Consistent evidence that brain serotonin 2A receptor binding is positively associated with personality-based risk markers of depression

Abstract

Background

Using [¹⁸F]altanserin, a serotonin 2A receptor (5-HT_2AR) antagonist Positron Emission Tomography (PET) tracer, a positive association between cortical 5-HT_2AR binding and the inward-directed facets of neuroticism has been demonstrated in healthy individuals. Psilocybin, a 5-HT_2AR agonist, shows promise for the treatment of depression, reducing neuroticism and mood symptoms potentially via hypothalamic-pituitary-adrenal (HPA) modulation. 5-HT_2AR and neuroticism are both modulated by HPA axis function.

Aims

In this study, we examined whether the association between 5-HT_2AR binding and the inward facets of neuroticism can be replicated in an independent healthy cohort using the new 5-HT_2AR agonist tracer [¹¹C]Cimbi-36, and if their association is moderated by cortisol awakening response (CAR), an index of HPA axis function. If so, this could advance mechanistic insights into interventions that target the 5-HT_2AR and reduce neuroticism.

Method

Eighty healthy volunteers underwent [¹¹C]Cimbi-36 PET scans and completed the NEO personality inventory (NEO-PI-R) for the assessment of neuroticism. Salivary samples were available for determination of CAR in 70 of the participants. Using linear latent variable models, we evaluated the association between 5-HT_2AR binding and inward facets of neuroticism, namely depression, anxiety, self-consciousness and vulnerability to stress, and whether CAR moderated this association.

Results

The study confirms the positive association between 5-HT_2AR binding and the inward facets of neuroticism (β = 0.01, 95% CI = [0.0005: 0.02], P = 0.04), and this association is independent of CAR (P = 0.33).

Conclusions

The findings prompt consideration of whether novel interventions such as psilocybin that actively targets 5-HT_2AR and causes changes in personality could be particularly beneficial if implemented as a targeted approach based on neuroticism profiles.

The personality trait of neuroticism shapes individuals’ reactions to stressful situations¹ and is a risk factor for the onset of psychiatric illnesses, especially major depressive disorder (MDD).² Research into the neurobiology of MDD strongly supports the role of the serotonin 2A receptor (5HT_2AR) system in the disorder's pathophysiology^3–6 and in behavioural phenotypes that are risk factors for depression including stress sensitivity,⁷ emotion processing,⁸ and personality traits such as pessimism (or negative styles of thinking) and neuroticism.^9,10 In one of the largest reported serotonin receptor imaging studies in healthy controls (n = 83), an association was shown between 5-HT_2AR binding and neuroticism using the [¹⁸F]altanserin tracer.¹⁰ Subsequently, it was also shown that this link between serotonergic function and neuroticism was stronger in individuals with a high familial risk of developing depression, relative to individuals with a low familial risk.¹¹ Interestingly, in these reports, the facets or dimensions of neuroticism that were closely linked to 5-HT_2AR binding were those that are directed ‘inwards’, namely depression, anxiety, self-consciousness and vulnerability to stress, as opposed to the impulsivity and angry hostility facets which are directed ‘outward’ when faced with frustration.¹¹ Thus, the inward-directed facets of neuroticism may be particularly sensitive to modulation by 5-HT_2AR function.

In the present study, we examined whether there exists an association between 5-HT_2AR and the inward facets of neuroticism in healthy individuals when imaged with a new 5-HT_2AR agonist Positron Emission Tomography (PET) tracer [¹¹C]Cimbi-36. If indeed the previously observed association between cortical 5-HT_2AR binding and inward neuroticism represents a generalisable phenomenon across different PET tracers, scanner conditions and populations, then this could point to a possible mechanism by which drugs that target the 5-HT_2AR change personality structures and alleviate depressive symptoms or other psychopathology. For instance, psilocybin exerts its psychoactive effect as an agonist of the 5-HT_2AR¹² and is considered a promising drug for the treatment of MDD. Psilocybin's stimulation of the 5-HT_2AR appears to significantly reduce neuroticism and leads to alterations in mood, perception and consciousness.¹³ It is posited that the hypothalamic-pituitary-adrenal (HPA) axis function may play a key role in psilocybin's salutary effects on mood.¹⁴ A method to characterise the HPA axis function is by measuring the cortisol awakening response (CAR). CAR indexes the increase in cortisol levels that occurs within the first hour upon morning awakening and is posited to be a marker of naturally occurring physiological (stimulated) cortisol response.¹⁵ The function of 5-HT_2AR and neuroticism are both closely related to cortisol dynamics. Evidence from preclinical studies indicates that 5-HT_2AR is upregulated by various stressors⁷ and stimulates the production of corticotropin-releasing factor which promotes the production of cortisol.¹⁶ Likewise, individuals with high neuroticism typically exhibit heightened stress reactivity,^17,18 and higher cortisol secretion.¹⁹ Together, these studies suggest that 5-HT_2AR and neuroticism are closely linked to cortisol dynamics and that their association may be influenced by variations in cortisol dynamics. Therefore, we include cortisol markers to examine if the association between 5-HT_2AR binding and the inward facets of neuroticism is dependent on cortisol dynamics in terms of CAR. We hypothesise that there is a significant association between 5-HT_2AR binding and inward facets of neuroticism, and furthermore that this association is moderated by CAR. Together, the study findings may advance mechanistic understanding of how interventions that target the 5-HT_2AR reduce neuroticism and mood symptoms.

