effect that has on personality.
Furthermore, future research on burnout can extend the intervention simulations to compare the effectiveness of various interventions on personality trait change and burnout reduction. An evidence review by Public Health England (2016) found that organisational-level interventions could produce longer-lasting positive effects against burnout compared to individual-level interventions. Alterations within the organisational culture and working practices have been found to significantly decrease burnout-related stressors. Therefore, it would be valuable to explore how organisational changes could impact one's perceptions of demands, resources and whether that can lead to any personality changes (Awa et al., 2010). Comparing the adaptive changes between different organisational and individual level interventions could better inform organisational psychologists and therapists on the most effective ways of tackling burnout depending on personality traits and the work environment.
Additionally, it would also be interesting to examine whether the current shift towards remote working and the general transition towards online professional communication has impacted the development of emotional burnout. Future research that uses the adaptive causal network modelling approach could examine the perceptions of job demands that are specific to remote working such as constant online presence and social isolation. The impact of these demands on the three components of emotional burnout can be evaluated, also examining whether certain personality traits could lead to larger burnout effect.
In summary, the analysis based on the designed adaptive causal network model presented in this paper makes a valuable contribution to the current understanding of burnout development modelling by representing personality traits as adaptive factors while also showing many possibilities for expansion in the future.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability
Data will be made available on request.