Introduction
Schizophrenia results from the complex interplay of genetic and environmental factors, which can impact the trajectories of brain development, particularly over the prenatal and perinatal stages, due to the important brain changes that occur during the intrauterine and the early postnatal period (Sasabayashi, Takahashi, Takayanagi, & Suzuki, Reference Sasabayashi, Takahashi, Takayanagi and Suzuki2021). Obstetric complications (OCs) have been associated with the neural network abnormalities observed in schizophrenia (Keshavan & Hogarty, Reference Keshavan and Hogarty1999). Exposure to OCs has been reported to influence the subsequent development of schizophrenia with an odds ratio between 1.17 and 3.5, and the largest effect sizes observed for polyhydramnios, premature rupture of membranes, low birth weight, and birth hypoxia (Davies et al., Reference Davies, Segre, Estradé, Radua, De Micheli, Provenzani and Fusar-Poli2020).
Reduced gray matter volume is a well-established feature of schizophrenia (De Peri et al., Reference De Peri, Crescini, Deste, Fusar-Poli, Sacchetti and Vita2012; Ellison-Wright, Glahn, Laird, Thelen, & Bullmore, Reference Ellison-Wright, Glahn, Laird, Thelen and Bullmore2008), especially in regions such as the frontal and temporal cortices, the cingulate, thalamus, and hippocampus (Matsuda & Ohi, Reference Matsuda and Ohi2018). However, measures of cortical morphology, such as surface area (SA), cortical thickness (CTh), and gyrification, exhibit distinct maturational trajectories (Smith et al., Reference Smith, Thornton, Lang, MacEwan, Kopala, Su and Honer2015). The latter refers to the development of brain surface folding patterns that, based on curvature and sulcal morphology, contribute to cortical SA (Fernández & Borrell, Reference Fernández and Borrell2023; Tonya White, Su, Schmidt, Kao, & Sapiro, Reference White, Su, Schmidt, Kao and Sapiro2010). Gyrification begins between 10 and 15 weeks of gestation and continues through the early postnatal period, remaining relatively stable thereafter (White et al., Reference White, Su, Schmidt, Kao and Sapiro2010). In contrast, maturational changes in SA and CTh exhibit a more protracted timeline, continuing throughout adolescence (Zilles, Palomero-Gallagher, & Amunts, Reference Zilles, Palomero-Gallagher and Amunts2013). Therefore, abnormalities in gyrification may reflect prenatal or perinatal adverse events (Papini et al., Reference Papini, Palaniyappan, Kroll, Froudist-Walsh, Murray and Nosarti2020; Smith et al., Reference Smith, Thornton, Lang, MacEwan, Kopala, Su and Honer2015) and could be used as markers of early neurodevelopment (Nelson et al., Reference Nelson, Kraguljac, White, Jindal, Shin and Lahti2020).
In healthy controls (HCs), cortical gyrification is related to general cognitive ability (Papini et al., Reference Papini, Palaniyappan, Kroll, Froudist-Walsh, Murray and Nosarti2020) and with language and working memory (WM) domains (Schmitt et al., Reference Schmitt, Ringwald, Meller, Stein, Brosch, Pfarr and Kircher2023). Exposure to obstetric risk factors, such as very preterm birth, impacts the Local Gyrification Index (LGI) and is related to poorer general cognitive ability, with cognitive deficits and higher rates of psychopathology (Papini et al., Reference Papini, Palaniyappan, Kroll, Froudist-Walsh, Murray and Nosarti2020; Schmitt et al., Reference Schmitt, Ringwald, Meller, Stein, Brosch, Pfarr and Kircher2023).
In schizophrenia, abnormal gyrification is considered an endophenotype reflecting aberrant connectivity (White & Gottesman, Reference White and Gottesman2012). Earlier sulci of the brain are thought to be correlated with genetic influences, while tertiary sulci are associated with nongenetic exposures to environmental risk factors (Sasabayashi, Takahashi, Takayanagi, & Suzuki, Reference Sasabayashi, Takahashi, Takayanagi and Suzuki2021). Abnormal gyrification has been described both in the at-risk mental state (ARMS) and first-episode psychosis (FEP), suggesting it could be an early imaging biomarker for psychosis (Matsuda & Ohi, Reference Matsuda and Ohi2018; Sasabayashi et al., Reference Sasabayashi, Takahashi, Takayanagi and Suzuki2021; White & Gottesman, Reference White and Gottesman2012).
In chronic schizophrenia, gyrification abnormalities were reported to be very heterogeneous (White & Gottesman, Reference White and Gottesman2012). Some studies have reported increased LGI, which refers to the ratio between the complete superficial contour and the outer contour of the cortex (Matsuda & Ohi, Reference Matsuda and Ohi2018) in regions such as the prefrontal cortex, insula, and temporo-parieto-occipital cortices (Sasabayashi et al., Reference Sasabayashi, Takahashi, Takayanagi and Suzuki2021; Spalthoff, Gaser, & Nenadić, Reference Spalthoff, Gaser and Nenadić2018; White & Gottesman, Reference White and Gottesman2012). In contrast, other studies have found decreased LGI (Matsuda & Ohi, Reference Matsuda and Ohi2018; Nelson et al., Reference Nelson, Kraguljac, White, Jindal, Shin and Lahti2020; Nesvåg et al., Reference Nesvåg, Schaer, Haukvik, Westlye, Rimol, Lange and Eliez2014; Palaniyappan & Liddle, Reference Palaniyappan and Liddle2012; Palaniyappan et al., Reference Palaniyappan, Marques, Taylor, Handley, Mondelli, Bonaccorso and Dazzan2013; Rychagov et al., Reference Rychagov, del Re, Zeng, Oykhman, Lizano, McDowell and Keshavan2024) predominantly in frontotemporal regions (Nanda et al., Reference Nanda, Tandon, Mathew, Giakoumatos, Abhishekh, Clementz and Keshavan2014; Palaniyappan, Mallikarjun, Joseph, White, & Liddle, Reference Palaniyappan, Mallikarjun, Joseph, White and Liddle2011; Sasabayashi et al., Reference Sasabayashi, Takahashi, Takayanagi and Suzuki2021), including the insula, superior temporal gyrus, Broca’s area, left precentral gyrus, right middle temporal gyrus, precuneus, and cingulate cortex (Madeira et al., Reference Madeira, Duarte, Martins, Costa, Macedo and Castelo-Branco2020; Matsuda & Ohi, Reference Matsuda and Ohi2018; Nanda et al., Reference Nanda, Tandon, Mathew, Giakoumatos, Abhishekh, Clementz and Keshavan2014; Nesvåg et al., Reference Nesvåg, Schaer, Haukvik, Westlye, Rimol, Lange and Eliez2014; Wheeler & Harper, Reference Wheeler and Harper2007). Inconsistencies were considered to reflect methodological differences, heterogeneity in demographic and clinical variables, or abnormalities being specific depending on the brain region examined (Si et al., Reference Si, Bi, Yu, See, Kelly, Ambrogi and Kempton2024).