Method

Participants

We identified individuals within the Center for Integrated Molecular Brain Imaging (CIMBI) database²⁰ that met the following criteria: (a) completed a baseline [¹¹C]CIMBI-36 scan, (b) completed an assessment of neuroticism using the NEO-Personality Inventory-Revised (NEO-PI-R),²¹ and (c) were healthy including no lifetime history of neurological and psychiatric disorders. Eighty healthy individuals met the above criteria and were included in the present study. Seventy of these individuals additionally provided salivary samples for the determination of CAR.

The included individuals had participated in neuroimaging studies at the Neurobiology Research Unit, Copenhagen University Hospital (Rigshospitalet) from 2012 to 2021. Written informed consent was obtained from all participants before their inclusion in the respective studies. The individual study protocols were approved by the Ethics Committee of Copenhagen and Frederiksberg, Denmark (protocol number: VEK H-4-2012-105, VEK (KF)01-2006-20, VEK H-15004506 and VEK H-16026898).

Procedures

Magnetic resonance imaging (MRI)

MRI scans were performed on a 3 T Verio scanner (n = 30) or a 3 T Prisma scanner (n = 50) using the following parameters: Verio/Prisma: repetition time = 1900 ms, echo time = 2.32/2.58 ms, inversion time = 900 ms, flip angle = 9°, in-plane matrix = 256 × 256, in-plane resolution = 0.9 × 0.9 mm, 224 slices and a slice thickness of 0.9 mm, no gap. The images were segmented using Voxel Based Morphometry (VBM8) implemented in Statistical Parametric Mapping (SPM8) (https://www.fil.ion.ucl.ac.uk/spm/software/spm12/)²² to mask grey matter in the subsequent extraction of tissue time-activity curves (TACs).

PET data acquisition

Procedures for PET acquisition have been described in detail elsewhere.²³ Briefly, scans were conducted using a high-resolution research tomography (HRRT) PET scanner with parameters: matrix = 256 × 256 × 207 voxels, voxel size = 1.22 × 1.22 × 1.22 mm. All participants underwent a transmission scan using the built-in source for attenuation and scatter correction followed by a 120 min emission scan which started at the time of the intravenous bolus injection of [¹¹C]Cimbi-36 (mean injected activity: 499 ± 109 MBq; injected mass of Cimbi-36: 0.72 ± 0.44 μg). [¹¹C]Cimbi-36 scan data were reconstructed into 45 dynamic frames (6 × 10 s, 6 × 20 s, 6 × 60 s, 8 × 120 s and 19 × 300 s) using the methods described previously.²⁴

PET processing and quantification

PET processing was performed as previously described.²³ Briefly, motion-corrected PET images were co-registered with the participants’ T1-weighted image using SPM8. Automatic delineation of regions of interest was performed using PVElab (MATrix LABoratory (MATLAB), Neurobiology Research Unit, Rigshospitalet, Copenhagen, Denmark).²⁵ A global neocortical region of interest (ROI) was defined as a volume-weighted mean of the following cortical ROIs: orbitofrontal cortex, superior, middle and inferior frontal gyrus; superior, middle and inferior temporal gyrus, sensorimotor cortex, parietal cortex and occipital cortex. A composite neocortex ROI was used because the signal is highly correlated between the neocortical regions.²⁶

Kinetic modelling using the simplified reference tissue model (SRTM)²⁷ with the cerebellum (excluding vermis) as the reference tissue was performed with in-house MATLAB scripts. The calculated non-displaceable binding potential (BP _ND) served as the outcome measure for quantifying 5-HT_2A binding. BP _ND is defined as:

$${BP}_{{\rm ND}} = \displaystyle{{V_T-V_{{\rm ND}}} \over {V_{{\rm ND}}}} $$

where V _T and V _ND are the distribution volumes in the ROI and reference region (cerebellum), respectively.

