Neuroticism is one of the Big Five dimensional domains of personality that encompasses individual differences in negative emotional responding to frustration, threat, or loss^1–4. It is operationally defined by items, which, in aggregate, reflect anger, sadness, irritability, self-consciousness, vulnerability, worry, and hostility that correlate well with one another in factor analyses^1–3. Although neuroticism is associated with manifestations of psychological illness, such as mood and anxiety disorders; adverse physical health outcomes; and poorer quality of life¹, comparably less is known about its neurochemical substrate, especially in common mental disorders. This gap in the literature hinders development of new therapeutics that could serve to mitigate the effects of high neuroticism.

Borderline personality disorder (BPD) is a debilitating illness characterized by severe emotional dysregulation, tumultuous interpersonal relationships, and self-harming and suicidal behavior⁵. According to meta-analytic review, BPD is the personality disorder most strongly associated with neuroticism^6,7. Additionally, longitudinal analysis points to robust associations between BPD symptomatology and neuroticism⁸. Antisocial personality disorder (ASPD) is another common condition that shares clinical similarities with BPD. Expression of neuroticism positively correlates with ASPD, especially with symptoms of anger and impulsivity⁹, although meta-analysis shows smaller effect sizes for ASPD and neuroticism than with BPD and the latter⁶. Clinically, a significant positive relationship exists between neuroticism in childhood or adolescence and adult antisocial behavior¹⁰. ASPD and BPD are highly comorbid¹¹. Higher neuroticism scores are seen in ASPD with comorbid BPD, especially among individuals with substance use disorders¹².

Advances in the study of endocannabinoid signaling pathways have furthered our understanding of the neurochemistry of myriad psychiatric disorders (for review, see¹³). The brain endocannabinoid system (ECS) is comprised of two main G protein-coupled cannabinoid receptors, cannabinoid receptor 1 (CB1R) and cannabinoid receptor 2 (CB2R) and two principal endogenous cannabinoids (e.g., endocannabinoids), N-arachidonylethanolamide (anandamide or AEA) and 2-arachidonylglycerol (2-AG). Other endocannabinoid-related signaling lipids, such as the N-acylethanolamides, N-palmitoylethanolamine and N-oleoylethanolamine, are also known to exist but are less well-studied¹⁴. In a distinctive signalling pathway, AEA and 2-AG are synthesized on demand in the post-synaptic neuron and released as required into the synaptic cleft and bind, in retrograde fashion, to pre-synaptic CB1R and CB2R located on axon terminals on the pre-synaptic neuron¹⁵. To terminate endocannabinoid signalling, fatty acid amide hydrolase (FAAH), an integral membrane serine hydrolase, degrades AEA in the post-synaptic neuron, while 2-AG is metabolized by the enzyme monoacylglycerol lipase in the pre-synaptic neuron¹³. CB1Rs are promiscuous in the central nervous system (CNS) with high levels in the cerebral cortex, hippocampus, amygdala, striatum, and substantia nigra, and they play functional roles in regulating mood, stress, and anxiety^16,17. FAAH is also widely distributed in the CNS, including the cerebral cortex, hippocampal formation, amygdala, and cerebellum with its regional distribution correlated with the density of CB1Rs^18–21. Therefore, enzymatic regulation of AEA by FAAH is able to indirectly control CB1R signaling and ostensibly modulate cognitive processes underlying mood and anxiety symptoms²².

Using [¹¹C]CURB positron emission tomography (PET), we reported that FAAH expression was elevated in the prefrontal cortex (PFC) of BPD and that PFC FAAH density correlated with measures of anger/hostility²³. Conversely, we also found utilizing [¹¹C]CURB PET that amygdala FAAH expression was lower in the amygdala of ASPD and that impulsive behaviors were negatively associated with FAAH density in other brain regions studied²⁴. These findings have enhanced our knowledge about the pathophysiology of BPD and ASPD and their symptom clusters, yet an unanswered question remains how FAAH expression may relate to personality traits such as neuroticism. There is precedent for investigating FAAH and neuroticism in other psychological disorders. For example, individuals with posttraumatic stress disorder and comorbid alcohol dependence who were wildtype for the FAAH gene displayed increased subjective anxiety responses to a stress challenge versus subjects with a single nucleotide polymorphism of the FAAH gene conferring lower levels of brain in vivo FAAH expression who reported decreased anxiety for the same stress paradigm²⁵. Importantly, all analyses in this study controlled for trait neuroticism. Based on these findings, we aimed to investigate FAAH brain expression in a combined sample of BPD and ASPD participants. The most consistent finding from structural magnetic resonance imaging (sMRI) studies of trait neuroticism has been structural alterations of the PFC and its subdivisions among individuals with higher neuroticism^26–31. Therefore, we focused on the PFC and hypothesized that trait neuroticism in BPD and ASPD would be associated with greater PFC FAAH binding, although we also explored other brain regions to test this association. It should also be noted that the association between ASPD and neuroticism is weaker than for ASPD and agreeableness³². Since the ECS may also be related to agreeableness³³, and AEA-mediated signaling at CB1R, driven by oxytocin, appears to control social reward³⁴, we also investigated whether the agreeableness dimension would be associated with FAAH binding in the PFC and other brain regions.