Heterogeneity in gyrification abnormalities has also been reported in the early stages of psychosis, with several studies in individuals with FEP and ARMS reporting hypergyria in various brain regions (Narr et al., Reference Narr, Bilder, Kim, Thompson, Szeszko, Robinson and Toga2004; Sasabayashi et al., Reference Sasabayashi, Takahashi, Takayanagi and Suzuki2021) and also hypogyria in the frontotemporal and insular cortices (Harris et al., Reference Harris, Yates, Miller, Best, Johnstone and Lawrie2004; Palaniyappan et al., Reference Palaniyappan, Marques, Taylor, Handley, Mondelli, Bonaccorso and Dazzan2013). Zhou et al. (Reference Zhou, Wang, Wang, Xu, Cao and Zhang2021) observed a higher LGI in regions such as the left lateral occipital cortex, and lower LGI in other regions such as the left transverse temporal cortex in drug-naïve patients with FEP (Matsuda & Ohi, Reference Matsuda and Ohi2018; H. Zhou et al., Reference Zhou, Wang, Wang, Xu, Cao and Zhang2021). Sasabayashi et al. reported increased LGI in the left medial occipital cortex of individuals with ARMS who subsequently developed psychosis and considered it as a possible endophenotype of subsequent transition (Sasabayashi et al., Reference Sasabayashi, Takayanagi, Takahashi, Koike, Yamasue, Katagiri and Suzuki2017).
In schizophrenia, several brain abnormalities have been related to OCs exposure (Costas-Carrera, Garcia-Rizo, Bitanihirwe, & Penadés, Reference Costas-Carrera, Garcia-Rizo, Bitanihirwe and Penadés2020), including reduced cortical volume (Cannon et al., Reference Cannon, van Erp, Rosso, Huttunen, Lönnqvist, Pirkola and Standertskjöld-Nordenstam2002; Neilson et al., Reference Neilson, Bois, Clarke, Hall, Johnstone, Owens and Lawrie2018; Smith et al., Reference Smith, Thornton, Lang, MacEwan, Kopala, Su and Honer2015), reduced hippocampal volume (McNeil, Cantor-Graae, & Weinberger, Reference McNeil, Cantor-Graae and Weinberger2000; Schulze et al., Reference Schulze, McDonald, Frangou, Sham, Grech, Toulopoulou and Murray2003; Stefanis et al., Reference Stefanis, Frangou, Yakeley, Sharma, O’Connell, Morgan and Murray1999), and an increased ventricle–brain ratio (Cannon et al., Reference Cannon, van Erp, Rosso, Huttunen, Lönnqvist, Pirkola and Standertskjöld-Nordenstam2002; Costas-Carrera et al., Reference Costas-Carrera, Verdolini, Garcia-Rizo, Mezquida, Janssen, Valli and Bernardo2024; McNeil et al., Reference McNeil, Cantor-Graae and Weinberger2000; Neilson et al., Reference Neilson, Bois, Clarke, Hall, Johnstone, Owens and Lawrie2018). Only three studies analyzed the relationship between OCs and gyrification in patients with schizophrenia (Falkai et al., Reference Falkai, Honer, Kamer, Dustert, Vogeley, Schneider-Axmann and Tepest2007; Haukvik et al., Reference Haukvik, Schaer, Nesvåg, McNeil, Hartberg, Jönsson and Agartz2012; Smith et al., Reference Smith, Thornton, Lang, MacEwan, Kopala, Su and Honer2015). Falkai et al. observed no effect of OC exposure on gyrification in schizophrenia when using a two-dimensional method, whereas Haukvik et al. (Reference Haukvik, Schaer, Nesvåg, McNeil, Hartberg, Jönsson and Agartz2012) and Smith et al. (Reference Smith, Thornton, Lang, MacEwan, Kopala, Su and Honer2015) reported a relationship between OCs and hypogyria . However, only Smith et al. (Reference Smith, Thornton, Lang, MacEwan, Kopala, Su and Honer2015) observed that the relationship between OCs and hypogyria was specific to patients with FEP and not in HCs (Smith et al., Reference Smith, Thornton, Lang, MacEwan, Kopala, Su and Honer2015).
Gyrification has been associated with cognition in both HCs (Gautam, Anstey, Wen, Sachdev, & Cherbuin, Reference Gautam, Anstey, Wen, Sachdev and Cherbuin2015; Green et al., Reference Green, Blackmon, Thesen, DuBois, Wang, Halgren and Devinsky2018; Schmitt et al., Reference Schmitt, Ringwald, Meller, Stein, Brosch, Pfarr and Kircher2023) and patients diagnosed with schizophrenia (Sasabayashi et al., Reference Sasabayashi, Takayanagi, Takahashi, Koike, Yamasue, Katagiri and Suzuki2017). Gautam et al. observed that greater gyrification of the lateral frontal cortex was associated with better performance in terms of executive function in HCs (Gautam et al., Reference Gautam, Anstey, Wen, Sachdev and Cherbuin2015), whereas Green et al. observed a positive relationship between measures in the parietal–frontal regions and WM in HCs (Green et al., Reference Green, Blackmon, Thesen, DuBois, Wang, Halgren and Devinsky2018). In patients with schizophrenia, greater gyrification of the right frontal cortex was associated with worse executive function (Sasabayashi et al., Reference Sasabayashi, Takahashi, Takayanagi and Suzuki2021), supporting an aberrant neurodevelopment in schizophrenia. Fetal growth is also correlated with cognition, and variables such as birth weight and gestational age are considered markers of the intrauterine environment (Schmitt et al., Reference Schmitt, Ringwald, Meller, Stein, Brosch, Pfarr and Kircher2023). Gestational age moderates the association between gyrification in regions such as the left supramarginal gyrus and the left superior frontal gyrus with attention/WM (Schmitt et al., Reference Schmitt, Ringwald, Meller, Stein, Brosch, Pfarr and Kircher2023). In adults born premature, increased frontal–temporal–parietal gyrification was related to worse cognitive performance (Hedderich et al., Reference Hedderich, Bäuml, Berndt, Menegaux, Scheef, Daamen and Sorg2019). Papini et al. reported widespread hypogyria in adults born preterm, especially in the middle and inferior frontal, superior temporal, and medial occipito-parietal regions, and found a relationship between LGI and cognition, with the spatial distribution of these associations substantially differing between preterm-born individuals and HCs (Papini et al., Reference Papini, Palaniyappan, Kroll, Froudist-Walsh, Murray and Nosarti2020).