In addition to extracting mean 5-HT_2AR BP _ND values in the neocortex using the methods described above, we generated vertex-wise (surface) cortical 5-HT_2AR BP _ND maps to visualise the distribution of cortical 5-HT_2AR BP _ND associated with the inward facets of neuroticism. The vertex-wise cortical 5-HT_2AR BP _ND maps were generated using the PETSurfer²⁸ processing stream in FreeSurfer (version 7.3.2; see http://surfer.nmr.mgh.harvard.edu),²⁹ as described previously.³⁰ FastSurfer (version 2.0.1; see https://doi.org/10.1016/j.neuroimage.2020.117012) was used in place of FreeSurfer to process the structural images (recon-all). Briefly, PET–MRI co-registration was estimated using linear registration between the structural MRI and the time-weighted average of the dynamic PET images. Cortical vertex-wise TACs were obtained by resampling the dynamic PET data onto the common fsaverage surface space registration.³¹ Cortical TACs were surface smoothed by 10 mm full-width at half-maximum.

Kinetic modelling of the parametric BP _ND was performed using the Multilinear Reference Tissue Model 2 (MRTM2)³² with cerebellar grey matter as the reference region. Cerebellar grey matter masks were obtained by taking the intersection of the cerebellar grey matter segmentation defined by FastSurfer and a cerebellar grey matter segmentation, excluding vermis, derived using SUIT 3.1 (_python3, Brain and Mind Institute, University of Western Ontario, London, UK, and Ontario, Canada).³³ The reference region washout rate (k ₂′) was computed using MRTM³⁴ and a noise-free high-binding TAC was obtained from a surface-weighted average of the neocortical regions for this purpose.

Neuroticism

Participants completed the Danish version of the 240-item NEO-Personality Inventory-Revised (NEO-PI-R); this version has been normed in a sample of 600 individuals.³⁵ The NEO-PI-R is a self-reported personality questionnaire that evaluates the broad personality traits of neuroticism, extraversion, openness, agreeableness and conscientiousness. Each trait is made up of six constituent facets which for the neuroticism trait include anxiety, depression, self-consciousness, vulnerability, angry hostility and impulsiveness. Each facet score is derived by summing the scores of six constituent items. The outcome measures of neuroticism for this study were the facets of anxiety, depression, self-consciousness and vulnerability to stress, as these facets have previously been shown to be associated with 5-HT_2AR binding.¹¹

Cortisol awakening response

For the assessment of CAR, participants were instructed to collect saliva samples at home immediately after awakening and after 15, 30, 45 and 60 min. Saliva samples were collected in Salivette tubes (Sarstedt, Neubringen, Germany) by chewing a swab until it was completely saturated with saliva. Participants were instructed to avoid food intake, drinking, brushing their teeth and smoking during the first hour after waking up. If the first saliva sample was collected more than 10 min after awakening, the data-set was not included in our analyses. The participants noted whether the CAR samples were collected on a work day, study day or weekend/day off, and also noted the time of awakening and collection of the saliva sample. The saliva samples collected at home were stored in the refrigerator and returned to the laboratory the day after the completion of sampling, and if collected during the weekend, a maximum of 3 days after collection. The samples were subsequently stored at −80°C until analysis. Samples were shipped on dry ice to an external laboratory, and cortisol concentrations were analysed in four different batches across the collection period (i.e. 2012–2021) to minimise laboratory-dependent noise. The analysis of the first batch (37% of the total sample) was performed using the electroluminescence immunoassay (ECLIA) method on Modular Analytics E170 equipment (Roche, Mannheim, Germany), whereas the three remaining batches were analysed using the chemiluminescence immunoassay (CLIA) method on the IDS-iSYS automatic analyser (IDS plc, Boldon, UK). The intra- and inter-assay variations of both immunoassay methods were less than 15%. We also had internal saliva samples that were run with each batch for quality control.³⁶ The CAR measurements were computed as the area under the curve (AUC) with respect to the increase from baseline at waking (AUCi). The AUCi captures change over time and is thereby an index of reactivity of the HPA axis.³⁷ There was no significant difference in AUCi values among the collection days (P = 0.14).

