Neuroticism-Withdrawal and Neuroticism-Volatility Differently Influence the Risk of Neuropsychiatric Symptoms in Alzheimer’s Disease

INTRODUCTION

Behavioral and psychological symptoms (BPS) are core manifestations of Alzheimer’s disease (AD). They may appear as early as the prodromal stage and tend to worsen as the disease progresses, with significant inter-individual variations [1]. The triggers and causes of BPS are complex, involving neurological, somatic, psychological, and environmental factors. Among these factors, personality is believed to play an important role [2]. Personality is defined as a set of cognitive, affective, and behavioral features stable over time, making one person recognizable among others. Clinical observations and a growing number of publications suggest that differences between individuals as to prior personality could significantly modulate and shape the behavioral expression of AD [3, 4]. Nevertheless, the way in which personality influences BPS remains to be clarified. Some BPS could be the enhanced manifestation of previous personality traits, or reflect inappropriate coping reactions in response to a stress generated by cognitive decline and its consequences. Most studies investigating the links between personality traits and BPS were based on the five-factors model, which provides a description of personality across five broad domains: neuroticism, conscientiousness, extraversion, agreeability, and openness. Neuroticism is recognized as the personality domain that is the most strongly associated with BPS [3]. This domain refers to the tendency to experience emotional instability and negative emotion. Individuals with a high level of neuroticism are more sensitive to minor stressors, tend to interpret day-to-day situations as threatening, feel guiltier or excluded, are more impulsive, and have difficulties in managing their emotions and frustrating situations. In non-cognitively impaired populations, neuroticism was associated with common mental disorders such as depression, anxiety, or substance abuse [5] as well as a number of personality disorders including borderline, avoidant, and dependent personality disorders [6]. In people with dementia, several studies highlighted a relationship between neuroticism and depression, anxiety, aggression, apathy, and psychotic symptoms [7–11]. Most of these studies had a cross-sectional design and were conducted in small populations of patients with advanced dementias, in whom premorbid personality assessment could be subject to bias [3]. Very few investigations have been conducted in mild cognitive impairment of AD and their findings deserve further investigation [12, 13]. Indeed, in the early stage of AD, the personality is still well preserved, and could therefore have a more specific impact on NPS than in later stages. A recent longitudinal survey of almost 2,000 patients with a non-specified dementia diagnosis confirmed the impact of neuroticism on various kind of BPS [14]. To date, a body of evidence in favor of a pathoplastic effect of neuroticism on the behavioral expression of AD have emerged [15]. Nevertheless, it is worthy to note that neuroticism can be viewed as a heterogeneous construct encompassing different psychological phenotypes reflecting vulnerability to stress. The five-factors model posits that personality traits are organized hierarchically, each of the broad domains being composed of distinct sub-domains called facets. These facets reflect more specific patterns of behavioral, emotional, and motivational characteristics than domains. The NEO-PI-R inventory [16], which is one of the most widely used in personality research, distinguishes six facets by domain, but no consensus on the number of facets in each domain has been established [17, 18]. In a research based on factor analytic studies and supported by genetic findings, Deyoung [18] has pointed out the relevance to add an intermediate level between domains and facets. Each Big Five domain can be broken down into exactly two sub-domains labelled aspects of personality and hierarchically located above the facets. The individualization of aspects is of interest because they are believed to be partially supported by distinct genetic bases and specific brain structures and more accurately, reflect the phenotypic variations existing within each of the personality domains [19]. According to this model, neuroticism is composed of two aspects, i.e., neuroticism-withdrawal (N-withdrawal) and neuroticism-volatility (N-volatility). N-withdrawal refers to the tendency to internalize negative emotions and is clinically associated with sadness, ruminations, and anxiety [18, 21]. N-volatility reflects the predisposition to externalize negative emotions and is related to behavioral manifestations such as anger, irritability, or emotional lability [18, 22]. The intensity of these two aspects determines how individuals react in stressful situations such as threats, uncertainties, or adverse life events. There is emerging evidence that the two aspects are related with different types of psychopathology. For example, N-withdrawal was linked to higher risk of depression, while N-volatility has been found to be correlated to bipolar disorder [20, 23].