OCs exert a differential impact on fetal outcomes regarding sex, suggesting a sexual dimorphism effect, which implies different outcomes regarding the timeframe of the event (Yu, Chen, Ge, & Wang, Reference Yu, Chen, Ge and Wang2021). This effect is particularly relevant in schizophrenia, as sex-related brain structural and cognitive differences have been described (Mendrek & Mancini-Marïe, Reference Mendrek and Mancini-Marïe2016).
The assessment of FEP patients provides valuable insights into how adverse prenatal and perinatal factors may contribute to neurodevelopmental vulnerabilities that increase the risk of psychosis without the potential confounding effect of protracted illness and medication exposure. We, therefore, sought to examine the impact of OCs on gyrification in FEP patients and their relationship with cognitive performance.
We hypothesized that there would be differences in gyrification between FEP patients and HCs, with a lower LGI in FEP patients, and that these differences would be related to OC exposure. We also hypothesized that cortical folding patterns would play a role in the relationship between OCs and cognition, and examined the potential effect of sex and the timeframe of OC exposure.
Methods and materials
This research was conducted as part of a multicenter longitudinal study examining gene–environment interactions on the pathway to psychosis (the PEPs study, ‘Phenotype–genotype and environmental interaction: Application of a predictive model in first psychotic episodes’, Bernardo et al. (Reference Bernardo, Bioque, Parellada, Saiz Ruiz, Cuesta, Llerena and Cabrera2013)).
Participants
The sample of the PEPs study included 335 FEP patients and 253 HCs recruited between January 2009 and December 2011. The inclusion criteria and characteristics of the study have been previously described in detail (Salagre et al., Reference Salagre, Arango, Artigas, Ayuso-Mateos, Bernardo, Castro-Fornieles and Vieta2019). Briefly, subjects with FEP aged 7–35 years, presenting psychotic symptoms for <12 months, were recruited from the inpatient and outpatient units of 16 participating Spanish centers, 14 of which are members of the Center of Biomedical Research Network on Mental Health (CIBERSAM). HCs were recruited at each site through advertisements and matched with patients by age (within ±10%), sex, and parental socioeconomic status, as measured by the Hollingshead–Redlich scale (within ±1 level). For the neuroimaging component of the study, a maximum of 6 months was established from inclusion to scan time. All centers received the approval of their respective Independent Ethics Committee. All procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Declaration of Helsinki of 1975, as revised in 2008. Written informed consent was obtained from all participants before they participated in the study and from parents/legal guardians for children under 16 years of age (children gave assent). In the present study, from the total sample, we included 139 individuals with FEP and 125 HCs based on the availability of data for magnetic resonance imaging, cognitive assessments, and OC exposure.
History of OC assessment
OCs were assessed using the Lewis–Murray scale through a family interview (Lewis & Murray, Reference Lewis and Murray1987). The scale groups OCs into three categories, A, B, and C, according to the type of complication defined as follows (M. Cannon, Jones, & Murray, Reference Cannon, Jones and Murray2002; Mezquida et al., Reference Mezquida, Fernandez-Egea, Treen, Mané, Bergé, Savulich and Garcia-Rizo2018):
-
A. Complications of pregnancy: syphilis or rubella, rhesus isoimmunization/Rh incompatibility, severe preeclampsia requiring hospitalization or induction of labor, and bleeding before delivery or threatened abortion.
-
B. Abnormal fetal growth and development: twin delivery, preterm birth before 37 weeks, or long-term after 42 weeks, weight at birth <2500 g, and any important physical abnormality.
-
C. Difficulties in delivery: premature rupture of membranes, duration of delivery more than 36 h or less than 3 h, umbilical cord prolapse, complicated cesarean delivery, abnormal fetal presentation, use of forceps, and being in an incubator for more than 4 weeks.
Participants were stratified based on exposure, which the Lewis–Murray scale classifies into definite and dubious based on the quality of the information. We included only definite events (yes/no) grouped as intrauterine (Lewis A and B) and delivery complications (Lewis C), as well as the total score (Lewis T).
Image acquisition and processing
After all DICOM images were converted to NIFTI format using the dcm2niix function from MRIcron software, the FreeSurfer analysis package (v7.1.1, https://surfer.nmr.mgh.harvard.edu/) was used to generate measurements of CTh, cortical and subcortical volumes, and LGI. The standard FreeSurfer processing pipeline was employed, which follows the workflow: motion and bias field correction, skull extraction, affine and nonlinear alignment to the Talairach atlas, subcortical division, and cortical segmentation using the Desikan–Killiany atlas. For quality assurance, a visual inspection of the segmentation was performed by a technician specialized in neuroimaging, following the quality control protocol 2.0 of the ENIGMA consortium (https://enigma.ini.usc.edu/protocols/imaging-protocols).
The LGI was calculated using FreeSurfer with default parameters (vertex-wise values were averaged for each lobe [frontal, parietal, temporal, and occipital] and area [cingulate cortex], considering the right and left hemispheres separately, Kernel size 10 mm; Madan & Kensinger (Reference Madan and Kensinger2017)).
In this multicenter study, data were collected from six distinct neuroimaging centers using different scanners (Siemens Magnetom Trio Tim 3T, Siemens Symphony 1.5T, Philips Achieva 3T, Philips Intera 1.5T, GE Signa Horizon MX 1.5T, and GE Signa Excite 1.5T). Detailed information on the acquisition parameters from each participating platform can be found in previous work from the collaborative study (Pina-Camacho et al., Reference Pina-Camacho, Del Rey-Mejías, Janssen, Bioque, González-Pinto, Arango and Parellada2016). To adjust for site, we employed the ComBat batch harmonization method (Fortin et al., Reference Fortin, Cullen, Sheline, Taylor, Aselcioglu, Cook and Shinohara2018).