Participants’ clinical and demographic information is reported in Table 1. Information on comorbid conditions is presented in Table 2.

This study investigated the relationship between brain FAAH binding and neuroticism in a sample of ASPD and BPD participants. Consistent with our main hypothesis, we found that PFC [¹¹C]CURB λk₃ correlated positively with trait neuroticism. Exploratory analyses determined that PFC FAAH binding was linked to greater anxiety and self-consciousness. A second main finding was that dorsal putamen [¹¹C]CURB λk₃ similarly correlated positively with neuroticism. In terms of facet-level associations, dorsal putamen FAAH binding was positively associated with anxiety, depression, and self-consciousness. These findings have relevance for understanding the neurochemistry of personality traits in BPD and ASPD and offer tentative guidance for testing FAAH inhibitors in personality-disordered individuals scoring high in neuroticism.

There are very few studies that have explored in vivo measures of personality functioning in ASPD or BPD and none to our knowledge that have specifically assayed components of the ECS, apart from one functional MRI study of healthy individuals that found no relationship between neuroticism and amygdala [¹¹C]CURB λk₃³⁵. Here, we focused on the relationship between FAAH binding and neuroticism in personality disorders. Our findings comport well with the animal literature linking pharmacological inhibition or genetic deletion of FAAH to the rescue of depressive and anxious phenotypes³⁶. According to this model, decreased FAAH activity increases brain AEA levels and stimulates anxiolytic-like responses in a CB1R-dependent manner³⁷. In support of this view, CB1R activation in the prefrontal cortex, ventral hippocampus, and periaqueductal grey elicits anxiolytic-like effects^38,39. Because maladaptive behavioral and neuroendocrine reactions to protracted stress contribute to manifestation of depressive-like behaviors⁴⁰, it is thought that facilitation of AEA signaling may counter these effects by enhancing adaptive stress coping behaviors⁴¹ and diminishing neuroendocrine responses to psychological stressors⁴². Overall, this research suggests that augmentation of AEA-CB1R signaling through inhibition of FAAH activity may be an effective strategy for alleviating depressive and anxious symptomatology. By contrast, dampened AEA-CB1R circuitry through increased FAAH expression may contribute to increased levels of neuroticism, including trait anxiety and depression, which was observed in our sample of ASPD and BPD participants.

Bolstered by sMRI data revealing alterations of the PFC in trait neuroticism^26–31, we similarly detected elevated PFC [¹¹C]CURB λk₃ as a function of increased neuroticism in ASPD and BPD. The PFC, which broadly plays a role in affective processing and emotion regulation⁴³, has been previously identified as an important locus for expression of facets of neuroticism in BPD. For example, Soloff and colleagues reported that trait impulsivity, one facet of neuroticism, was related to decreased serotonin-2A receptor (5-HT2AR) binding potential using [¹⁸F]altanserin PET in the medial frontal cortex among female BPD participants⁴⁴. Studies have also examined the association of PFC 5-HT2R binding with trait neuroticism in healthy cohorts, with some findings showing contradictory findings that could be a reflection of differences in the neural substrates between clinical and healthy populations. One investigation found a positive correlation between frontolimbic 5-HT2AR and neuroticism in healthy volunteers that was mainly driven by the anxiety and vulnerability facets⁴⁵. In another PET study of healthy participants using the radiotracer [¹¹C]WAY-100635, trait anxiety was inversely correlated with 5-HT1AR binding in the dorsolateral PFC (DLPFC)⁴⁶. A similar negative correlation was observed between DLPFC [¹¹C]WAY-100635 binding potential and trait neuroticism in another sample of healthy subjects⁴⁷. Since prominent linkages between the ECS and serotonergic system have been described⁴⁸, we suggest that the positive association of PFC [¹¹C]CURB λk₃ with neuroticism may be partially explained by interactions of the ECS with serotonergic signaling pathways. This line of reasoning is supported by preclinical evidence showing that basal serotonergic tone is higher in the frontal cortex of FAAH knockout mice following potassium-induced depolarization⁴⁹ and that downregulation of 5-HT2AR and 5-HT2CR occurs in the PFC⁴⁸, perhaps as a result of higher endogeneous levels of 5-HT. These neurochemical changes were paralleled by anxiolytic-like effects. Other reports have signaled that pharmacological inhibition of FAAH leads to increased brain-derived neurotrophic factor and neurogenesis in other brain regions functionally linked to the PFC⁵⁰. In light of this evidence, we propose that increased PFC FAAH could lead to decreased endogenous 5-HT tone in affected regions that predisposes to greater expression of neuroticism. Dual radiotracer PET experiments that simultaneously measured PFC [¹¹C]CURB λk₃ and 5-HT1AR or 5-HT2AR binding potential could shed light on this question.