Previously, we showed that elevated neuroticism predicted severe BPS occurrence, particularly with respect to affective symptoms, sleep disorder, and apathy [4]. The objective of the current investigation was to complete and extend our prior study by exploring the impact of N-volatility and N-withdrawal on BPS. Our analyses have focused only on neuroticism since this domain of personality is the most strongly associated with BPS and psychopathology [3, 5]. We hypothesized that N-withdrawal and N-volatility would differently modulate the behavioral expression of AD. We expect N-withdrawal to be positively associated with depression and anxiety, whereas N-volatility should be positively linked to BPS reflecting impulsiveness as manifested by agitation, aggression, irritability or emotional lability.

METHODS

RESULTS

DISCUSSION

The goal of this research was to investigate prospectively the impact of N-withdrawal and N-volatility on BPS risk in patients with AD from prodromal to mild stage. To our knowledge, these personality traits have never been examined in AD. This work completes prior research on the same cohort, which had demonstrated a positive association between elevated neuroticism and the NPI-Q global score as well as sleep disorders, affective and apathy symptoms [4]. Similarly, our current results show that N-volatility and N-withdrawal are independently related to the worsening of the NPI-Q global score. Moreover, these two aspects influence the typology of BPS in different ways. Thus, N-withdrawal was positively associated with the risk of anxiety and depression, while N-volatility was positively related to the occurrence of delusions, agitation/aggression, irritability/lability, and apathy. Appetite and sleep disorders were linked with both aspects. In our previous study, agitation/aggression, irritability, and delusional ideas were not associated with neuroticism, unlike the aspects underlying neuroticism [4]. This demonstrates that the analysis of the broad domains of the five-factors model can smooth out significant personality features that are only captured when considering lower levels of personality. This observation reinforces the relevance of not limiting personality research to the broad dimensions of the five-factors model alone [20, 23].

Our results clarify the interaction between premorbid neuroticism and BPS as they demonstrate that aspects of neuroticism differently influence BPS. Numerous studies have highlighted an association between neuroticism and various BPS, including depression, anxiety, agitation, apathy, delusion, or hallucinations [3, 14]. Nevertheless, BPS previously found to be influenced by neuroticism were different in nature raising questions about the etiological pathways mediating these associations. Our findings suggest that it is possible to group these BPS in a more coherent way after accounting on N-withdrawal and N-volatility.

Unsurprisingly we found a very significant relationship between N-withdrawal and both depressive and anxious symptoms. N-withdrawal reflects the predisposition to internalize negative emotions such as sadness, ruminations, and feeling of guilt [18, 22]. Links between elevated neuroticism and affective disorders have been well established in adult [5] and elderly population [30]. In AD, the relationship between neuroticism and affective disorders has also been shown by a number of studies [3, 31]. In a study including 208 patients with probable AD, Archer et al. showed an association between neuroticism and anxiety but not with depression [7]. At the facet level, Chatterjee [8] reported that depressive symptoms were linked with an association of facets comprising depressive, anxiety, and self-conscious facets. This combination of facets comes very close to N-withdrawal as defined in our study. Furthermore, recent research in non-cognitively impaired populations revealed that the impact of neuroticism on depression was probably mediated by N-withdrawal, but not N-volatility [20, 23]. Our results suggest that the same is true in AD and promote the interest of considering N-withdrawal when examining affective symptoms.

We had not predicted the link between N-volatility and delusions. In people without any neurocognitive disorder, the relationship between psychotic symptoms and neuroticism have already been stressed. High neurotic individuals with schizotypal traits or suffering from schizophrenia have higher rejection sensitivity or more severe delusions [32, 33]. Moreover, neuroticism is also considered as a contributing factor for minor delusional ideation experienced by individuals without any psychiatric condition [34]. Delusions are common manifestation of AD that may worsen over time [35]. Some studies suggest neuroticism could favor delusional ideas of AD [8, 10]. In line with our results, Chatterjee [8] found a significant association between hostility, a facet encompassed by N-volatility, and delusions. In clinical practice, individuals with hallucinations often experience delusional ideas, but the opposite is not true. Our results indicate that N-volatility specifically influences delusion but not hallucination risk. Individuals with high-levels of N-volatility are hypervigilant to negative environmental cues, have higher reactivity to stress, tend to interpret ambiguous or neutral stimuli in a negative way and are more prone to generate negative attributions [21, 22]. Thus, psychological distress induced by AD might accentuate the premorbid negative attribution bias characterizing individuals with high N-volatility and increase the likelihood of delusion emergence.