Cognitive assessment
In the PEPs study, the cognitive assessment at baseline was performed in the second month after inclusion to ensure the clinical stability of patients after the FEP (Bernardo et al., Reference Bernardo, Bioque, Parellada, Saiz Ruiz, Cuesta, Llerena and Cabrera2013; Cuesta et al., Reference Cuesta, Sánchez-Torres, Cabrera, Bioque, Merchán-Naranjo, Corripio and Bernardo2015). For our hypothesis, we focused on verbal memory (VM) and WM, two cognitive domains that we had previously observed to be associated with OC exposure in schizophrenia (Amoretti, Rabelo-da-Ponte, et al., Reference Amoretti, Rabelo-da-Ponte, Garriga, Forte, Penadés, Vieta and Garcia-Rizo2022).
VM was assessed using the Verbal Learning Test España-Complutense (Benedet, Reference Benedet and Madrid1998), while WM was evaluated using the Digit Span and Letter-Number Sequencing subtests of the WAIS-III (Wechsler, Reference Wechsler2019). These cognitive domains were derived from our previous work using principal component analysis (Amoretti, Rosa, et al., Reference Amoretti, Rosa, Mezquida, Cabrera, Ribeiro, Molina and Bernardo2022), with higher scores reflecting better performance in both domains. Further details are described in a previous work (Cuesta et al., Reference Cuesta, Sánchez-Torres, Cabrera, Bioque, Merchán-Naranjo, Corripio and Bernardo2015).
Statistical analysis
Descriptive statistics were calculated for each sociodemographic, neuropsychological, and clinical variable. Continuous variables are presented as mean value ± standard deviation and compared using Student’s t-tests. Categorical variables are expressed as total numbers (or percentages) and compared between groups using χ 2-tests. OCs were reported as dichotomous variables (yes/no).
Student’s t-tests were used to examine differences in LGI between FEP patients and HCs. General linear model (GLM) analyses were performed to test the relationship between each independent variable and the LGI of each brain area. Independent variables were sex, age, and chlorpromazine equivalent antipsychotic dose (Gardner, Murphy, O’Donnell, Centorrino, & Baldessarini, Reference Gardner, Murphy, O’Donnell, Centorrino and Baldessarini2010), diagnostic group (FEP/HCs), OCs (presence/absence), and the interaction between OCs and diagnostic group. We report results before and after false discovery rate (FDR) correction for multiple comparisons.
For cognitive measures, parametric mediation analyses were conducted in both groups (HCs and patients with FEP) to test whether gyrification mediates the effect of OCs on cognitive outcomes. We employed the LGI of areas where we had observed a significant interaction between diagnosis and OCs as the mediation variable between the independent variable (OCs) and cognitive outcomes, including WM and VM. The estimation of average causal mediation effects, average direct effects, total effects, and proportion-mediated effects was computed for each of the 1,000 bootstrapped resamples, and the 95% confidence interval (95% CI) was computed by determining the 2.5th and 97.5th percentiles for the resamples. The mediation analysis was performed with the mediation package version 4.5.03 (Tingley, Yamamoto, Hirose, Keele, & Imai, Reference Tingley, Yamamoto, Hirose, Keele and Imai2014).
Statistical analyses were performed with Statistical Package for the Social Sciences (Version 25) and R version 4.5.0.
Results
Sociodemographic and clinical variables
The sociodemographic characteristics of the sample are described in Table 1. There were no significant between-group differences in sex or age. However, as expected, the groups differed significantly in terms of educational level (χ 2 33.32, p < 0.001), with higher levels in HCs than in patients. There was a trend toward significance in the prevalence of OCs (Lewis total score) (χ 2 = 2.91, p = 0.089) between FEP and HCs. The median duration of untreated psychosis for the FEP patients was 45 days, and the daily equivalent dose of chlorpromazine was 450 (mean value: 557.62).
Table 1. Clinical and sociodemographic characteristics of the sample
Lewis–Murray total score (any difficulty during delivery and pregnancy).
Lewis–Murray A: Complications of pregnancy; Lewis–Murray B: Abnormal fetal growth and development; Lewis–Murray C: Difficulties in delivery.
Differences in LGI between FEP and HC
We observed significant differences in LGI between FEP patients and HCs in the left cingulate (Wald’s χ 2 = 4.24, p = 0.04), left occipital (Wald’s χ 2 = 6.47, p = 0.01), left parietal (Wald’s χ 2 = 3.93, p = 0.05), right cingulate (Wald’s χ 2 = 3.77, p = 0.05), and right occipital (Wald’s χ 2 = 5.95, p = 0.01) cortices (Table 2).
Table 2. Generalized linear model gyrification index lobes by diagnosis (psychosis/control) and stratified by the presence/absence of difficulties during the intrauterine/delivery period (OC presence/absence)
FDR, false discovery rate; OCs, obstetric complications (during the intrauterine period and delivery; LGI, Local Gyrification Index.
The effect of the estimates (Wald’s χ 2 and p-value) for diagnosis and OCs refers to the outcome of the regression analysis without interaction.
The effect of estimates (Wald’s χ 2 and p-value) for diagnosis by OCs refers to the outcome of the regression analysis where the interaction was included in addition to the other independent variables. Estimated marginal means of predicted gyrification index values are adjusted by covariates, age, sex, and chlorpromazine equivalent mean dose.
*p<.05; ^p=0.05-.10
Differences in LGI between subjects with and without OCs
Within the whole sample, there were significant differences between subjects with and without OCs in the left frontal (t = −2.18, p = 0.03); left cingulate (t = −2.03, p = 0.04); and trend level differences in the right cingulate (t = −1.9, p = 0.06) cortices (data not shown). However, when we adjusted for covariates (age, sex, and chlorpromazine equivalent mean dose), none of these differences maintained statistical significance (Table 2).
Association between LGI, diagnosis, and OCs
A GLM was applied to analyze whether OC exposure (Lewis total score) interacted with diagnosis to predict differences in LGI. We observed a significant interaction between OCs and diagnosis in the left cingulate cortex (LCC) (Wald’s χ 2 = 6.65, p = 0.01), and trend level differences in the right parietal cortex (Wald’s χ 2 = 2.98, p = 0.08). However, there was no significant effect on LGI after adjusting for multiple comparisons.
In the LCC, FEP patients with OCs displayed the lowest LGI (2.28 ± 0.02), followed by FEP patients without OCs (2.30 ± 0.01), HCs without OCs (2.32 ± 0.01), and finally, HCs with OCs (2.41 ± 0.03) (Table 2 and Figure 1).