Statistical analyses

Analyses of demographic, clinical, scanner and radiotracer data

Analyses were conducted in R (version 4.3.12022.2.0)³⁸ using RStudio. Group differences between males and females in continuous (age, weight-adjusted [¹¹C]Cimbi-36 injected mass (μg/kg)) and categorical variables (magnetic resonance scanner type) were evaluated using Mann–Whitney U-tests and Fisher's exact tests, respectively. All results were considered significant at P < 0.05.

Analysis of the association between neocortical 5-HT_2AR BP _ND and neuroticism

We applied a linear latent variable model (LVM) using the lava package (version 7.2.1)³⁹ implemented in R to first examine the association between 5-HT_2AR binding in the neocortex and the inward-directed facets of neuroticism in 80 healthy individuals. LVM is used to model the association between neuroticism and PET with a single parameter via a latent variable.

This latent variable, denoted Inward-Neuroticism_LV captures the shared signal between facets and can be seen as a weighted average of the facets where the weights are estimated via maximum likelihood. The PET outcome variable was log-transformed 5-HT_2AR BP _ND in the neocortex, and the neuroticism outcome variable was a latent variable (termed Inward-Neuroticism_LV) representing the inward-directed facets of neuroticism, namely anxiety, depression, self-consciousness and vulnerability. Inward-Neuroticism_LV was adjusted for age and gender, while neocortical 5-HT_2AR BP _ND values were adjusted for age, gender and magnetic resonance scanner type. The association between 5-HT_2AR BP _ND and the Inward-Neuroticism_LV is described in terms of regression coefficients (‘parameter estimate’). The 5-HT_2AR BP _ND vertex maps were input into the same LVM model to aid the visualisation of this association at each cortical vertex. The parameter estimate and P-value maps were created using the oro.nifti package⁴⁰ in R, and visualised using FreeSurfer.²⁹

A sensitivity analysis using structural equation modelling examined the association between 5-HT_2AR BP _ND and the inward facets of neuroticism when the latter was calculated as the sum of its constituent traits (rather than as a latent variable). As in the LVM, neuroticism was adjusted for age and gender and 5-HT_2AR BP _ND was adjusted for age, gender and magnetic resonance scanner type.

CAR as a moderator of the association between neocortical 5-HT_2AR BP _ND and neuroticism

In individuals with data on CAR, secondary analysis with LVM was performed to determine if cortisol dynamics, as measured with CAR, moderated the association between 5-HT_2AR BP _ND and Inward-Neuroticism_LV. To perform this moderation analysis with LVM, CAR values were split into three equally sized categories.

Gender as a moderator of the association between neocortical 5-HT_2AR BP _ND and neuroticism

Lastly, an exploratory analysis using an LVM investigated whether gender moderated the association between 5-HT_2AR BP _ND and Inward-Neuroticism_LV. Gender differences in neuroticism have been repeatedly observed, with females reporting higher levels of this trait relative to males.^41,42 Although 5HT_2AR binding has not consistently been shown to differ by gender,^43–45 5HT_2AR binding has been found to be positively associated with endogenous estradiol levels.⁴⁶ The evidence for the effect of gender on neuroticism and potentially on 5-HT_2AR BP _ND motivated this exploratory analysis.

For all the above-mentioned LVMs, the relevance of additional model paths was assessed using score tests (via the modelsearch function within the lava package). The path with the lowest P-value was added to the model if the adjusted P-value was below 0.05. A Bonferroni adjustment was used over all possible paths to control the family-wise error rate at 0.05. This procedure was repeated until no new model paths were supported.

Results

Demographic, clinical, scanner and radiotracer information

The sample included 80 individuals aged 18–58 years (median age: 27.3 years; inter-quartile range: 8.04 years). Our sample was well-balanced with 55% males and 45% females, and there were no significant differences in age (P = 0.20), weight-adjusted [¹¹C]Cimbi-36 injected mass (P = 0.38) or MR scanner type distribution (P = 0.50) between males and females. CAR AUCi was calculated for 70 of the 80 individuals included in the study (median AUCi: 58.84; inter-quartile range: 248.18). AUCi was not calculated for the remaining 10 individuals because of factors such as failure to provide saliva samples (n = 2), missing data points for the calculation of AUCi (n = 5), incorrect or missing collection time (n = 2) and inadequate saliva volume for AUCi calculations (n = 1). Additional demographics and radiotracer information for healthy individuals are detailed in Table 1.