The association between dorsal striatum [¹¹C]CURB λk₃ and neuroticism may provide a mechanism to understand impaired decision-making in personality disorders. While the involvement of the dorsal striatum in control of motor movement has been well-established, evidence further implicates the dorsal striatum and its corticostriatal network in action selection and initiation⁵¹. Poor decision making has been demonstrated in ASPD and BPD⁵², and poor decision making has been linked to increased neuroticism in some samples⁵³. Therefore, ASPD and BPD with high neuroticism may be poor decision makers. Additionally, some sMRI investigations have also shown connections between trait neuroticism and functional alterations of the striatum⁵⁴. Preclinical research indicates that receptors involved with striatal synaptic plasticity are crucial for striatal-based learning⁵⁵. One investigation determined that striatal CB1R receptor deletion impaired habit learning⁵⁶. Preclinical models further demonstrated that FAAH inhibition can preserve the structural integrity of the striatum and prevent neuronal loss following an excitotoxic lesion⁵⁷. These results indicate a role for endocannabinoids in habit learning⁵⁵ and suggest that higher FAAH binding associated with greater neuroticism may produce deficits resulting in impaired decision making in ASPD and BPD.

When we examined the relationship between PFC [¹¹C]CURB λk₃ and neuroticism in individuals who had comorbid MDD, we did not find a significant correlation. However, a trend association emerged for the relationship between PFC [¹¹C]CURB λk₃ and neuroticism among participants with current anxiety disorders, which aligns with our other finding of a significant correlation between facet anxiety, but not facet depression, and PFC [¹¹C]CURB λk₃. Thus, the association between PFC FAAH binding and neuroticism may be more driven by facet-level anxiety. These results coincide with clinical observations that FAAH inhibitors have shown some benefit for treatment of anxiety disorders⁵⁸.

Several limitations of the present study must be acknowledged. First, we combined ASPD and BPD participants to investigate them as a single group. We justified collapsing ASPD and BPD into a single group, because the conditions are both Cluster B personality disorders, both are highly comorbid with one another¹¹, and both show high levels of neuroticism. In fact, there was no difference in trait levels of neuroticism between the ASPD and BPD groups (ASPD = 71.4 ± 6.9, BPD = 74.8 ± 8.3; t =  − 1.1, df = 29, p = 0.26), which suggests that neuroticism levels did not relate to gender in this sample. There is also precedent for studying Cluster B personality disorders as a single group in biological studies^59,60. Finally, using standardized T scores of neuroticism for all of the correlational analyses allowed for some comparison across genders. Collapsing ASPD and BPD into one group did enable us to increase the study power. Still, our investigation was likely underpowered to detect effects for the secondary ROIs; all regions showed correlations between [¹¹C]CURB λk₃ and neuroticism in the expected direction but did not survive correction for multiple comparisons. Second, as this study was cross-sectional, the design cannot inform on whether PFC [¹¹C]CURB λk₃ is stable to changes in trait neuroticism over time. Third, we did not sample for serum levels of endocannabinoids, such as AEA and the endocannabinoid-related signaling lipids N-palmitoylethanolamine and N-oleoylethanolamine that may be correlated with in vivo brain levels of FAAH expression. While measuring serum endocannabinoids could provide peripheral markers of FAAH expression, no clear-cut relationship has been demonstrated between peripheral and central FAAH biomarkers⁶¹. A fourth limitation is that none of our BPD or ASPD participants had an active substance use disorder. We purposely excluded individuals with active substance use disorders to avoid confounds of psychoactive substances on the central marker of interest. However, individuals with BPD or ASPD are noted to have elevated rates of active substance use disorders, perhaps upwards of 70% in some samples⁶². Thus, our results may not be generalizable to the larger, more common, group of BPD and ASPD individuals with comorbid substance use disorders. Finally, since [¹¹C]CURB is an irreversible tracer, λk₃ is a proxy measure of FAAH availability, as absolute quantification is not possible.

In summary, we have confirmed previous results of high neuroticism in ASPD and BPD and report that PFC and dorsal caudate [¹¹C]CURB λk₃ are positively correlated with trait neuroticism. With the investigation of FAAH inhibitors for anxiety disorders and anxious endophenotypes, our data support the careful testing of these pharmacological agents in personality disorders with high neuroticism. Establishing an evidence base for novel therapeutics in personality disorders would represent a positive step forward, as authorized pharmacological treatments for ASPD and BPD are currently lacking.

All participants provided written informed consent after all study components were fully explained to them. The Research Ethics Board of the Centre for Addiction and Mental Health (CAMH) in Toronto, Ontario, Canada, approved all procedures of this investigation. All methods were performed in accordance with the relevant guidelines and regulations.

N.J.K., I.B., and R.M.B. conceived and designed the study idea. N.J.K. participated in data collection. N.J.K. wrote the first draft of the manuscript. N.J.K., I.B., and R.M.B. critically reviewed the manuscript. All authors contributed to and approved the final version of the manuscript.

This work was funded by Canadian Institutes of Health Research (CIHR) Project Grant awarded to NJK and a CIHR Clinician-Scientist Salary Award also awarded to NJK.

The authors declare no competing interests.