Figure 1. General linear model mean predicted estimated value for Local Gyrification Index for the left cingulate area covariated by sex, age, chlorpromazine equivalent dose, diagnostic group (FEP patients/HCs), OCs (presence/absence), and the interaction between OCs and the diagnostic group. Note: FEP, first episode psychosis; HC, healthy controls; OCs, obstetric complications.
We then separately analyzed the role of gestational (Lewis A + B) and delivery complications (Lewis C) on LGI and observed that the interaction we had found in LCC was specific to those participants who had suffered complications during gestation (Wald’s χ 2 = 5.31, p = 0.02) rather than delivery (Wald’s χ 2 = 3.231, p = 0.072).
Finally, we examined whether sex influenced this interaction effect on LGI in the LCC. In the MLG model, sex had a significant effect on LGI (Wald’s χ 2 = 14.89, p < 0.001) and the interaction of Lewis T × diagnosis × sex was significant (Wald’s χ 2 = 9.70, p = 0.021). Therefore, we stratified the sample according to sex and observed that the effect of Lewis T significantly interacted with diagnosis only in males (Wald’s χ 2 = 8.12, p = 0.004) and remained significant after FDR correction (p = 0.04) (Supplementary Table 2A). In this male subsample, FEP with OCs displayed the lowest LGI (2.33 ± 0.03), followed by FEP without OCs (2.31 ± 0.03), HCs without OCs (2.36 ± 0.02), and finally, HCs with OCs (2.50 ± 0.04). This effect was not observed in the female group (Wald’s χ 2 = 0.253; p = 0.615) (Supplementary Table 2B). However, the female sex was underrepresented, with only 17 women with OC history.
Association between gyrification, OCs, and cognition
To analyze the relationship between OCs, cognition, and gyrification, we focused on the LCC, where we observed a significant interaction between diagnosis and OCs.
We tested a mediation model, with OCs as the independent variable, cognitive functioning (independently examining VM and WM) as the dependent variable, and the LGI in the LCC as the mediator. We stratified the sample between patients and HCs, as previous studies have shown different profiles.
In HCs (Supplementary Figure S2), the total effect of OCs (Lewis T) on WM was not significant (β = −5.46, 95% CI: −13.45 to 1.86; p = 0.14), nor was the average causal mediation effect (ACME) (β = 1.68, 95% CI: −0.11 to 4.32; p = 0.10), whereas the average direct impact of OCs on WM (ADE) (β = −7.14, 95% CI: −15.16 to −0.26; p = 0.048*) was significant. On the contrary, no significant effects were found for VM: total effect (β = 10.92, 95% CI: −10.94 to 28.81; p = 0.32), ACME (β = −1.79, 95% CI: −6.20 to 2.40; p = 0.38), or ADE (β = 12.71, 95% CI: −9.26 to 30.41; p = 0.29).
In FEP patients (Supplementary Figure 3), none of the effects were significant for WM: total effect (β = −0.66, 95% CI: −7.26 to 6.54; p = 0.88), ACME (β = −0.18, 95% CI: −1.32 to 0.41; p = 0.63), and ADE (β = −0.48, 95% CI: −7.11 to 6.77; p = 0.92). Similarly, no significant effects were found for VM: total effect (β = −10.92, 95% CI: −40.16 to 16.41; p = 0.40), ACME (β = −0.27, 95% CI: −5.80 to 4.31; p = 0.92), or ADE (β = −10.65, 95% CI: −39.99 to 16.98; p = 0.41).
When mediation analyses were stratified by sex (Supplementary Figures 4 and 5), none of the effects were significant for WM in HC males: total effect (β = −2.83, 95% CI: −15.49 to 6.84; p = 0.66), ACME (β = 2.75, 95% CI: −0.28 to 8.41; p = 0.10), and ADE (β = −5.58, 95% CI: −19.25 to 4.48; p = 0.38), while in HC females, a trend was seen in the total effect (β = −7.36, 95% CI: −15.35 to 0.54; p = 0.07). However, the ACME (β = −0.28, 95% CI: −2.69 to 1.44; p = 0.75) and ADE (β = −7-08, 95% CI: −15.31 to 1.24; p = 0.11) in HC females were not significant.
In FEP patients, none of the effects were significant for WM in males: ACME (β = −0.11, 95% CI: −1.10 to 0.80; p = 0.81), ADE (β = 4.15, 95% CI: −3.34 to 12.03; p = 0.32), and total effect (β = 4.04, 95% CI: −3.47 to 11.85; p = 0.33). On the other hand, in female FEP patients, the total effect (β = −12.30, 95% CI: −21.57 to −2.15; p = 0.03) and ADE (β = −11.45, 95% CI: −19.92 to −2.15; p = 0.03) were significant, while the ACME (β = −0.85, 95% CI: −4.95 to −3.32; p = 0.62) was not significant.
Discussion
This study examined the relationship between exposure to OCs, cortical folding, and cognitive function.
We observed significant differences in LGI measures between FEP and HC participants in the left parietal, bilateral cingulate, and bilateral occipital cortices. These results are consistent with studies that reported hypogyria in patients with psychotic disorders compared to controls (Nanda et al., Reference Nanda, Tandon, Mathew, Giakoumatos, Abhishekh, Clementz and Keshavan2014; Nesvåg et al., Reference Nesvåg, Schaer, Haukvik, Westlye, Rimol, Lange and Eliez2014), also in the early stages of the illness (Palaniyappan et al., Reference Palaniyappan, Marques, Taylor, Handley, Mondelli, Bonaccorso and Dazzan2013).
In terms of exposure to OCs, regardless of diagnosis, we observed no significant gyrification differences between subjects exposed to OCs and those who were not exposed.
Results of previous studies are inconsistent, with some reporting a relationship between exposure to OCs and a reduction in gyrification (Engelhardt et al., Reference Engelhardt, Inder, Alexopoulos, Dierker, Hill, Van Essen and Neil2015; Haukvik et al., Reference Haukvik, Schaer, Nesvåg, McNeil, Hartberg, Jönsson and Agartz2012; Smith et al., Reference Smith, Thornton, Lang, MacEwan, Kopala, Su and Honer2015; Yehuda et al., Reference Yehuda, Rabinowich, Zilberman, Wexler, Haratz, Miller and Bashat2024) and others describing increased gyrification (Hedderich et al., Reference Hedderich, Bäuml, Berndt, Menegaux, Scheef, Daamen and Sorg2019; Wu, De Asis-Cruz, & Limperopoulos, Reference Wu, De Asis-Cruz and Limperopoulos2024). Such differences may be related to the heterogeneity of OCs (hypoxia, premature birth, fetal growth restriction, etc.).