Table 1 Demographic, clinical, scanner and radiotracer information

IQR, inter-quartile range; MRI, magnetic resonance imaging; CAR, cortisol awakening response; AUCi, area under the curve with respect to increase from baseline at awakening.

The estimate loadings (regression coefficients) of the latent variable (summarised by the Inward-Neuroticism_LV) onto its constituent facets ranged from 0.80 to 1.52, all P < 0.001, indicating that all facets significantly correlated with the latent variable. The constituent facet of depression loaded the most (β = 1.52) and self-consciousness the least (β = 0.80) on the Inward-Neuroticism_LV, as visualised in Fig. 1. The first LVM-based model showed a positive relation between neocortex 5-HT_2AR BP _ND and Inward-Neuroticism_LV across all participants (β = 0.01, 95% CI = 0.0005: 0.02, P = 0.04; Fig. 1). Figure 2 is a visualisation of this association using vertex-wise cortical 5-HT_2AR BP _ND maps. Additionally, the LVM showed a negative association between 5-HT_2AR BP _ND and age (β = −0.012, 95% CI = −0.008: −0.016, P < 0.001). The sensitivity analysis also showed a significant positive association between 5-HT_2AR BP _ND and the total inward neuroticism score (P = 0.04).

Figure 1 Latent variable model to estimate the association between 5-HT_2AR binding and the inward facets of neuroticism.

This figure shows the estimated latent variable model of the association between 5-HT_2AR binding and the inward facets of neuroticism. The ellipse represents the latent variable capturing Inward-Neuroticism_LV. Circular dashed lines represent error estimates included in the model. λ represents the loading of a given neuroticism sub-facet estimate onto its respective latent variable. Both Inward-Neuroticism_LV and neocortex BP _ND were adjusted for age and gender, and neocortex BP _ND was additionally adjusted for magnetic resonance (MR) scanner type. Self-cons, self-consciousness; BP _ND, non-displaceable binding potential.

Figure 2 Visualisation of the distribution of cortical 5-HT_2AR and associations with the inward facets of neuroticism. (a) The regional association between 5-HT_2AR binding with the latent estimates of inward neuroticism in the cortex. The maps are thresholded at P = 0.05. The colour-bar represents the range of estimate values. (b) The distribution of 5-HT_2AR in the cortex (reproduced with permission from Beliveau et al²⁸).

We did not find evidence that CAR moderated the association between 5-HT_2AR BP _ND and Inward-Neuroticism_LV (P = 0.33). There was also no evidence for a moderating effect of gender on the association between 5-HT_2AR BP _ND and Inward-Neuroticism_LV (P = 0.99).

Discussion

This study replicates previously reported findings of an association between brain 5-HT_2AR and the inward facets of neuroticism in healthy individuals. Using a latent variable model framework, we showed a positive association between neocortical 5-HT_2AR binding, measured using the [¹¹C]Cimbi-36 PET tracer, and the inward facets of neuroticism, namely vulnerability, depression, self-consciousness and anxiety. Furthermore, there was no evidence that this association was moderated by cortisol dynamics or gender.

Using a 5-HT_2AR PET antagonist tracer [¹⁸F]altanserin, a significant positive association between 5-HT_2AR binding and the inward facets of neuroticism was previously found in healthy individuals,¹⁰ and this association was stronger in individuals with a high familial risk for depression relative to those with low familial risk, suggesting a link to risk brain architecture.¹¹ In further support of this association, a longitudinal study in healthy individuals found that those with the phenotype for high inward neuroticism and high 5-HT_2AR binding have an elevated risk of developing depression in the future.⁴⁷ Treatment with antidepressant drugs downregulates 5-HT_2AR, suggesting that 5-HT_2A upregulation is a compensatory mechanism for serotonin deficiency.⁴⁸ To what extent genetic predisposition and environmental stress shape 5-HT_2AR density is not well understood. However, in light of the non-significant findings with stress response (CAR), we speculate that the association between 5-HT_2AR binding and the inward facets of neuroticism reflects a partly shared biological mechanism by which a genetic predisposition to 5-HT_2AR availability may contribute to increased neuroticism even in the absence of environmental stressors.⁴⁷ Our study findings lend support to the use of novel interventions such as psilocybin which actively targets the 5-HT_2AR and causes long-lasting changes in neuroticism¹⁰ for the treatment of depression.