An interesting finding was the association of N-volatility with agitation/aggression and irritability/lability. Previous research has provided mixed information on the links between personality and these challenging behaviors. For example, agitation, aggression, and impulsivity have been predicted by either a low or a high level of agreeableness [7, 11]. A large prospective study focusing on a population with non-specified dementia found a significant relationship between premorbid neuroticism and uncontrolled temper [14]. People with elevated levels of N-volatility tend to externalize their negative emotion and are characterized by anger, hostility, emotional instability, and impulsivity [21, 22]. Agitation/aggression and irritability/lability in AD are often grouped under the umbrella term of hyperactivity behavior, which may be linked to impulsivity [36]. Impulsivity is a complex construct including various behavioral and emotional manifestations. Whiteside and Lynam [37] identified four different components of impulsivity: sensation-seeking, lack of planning, lack of perseveration, and urgency. Urgency is defined as the tendency to act rashly when experiencing negative emotions. It has been found to load highly on the impulsiveness facet of neuroticism, which is part of N-volatility [37]. Cognitive decline in AD may alter the modulation of emotion-based reactivity and so, may potentiate the impulsive tendency and emotional lability of individuals with high premorbid N-volatility. Agitation/aggression and irritability/lability reported in our study and found to be linked with N-volatility could be a clinical expression of the concept of urgency described by Whiteside and Lynam [37]. Other manifestations of impulsive behavior, such as disinhibition frequently encountered in AD with frontal syndromes could be favored by personality traits such as extraversion, involved with positive emotions and sensation-seeking. A recent meta-analysis demonstrated that impulsivity is modulated by different personality traits including conscientiousness and extraversion [38]. It would therefore be interesting to extend our researches with further analyses accounting for the others domains and their relative sub-components.

Apathy is a common manifestation of AD whose links with personality traits remain unclear. Previous studies have provided controversial findings. High level of extraversion and agreeability have been found to be potentially protective factors against apathy, whereas high openness could have a negative impact [7, 12]. We have previously found that neuroticism could promote the development of apathy [4]. The current study goes further by revealing that only N-volatility seems to mediate this relationship. At first glance, this finding may seem unexpected and counter-intuitive. Nevertheless, the clinical association between apathy and externalized behaviors, such as uncooperative agitation or anger, have been reported [39]. Apathy and impulsive behavior may share common pathophysiological mechanisms, involving prefrontal brain structures, subcortical regions and limbic motivational loops [40]. Patients with high levels of N-volatility could be more sensible to dysregulation in this system, leading to altered motivational control and to an alternation between apathy and externalizing behaviors.