Therefore, we examined whether prenatal and perinatal complications interacted differently with diagnosis to determine differences in LGI. We observed that a significant interaction with diagnosis to determine LGI differences in the LCC specifically pertained to antepartum complications. This result did not survive correction for multiple comparisons within the entire group but it maintained significance within the male subsample when we stratified our participants based on sex. In this region, male patients exposed to OCs displayed the lowest LGI, while in male HC participants, OCs acted inversely, and exposure was associated with the highest LGI. This finding may reflect how early environmental stressors affect patients and controls differently, suggesting that mere exposure to OCs may not be sufficient to explain the changes observed in psychosis, in keeping with the gene–environment interaction hypothesis. Previous research also reported that exposure to OCs can affect patients with schizophrenia differently compared to HC and specifically identified the smallest hippocampal volumes in patients with OCs (Stefanis et al., Reference Stefanis, Frangou, Yakeley, Sharma, O’Connell, Morgan and Murray1999). Similarly, Cannon et al. observed that OCs interacted with schizophrenia diagnosis to determine brain volumetric abnormalities such as ventricular enlargement (Cannon et al., Reference Cannon, van Erp, Rosso, Huttunen, Lönnqvist, Pirkola and Standertskjöld-Nordenstam2002).
Ducsay et al. (Reference Ducsay, Goyal, Pearce, Wilson, Hu and Zhang2018) suggested that hypoxia, a common mechanism of most OCs, is implicated in the disruption of neural plasticity through epigenomic modifications. These alterations in the epigenetic landscape are believed to underlie phenotypic programming mechanisms that influence long-term health outcomes and disease susceptibility (Ducsay et al., Reference Ducsay, Goyal, Pearce, Wilson, Hu and Zhang2018). Furthermore, evidence from genome-wide association studies indicates that numerous genes implicated in schizophrenia risk are not only active during fetal neurodevelopment (Hall & Bray, Reference Hall and Bray2022) but are also regulated by hypoxic conditions (Semenza, Reference Semenza2001). In our sample, we observed a specific effect of antepartum complications, which we had previously observed to be significantly associated with case–control status (Valli et al., Reference Valli, Gonzalez Segura, Verdolini, Garcia‐Rizo, Berge, Baeza and Vieta2023) and are often characterized by protracted hypoxia-associated placental pathology (Valli & McGuire, Reference Valli and McGuire2023).
Collectively, these findings support the hypothesis that exposure to OCs may induce maladaptive epigenetic changes in individuals with a genetic predisposition, thereby altering neurodevelopmental trajectories and heightening vulnerability to subsequent environmental risk factors. On the contrary, the increased LGI in the LCC that we observed in HC exposed to OCs might represent an adaptive mechanism afforded to those at low genetic risk.
Furthermore, within our sample, male participants appeared to be more susceptible to the effects of OCs, as evidenced by a more pronounced reduction in LGI of the LCC in male FEP patients compared to other groups. This finding is consistent with previous studies demonstrating sex-specific differences in brain gyrification patterns in schizophrenia. For instance, one study reported differential gyrification alterations based on sex (Mancini-Marïe et al., Reference Mancini-Marïe, Yoon, Jiminez, Fahim, Potvin, Grant and Mendrek2018), while another found that hypogyrification occurs more frequently in males with schizophrenia than in females (Vogeley et al., Reference Vogeley, Schneider-Axmann, Pfeiffer, Tepest, Bayer, Bogerts and Falkai2000). In addition, studies focusing on individuals exposed to OCs have also identified sex-dependent effects, with male patients exhibiting greater LGI reductions across several brain regions (Mareckova, Miles, Andryskova, Brazdil, & Nikolova, Reference Mareckova, Miles, Andryskova, Brazdil and Nikolova2020; Papini et al., Reference Papini, Palaniyappan, Kroll, Froudist-Walsh, Murray and Nosarti2020). Overall, males seem to be more vulnerable to intrauterine insults, as also suggested by the differential placental upregulation of genes involved in schizophrenia between males and females exposed to obstetric risk (Sutherland & Brunwasser, Reference Sutherland and Brunwasser2018; Ursini et al., Reference Ursini, Punzi, Chen, Marenco, Robinson, Porcelli and Weinberger2018).
These findings emphasize the importance of considering sex-specific neurodevelopmental trajectories in understanding the impact of OCs on cortical morphology in schizophrenia.
The left cingulate area is involved in affect, attention, memory, and executive function (Haznedar et al., Reference Haznedar, Buchsbaum, Hazlett, Shihabuddin, New and Siever2004) and is often reported to be altered in schizophrenia (Bersani et al., Reference Bersani, Minichino, Fojanesi, Gallo, Maglio, Valeriani and Fitzgerald2014; Feng et al., Reference Feng, Palaniyappan, Robbins, Cao, Fang, Luo and Luo2023). In this region, we observed that FEP with OCs displayed the lowest LGI. The neurodevelopmental model of psychosis (Murray & Lewis, Reference Murray and Lewis1987) postulates that genetic predisposition combined with environmental insults may be involved in the aberrant brain development observed in psychosis. In utero insults were identified as key disruptors of neurodevelopmental processes (Keshavan & Hogarty, Reference Keshavan and Hogarty1999). In particular, cingulate dysfunction has been suggested to disrupt the modulation of prefrontal–temporal integration (Dazzan et al., Reference Dazzan, Lawrence, Reinders, Egerton, van Haren, Merritt and McGuire2021; Nanda et al., Reference Nanda, Tandon, Mathew, Giakoumatos, Abhishekh, Clementz and Keshavan2014) with an effect on cognitive functions such as memory and executive function (Fletcher, McKenna, Friston, Frith, & Dolan, Reference Fletcher, McKenna, Friston, Frith and Dolan1999; L. Zhou et al., Reference Zhou, Pu, Wang, Liu, Wu, Liu and Liu2016), and a relationship to poorer WM in individuals with FEP (Feng et al., Reference Feng, Palaniyappan, Robbins, Cao, Fang, Luo and Luo2023; L. Zhou et al., Reference Zhou, Pu, Wang, Liu, Wu, Liu and Liu2016).