Stress responses are intimately linked with serotonergic neurotransmission, and vulnerability to stress represents a feature of neuroticism. However, we did not detect a moderating effect of CAR on the association between serotonin function and neuroticism. This is the first study to examine the interplay between serotonin, neuroticism and CAR in the healthy brain, although previous work has examined the association of CAR separately with serotonin function and neuroticism. For instance, studies have previously found associations between CAR and several serotonergic markers, including negative associations of CAR with 5-HT₄R binding,⁴⁹ and positive associations of CAR with serotonin transporter binding.⁵⁰ In recent work, however, the association between CAR and serotonin transporter binding was not replicated.⁵¹ This may in part be because of differences in sample characteristics, as the latter study included predominantly young women relative to earlier studies,⁵² and calculated AUCi over 60 min after awakening relative to earlier studies that calculated AUCi over a 15 h time span.⁴⁹ Likewise, the relation of neuroticism with cortisol has also been previously reported,⁵³ although less consistently when neuroticism was studied in relation to measures of cortisol change (i.e. CAR).^19,53 To consider the role of differences in the quantification of cortisol response in the relation between neuroticism and cortisol, a study evaluated cortisol in three different ways (CAR, diurnal slope and total output) over multiple days, but did not find associations between cortisol dynamics and neuroticism for any of the measures.⁵² These studies highlight inconsistent findings on the relation of cortisol dynamics to 5-HT_2AR binding and neuroticism in healthy individuals.

In the present study, we found no evidence that gender moderated the association between 5-HT_2AR binding and the inward facets of neuroticism. However, we found a negative association between 5-HT_2AR binding and age, replicating previously reported finding with the [¹⁸F]altanserin tracer.^44,45,51 Evidence for gender differences in 5-HT_2AR binding in healthy individuals in the literature has been inconclusive with reports of no significant differences between the genders,^43–45 as well as higher binding in males relative to females.⁴³

Methodological inconsistencies, and, importantly, differences in tracers are often cited as reasons for the mixed findings in the PET literature,⁵⁴ and the replicability crisis in neuroscience is one of the main barriers to the translation of research findings into clinical practice. The current independent replication with the new agonist tracer [¹¹C]Cimbi-36 provides strong support to the link between 5-HT_2AR binding and inward-directed neuroticism in healthy adults and to interventions that target the 5-HT_2AR system, such as psilocybin, in reducing neuroticism and alleviating mood symptoms.¹⁰ The findings of the study, however, should be considered in the context of its limitations. 5-HT_2AR is most abundantly found in the neocortex where the signal with the applied PET method is also best captured, and therefore we limited our investigation to neocortical 5-HT_2AR binding and its association with the inward facets of neuroticism. We did not include individuals older than 60 years, therefore our results may not be generalisable to older populations. One of the reasons for the absence of a moderating effect of CAR on the association between 5-HT_2AR and neuroticism in our study could be because variations in HPA dynamics may be difficult to capture in a healthy sample. HPA axis dysfunction has been shown to be present in severe or acute major depressive episodes rather than mild or atypical depression.⁵² Hence, HPA dynamics may have a distinct impact in relation to environmental stress such as stressful life events or the onset of a mood disorder, but might not be evident in a healthy sample,⁵⁴ in which case we would not expect to be able to detect any association with underlying brain architecture. Thus, further studies on the influence of cortisol dynamics on 5-HT_2AR binding and neuroticism in a depressed sample, a high-risk sample or a sample that recently experienced stressful life events are warranted. The absence of a moderating effect of CAR on the association between 5-HT_2AR and neuroticism in our study could also be because of our choice of cortisol measure. We aimed to characterise the HPA axis function using the CAR AUCi. While this measure serves as a relevant composite indicator of cortisol responses and has been previously demonstrated to relate to serotonergic brain architecture,^49,50 it does not encompass the full spectrum of dynamic parameters of cortisol release. Alternative metrics, such as total CAR, mean cortisol and peak cortisol levels, were not included in our analysis.

In conclusion, we provide robust and consistent evidence for the link between 5-HT_2AR binding and the inward-directed personality facets of neuroticism. There was no evidence that cortisol dynamics moderated the association between 5-HT_2AR binding and inward-directed neuroticism in healthy individuals. Further, the association persists across gender. The confirmatory positive association between 5-HT_2AR binding and neuroticism prompts consideration of whether novel interventions such as psilocybin that actively targets 5-HT_2AR and causes changes in personality could be particularly beneficial if implemented as a targeted approach based on neuroticism profiles.