Neurobiological models of personality may provide other insights. Deyoung [18] drew a parallel between N-volatility, N-withdrawal, and the neurobiological networks governing response to threat and punishment proposed by the revised Gray’s reinforcement sensitivity theory [41]. Gray and McNaughton [41] distinguished two brain networks processing aversive stimuli, called Fight-Flight-Freeze System (FFFS) and Behavioral Inhibition System (BIS), respectively. The FFFS, which is activated when a negative cue has been detected, generates defensive behaviors such as escape, freezing, or attack depending on the nature of the stimulus or the imminence of the danger [42]. The activation of this system results in experiencing fear mainly, but also manifests with emotional lability and reactions of anger or aggressiveness. The triggering of the BIS occurs in situations of uncertainty with potential threats requiring an accurate environmental assessment. A sensitive BIS is accompanied by inhibition, increased arousal and is associated with experience of anxiety, worrying and rumination. The level of sensitivity of these systems varies from one individual to another, explaining the differences observed within neuroticism. From this perspective, N-volatility is supposed to reflect the sensitivity of the FFFS, while N-withdrawal should depend on the sensitivity of BIS [18, 42]. In light of these elements, a new understanding of links between neurotic personality, brain networks and BPS could be put forward. Anxiety and depression associated with N-withdrawal could be due to an overactivation of the BIS, while agitation, aggression, irritability, lability, impulsiveness, and delusions associated with N-Volatility could be the behavioral expression of an overactivation of the FFFS. Links between N-Volatility and apathy could be interpreted as an inappropriate freezing behavior, but this hypothesis remains highly speculative since no big emotional load is usually associated to apathy. The brain structures underlying neuroticism, BIS and FFFS are partly common and rely on distributed areas including, but not exclusively, amygdala, hippocampus, and anterior cingulate cortex [17, 41]. The same brain areas are typically involved in AD [43]. We posit that AD lesions combined with a pre-existing vulnerability of limbic brain structures characterizing individuals with high levels of N-withdrawal or N-volatility would favor the development of specific BPS. Neuroimaging investigations of personality provide additional evidences. Reduced volumes of prefrontal regions, decreased connectivity within the limbic system, and between prefrontal cortex and limbic systems have been linked with neuroticism [44–47]. The extension of AD lesions toward these strategic regions known to be crucial in the top-down regulation of emotions could also potentiate behavioral, emotional, and cognitive deregulation in highly neurotic people and favor the exacerbation of personality predispositions.

Several limitations must be taken into consideration. Personality assessment in context of AD is subject to caution. Indeed, patients with AD have metacognitive impairments that can bias their perception and self-evaluation of their prior personality. Nevertheless, patients with mild AD tend to report their prior personality when asked to describe their current personality due to self-image deficit updating [27]. Patients included in our study were suffering from mild AD; we therefore considered the self-assessment of personality to be reliable. Our method to extract N-withdrawal and N-Volatility from the facets of neuroticism using a PCA can be discussed. Nonetheless, factors of our PCA were relevant with Deyoung’s findings [18], reinforcing his assumption of a multi-level hierarchical organization of personality. Due to the study design, only participants with no or mild BPS were included in this research. We cannot therefore conclude to what extent personality influences BPS in beginning AD. Another limitation was the short duration of the longitudinal follow-up, which does not allow determining whether N-withdrawal and N-volatility may have an influence on BPS in patients with advanced AD. Finally, we did not control our statistical analyses on psychotropic drugs. Indeed, the relationship between psychotropic drugs and the intensity BPS is difficult to interpret. Individuals experiencing BPS are more are risk to be prescribed psychotropic medications, so prescriptions could be associated with higher BPS intensity. On the opposite, psychotropic drugs when efficient could diminish BPS. Previous large studies investigating links between BPS and personality used a similar methodology to ours [7, 14].

The large number of participants included, the prospective design, the diagnosis of AD based on standardized criteria, and the detailed personality assessments are strengths of this study. The originality of our research was to specifically focus on the N-Withdrawal and N-Volatility which have never been explored before.

To conclude, we have found that individuals with mild AD were at risk to develop different types of BPS depending on their level of N-withdrawal or N-volatility. In adulthood high level of neuroticism may be by itself not sufficient to induce psychopathology [48]. AD in highly neurotic individuals may prompt the expression of a latent psychiatric vulnerability hardly expressed over the previous life course. So, some synergy must be considered between neuropathological lesions, environment, psychological status, and personality in the genesis of BPS in AD. Finally, our findings could be relevant to clinical practice. BPS are disabling manifestations that are among the most difficult symptoms to manage in AD. Drug treatments have to be avoided in this frailty population because of the high risk of side effects. Non-pharmacological treatments of BPS are recommended as first-line treatments but their effectiveness is debated. Our study suggests that early personality assessment, with a particular focus on neuroticism and its sub-components, could help to identify individuals at high risk of developing BPS. Therefore, appropriate tailored preventive interventions based on the personality profile could be developed.