In our sample, FEP patients performed significantly worse than HC participants in both the cognitive domains examined (WM and VM). This is consistent with previous studies reporting these deficits from the early phases of the disorder (Catalan et al., Reference Catalan, McCutcheon, Aymerich, Pedruzo, Radua, Rodríguez and Fusar-Poli2024; Rodriguez-Jimenez et al., Reference Rodriguez-Jimenez, Santos, Dompablo, Santabárbara, Aparicio, Olmos and García-Fernández2019).
Finally, we examined whether the LGI of the LCC mediated the relationship between OCs and cognitive functioning. In our FEP sample, we did not observe a significant mediation effect for either WM or VM. Meanwhile, a significant direct effect of OCs on WM was observed in HCs. Similarly, the mediation effect of the LCC LGI showed a positive indirect effect on cognitive performance. Although it could be considered a trend toward significance, there is no strong evidence that LCC LGI explains the relationship between OCs and WM. A possible explanation is that cognitive performance in FEP patients depends on a more complex interaction between genetic and environmental risk factors, with several environmental exposures involved in shaping the relationship between OCs and cognition. For instance, a family history of psychosis confers a higher risk for developing psychosis than OCs (Davies et al., Reference Davies, Segre, Estradé, Radua, De Micheli, Provenzani and Fusar-Poli2020) and can contribute to brain changes observed in patients with psychosis or even high-risk populations (Neilson et al., Reference Neilson, Bois, Clarke, Hall, Johnstone, Owens and Lawrie2018). In addition, other environmental risk factors, such as childhood maltreatment (Sideli et al., Reference Sideli, Schimmenti, La Barbera, La Cascia, Ferraro, Aas and van der Ven2022) and cannabis use (Yucel et al., Reference Yucel, Bora, Lubman, Solowij, Brewer, Cotton and Pantelis2012), are related to cognitive impairment in FEP patients and may attenuate the isolated effect of OC exposure on cognition.
When stratifying our sample by sex, we observed no significant effect of OCs in relation to cognition in males, while in females, a direct effect of OCs on WM was observed in FEP patients. These differential effects of OCs on cognition by sex were unexpected, especially considering that previous evidence suggests that the male fetus may be more vulnerable to the effects of OCs than the female fetus, and that males who go on to develop psychosis may possess fewer compensatory resources compared to other groups (DiPietro & Voegtline, Reference DiPietro and Voegtline2017; Sutherland & Brunwasser, Reference Sutherland and Brunwasser2018). One possible explanation for this unexpected pattern lies in the underrepresentation of women with a history of OCs within our sample. This limitation increases the likelihood of unstable parameter estimates, inflated effect sizes, and spurious significance, especially in stratified models. Thus, this effect should be interpreted with caution, and future studies with larger and sex-balanced samples are necessary to clarify the role of sex in moderating the relationship between early adversity and cognition in psychosis.
This research should be interpreted in the context of several other limitations. First, OCs have a small effect size to explain psychosis, so larger sample sizes are needed to increase the statistical power and detect significant effects (De Prisco & Vieta, Reference De Prisco and Vieta2024). In addition, this was a cross-sectional study conducted at the onset of psychosis (despite patients being minimally treated, they were not antipsychotic-naïve), so we were not able to control other factors, including treatment adherence (Vieta & De Prisco, Reference Vieta and De Prisco2024). Longitudinal studies of the population could track the long-term impact of OCs on brain and cognitive development, enabling measures to prevent and reduce the risk of developing psychosis and cognitive impairment. In our mediation analysis, we did not include potential confounders, such as educational level or chlorpromazine equivalents (which might be associated with both LGI and cognitive functioning) (Ilzarbe & Vieta, Reference Ilzarbe and Vieta2023) to reduce the complexity and avoid multicollinearity in the specific case of educational level.
Several strengths counterbalance the aforementioned limitations, such as a relatively large sample size and the lack of confounders associated with the protracted duration of illness and treatment exposure. To the best of the author’s knowledge, this is the first study to examine the potential impact of OCs on brain gyrification processes and how they can affect cognitive performance in FEP.
In summary, we observed that FEP patients differ from HCs in terms of gyrification in the parietal, occipital, and cingulate cortices. Our findings suggest that OCs are associated with altered cortical gyrification, particularly in the LCC, with a differential effect of OC exposure between patients and HCs, especially evident in males and driven by antepartum complications.
Among the HCs, OCs were negatively associated with WM. This pattern potentially reflects compensatory neurodevelopmental mechanisms in response to early adversity, which is also supported by the differential effects of OCs on the brain. In contrast, neither direct nor mediated effects of OCs on cognition were observed among individuals with FEP, suggesting a more complex interplay between early-life adversity, cortical neurodevelopment, and cognition in psychosis. Sex differences may modulate this complex interplay, and the underrepresentation of female participants in our sample limits the strength of our conclusions. Future studies with larger, sex-balanced samples are essential to clarify the role of sex in shaping the neurodevelopmental impact of OCs in psychosis.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/S0033291725100974.
PEPs Group
Nora Guasch-Capella5,6, Anaid Perez-Ramos5,6,7,19, Jessica Merchán-Naranjo6,9, Laura Pina-Camacho6,9, Alexandra Roldán6,20, Anna Alonso-Solís6,20, Ana González-Pinto6,21, Iñaki Zorrilla6,21, Pedro M. Ruiz-Lázaro6,22, David Vaquero-Puyuelo6,22, Maria José Escarti6,23,24,25, Marta Perez-Rando6,24,26, Anna Mané6,12,13,14, Angeles Malagon6,12,13,14, Maria Serra-Navarro7,8, Eduard Vieta2,6,7,8, Josefina Castro-Fornieles2,6,7,15, Inmaculada Baeza2,6,7,15, Francesco Dal Santo6,27, Paz García-Portilla6,27,28, Rafael Segarra Echevarria6,29,30, Arantzazu Zabala Rabadán6,29,30, Roberto Rodriguez-Jimenez6,31,32, Isabel Martínez-Gras31,32,33, Judith Usall34, Anna Butjosa6,34, Edith Pomarol-Clotet6,17, Raymond Salvador6,17, Angela Ibañez6,35, Manuel J. Cuesta10,11, Vicent Balanzá-Martínez36
19Neuropsychopharmacology and Psychobiology Research Group, Department of Neuroscience, University of Cadiz, Cadiz, Spain
20Department of Psychiatry, Institut d’Investigació Biomèdica Sant Pau, Hospital de la Santa Creu i Sant Pau, Barcelona, Spain
21Department of Psychiatry, Hospital Universitario de Alava, UPV/EHU, BIOARABA, Vitoria, Spain
22Department of Medicine and Psychiatry, Universidad de Zaragoza, Instituto de Investigación Sanitaria Aragón (IIS Aragón), Zaragoza, Spain
23Servicio de Psiquiatria, Hospital Clínico Universitario de Valencia, Valencia, Spain
24Fundación Investigación Hospital Clínico de Valencia, INCLIVA, Valencia, Spain
25Department of Medicine, University CEU-UCH, Valencia, Spain
26Neurobiology Unit, Institute for Biotechnology and Biomedicine (BIOTECMED), Universitat de València, Valencia, Spain
27Servicio de Salud del Principado de Asturias (SESPA), Instituto de Investigación Sanitaria del Prinicipado de Aturias (ISPA), Oviedo, Spain
28Área de Psiquiatría, Departamento de Medicina, Universidad de Oviedo, Oviedo, Spain
29Área de Psiquiatría, Departamento de Neurociencias, Universidad del País Vasco UPV/EHU, Bilbao, Spain
30Instituto de Investigación Sanitaria BioCruces Bizkaia, Barakaldo, Spain
31Faculty of Medicine, Complutense University of Madrid (UCM), Madrid, Spain
32Department of Psychiatry, Instituto de Investigación Sanitaria Hospital 12 de Octubre (imas12), Madrid, Spain
33RETIC (Network of Addictive Conditions), Institute of Health Carlos III, Madrid, Spain
34Institut de Recerca Sant Joan de Déu, Department of Child and Adolescent Mental Health, Parc Sanitari Sant Joan de Déu, Sant Boi de Llobregat, Spain
35Department of Psychiatry, Hospital Universitario Ramon y Cajal, Universidad de Alcala, Instituto Ramon y Cajal de InvestigacionSanitaria (IRYCIS), Madrid, Spain
36Teaching Unit of Psychiatry and Psychological Medicine, Department of Medicine, University of Valencia, CIBERSAM, INCLIVA, Valencia, Spain
Acknowledgments
The authors would like to thank all participants for their contribution.
Funding statement
This study is part of a coordinated multicenter Project, PEPs study, funded by the Ministerio de Economía y Competitividad (PI08/0208; PI11/00325; PI14/00612), Instituto de Salud Carlos III – Fondo Europeo de Desarrollo Regional. Unión Europea. Una manera de hacer Europa, Centro de Investigacion Biomédica en Red de salud Mental, CIBERSAM, Instituto de Salud Carlos III, by the CERCA Program/Generalitat de Catalunya and Secretaria d’Universitats i Recerca of the Departament d’Economia I Coneixement (2017SGR1355). Departament de Salut de la Generalitat de Catalunya (2017), funds from the Pla Estratègic de Recerca i Innovacio en Salut (PERIS 2016–2020), Projectes de recerca orientats al’atencio primària, SLT006/17/00345. This study has been funded by the Instituto de Salud Carlos III (ISCIII) through the projects ‘PI20/00661 and PI24/00196’ and co-funded by the European Union. CGR received funding from the Maria i Núria Cunillera legacy (FCRB_CU1_2024). NV and IV were supported by a BITRECS fellowship that received funding from the European Union’s Horizon 2020 research and innovation program under the grant agreement No. 754 550 and from ‘La Caixa’ Foundation under the agreement LCF/PR/GN18/5031000. IB received funding from the Pons-Bartran legacy (FCRB_IPB2_2023), as well as the Spanish Ministry of Science, Innovation and Universities, Instituto de Salud Carlos III (PI21/00391, co-funded by the European Union, ‘One way to make Europe’). RRJ is supported by the Instituto de Salud Carlos III, Fondo de Investigación Sanitaria (FIS), Centro de Investigación Biomédica en Red de Salud Mental (CIBERSAM), Madrid Regional Government (S2010/ BMD-2422 AGES; S2017/BMD-3740; S2022/BMD-7216 AGES 3-CM), the Madrid Regional Government (R&D activities in Biomedicine S2022/BMD-7216 (AGES 3-CM), and CIBERSAM-ISCIII. AI was supported by CIBER – Consorcio Centro de Investigación Biomédica en Red – (CB/07/09/0025), Instituto de Salud Carlos III, Ministerio de Ciencia e Innovación; by the Madrid Regional Government (S2022/BMD-7216 (AGES 3-CM) and European Union Structural Funds; and by grants PI22/01183 and ICI21/00089, integrated into the Plan Nacional de I + D + I and co-financed by the ISCIII-Subdirección General de Evaluación and the Fondo Europeo de Desarrollo Regional (FEDER).
Competing interests
NV has received financial support for CME activities and travel funds from the following entities: Angelini, Janssen-Cilag, Lundbeck, and Otsuka. CGR has received grants from/or served as a consultant, advisor, or speaker for the following entities: Adamed, Angelini, Casen-Recordati, Janssen-Cilag, Lunbeck, and Newron. IB has received honoraria or travel support from Otsuka-Lundbeck and Angelini. EV has received grants and served as a consultant, advisor, or CME speaker for the following entities: AB-Biotics, AbbVie, Adamed, Alcediag, Angelini, Biogen, Beckley-Psytech, Biohaven, Boehringer-Ingelheim, Celon Pharma, Compass, Dainippon Sumitomo Pharma, Esteve, Ethypharm, Ferrer, Gedeon Richter, GH Research, Glaxo-Smith Kline, HMNC, Idorsia, Johnson & Johnson, Lundbeck, Luye Pharma, Medincell, Merck, Newron, Novartis, Orion Corporation, Organon, Otsuka, Roche, Rovi, Sage, Sanofi-Aventis, Sunovion, Takeda, Teva, and Viatris, outside the submitted work. RRJ has been a consultant for, spoken in activities of, or received grants from JanssenCilag, Lundbeck, Otsuka, Pfizer, Ferrer, Juste, Takeda, Exeltis, Casen-Recordati, Angelini, and Rovi. AI has received research support from or served as a speaker or advisor for JanssenCilag, Lundbeck, Otsuka Pharmaceutical, Alter, Rovi, Casen Recordati, and Viatris.
All authors report no financial relationship relevant to the subject of this article.
Open access
