Schizophrenia Bulletin vol. 48 no. 2 pp. 514–523, 2022 https://doi.org/10.1093/schbul/sbab122 Advance Access publication October 8, 2021 Neural Substrates of Psychotic Depression: Findings From the Global ECT-MRI Research Collaboration Akihiro Takamiya1,2, Annemiek Dols3,4, Louise Emsell2, Christopher Abbott5, Antoine Yrondi6, Carles Soriano Mas7,8,9, Martin Balslev Jorgensen10,11, Pia Nordanskog12, Didi Rhebergen13, Eric van Exel3,4, Mardien L. Oudega3,4, Filip Bouckaert2, Mathieu Vandenbulcke2, Pascal Sienaert14, Patrice Péran15, Marta Cano8,16,17, Narcis Cardoner16, Anders Jorgensen10,11, Olaf B. Paulson18, Paul Hamilton19, Robin Kampe19, Willem Bruin20, Hauke Bartsch21,22,23, Olga Therese Ousdal21,24, Ute Kessler25,26, Guido van Wingen20, Leif Oltedal21,25, and Taishiro Kishimoto*,1 1 Department of Neuropsychiatry, Keio University School of Medicine, Tokyo, Japan; 2Department of Neurosciences and Neuropsychiatry, Leuven Brain Institute, KU Leuven, Leuven, Belgium; 3GGZ inGeest Specialized Mental Health Care, Amsterdam, the Netherlands; 4Department of Psychiatry, Amsterdam Neuroscience, Amsterdam UMC, Vrije Universiteit Amsterdam, Amsterdam, the Netherlands; 5Department of Psychiatry, University of New Mexico, Albuquerque, NM, USA; 6Service de Psychiatrie et de Psychologie Médicale, Centre Expert Dépression Résistante FondaMental, CHU Toulouse, Hospital Purpan, ToNIC, Toulouse NeuroImaging Center, Université de Toulouse, Inserm, UPS, Toulouse, France; 7Department of Psychiatry, Bellvitge Biomedical Research Institute-IDIBELL, Barcelona, Spain; 8CIBERSAM, Carlos III Health Institute, Madrid, Spain; 9Department of Psychobiology and Methodology in Health Sciences, Universitat Autònoma de Barcelona, Barcelona, Spain; 10Psychiatric Centre Copenhagen, Copenhagen, Denmark; 11Institute of Clinical Medicine, University of Copenhagen, Copenhagen, Denmark; 12Department of Clinical and Experimental Medicine, Center for Social and Affective Neuroscience (CSAN), Linköping University, Linköping, Sweden; 13 Mental Health Care Institute, GGZ Centraal, Amersfoort, the Netherlands; 14Academic Center for ECT and Neurostimulation (AcCENT), University Psychiatric Center (UPC)—KU Leuven, Kortenberg, Belgium; 15ToNIC, Toulouse NeuroImaging Center, Université de Toulouse, Inserm, UPS, Toulouse, France; 16Mental Health Department, Unitat de Neurociència Traslacional, Parc Tauli University Hospital, Institut d’Investigació i Innovació Sanitària Parc Taulí (I3PT), Barcelona, Spain; 17Department of Psychobiology and Methodology of Health Sciences, Universitat Autònoma de Barcelona, Barcelona, Spain; 18Neurobiological Research Unit, Rigshospitalet, Copenhagen, Denmark; 19Department of Biomedical and Clinical Sciences, Center for Social and Affective Neuroscience (CSAN), Linköping University, Linköping, Sweden; 20Department of Psychiatry, Amsterdam Neuroscience, Amsterdam UMC, Amsterdam, the Netherlands; 21Department of Radiology, Mohn Medical Imaging and Visualization Centre, Haukeland University Hospital, Bergen, Norway; 22Department of Research and Innovation, Haukeland University Hospital, Bergen, Norway; 23Department of Informatics, University of Bergen, Bergen, Norway; 24Faculty of Psychology, Centre for Crisis Psychology, University of Bergen, Bergen, Norway; 25Department of Clinical Medicine, University of Bergen, Bergen, Norway; 26Division of Psychiatry, NORMENT, Haukeland University Hospital, Bergen, Norway *To whom correspondence should be addressed; Department of Neuropsychiatry, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan; tel: +81-3-5363-3829; fax: +81-3-5379-0187; e-mail: tkishimoto@keio.jp Psychotic major depression (PMD) is hypothesized to be a identify longitudinal effects of ECT on GMV and longitudinal distinct clinical entity from nonpsychotic major depression differential effects of ECT between PMD and NPMD, respec- (NPMD). However, neurobiological evidence supporting this tively. Compared with NPMD, PMD showed lower GMV in notion is scarce. The aim of this study is to identify gray matter the prefrontal, temporal and parietal cortex before ECT; PMD volume (GMV) differences between PMD and NPMD and showed lower GMV in the medial prefrontal cortex (MPFC) their longitudinal change following electroconvulsive therapy after ECT. Although there was a significant main effect of time (ECT). Structural magnetic resonance imaging (MRI) data on GMV in several brain regions in both PMD and NPMD, from 8 independent sites in the Global ECT-MRI Research there was no significant group-by-time interaction. Lower Collaboration (GEMRIC) database (n = 108; 56 PMD and GMV in the MPFC was consistently identified in PMD, 52 NPMD; mean age 71.7 in PMD and 70.2 in NPMD) were suggesting this may be a trait-like neural substrate of PMD. analyzed. All participants underwent MRI before and after Longitudinal effect of ECT on GMV may not explain superior ECT. First, cross-sectional whole-brain voxel-wise GMV com- ECT response in PMD, and further investigation is needed. parisons between PMD and NPMD were conducted at both time points. Second, in a flexible factorial model, a main ef- Key words: psychosis/depression/magnetic resonance fect of time and a group-by-time interaction were examined to imaging/gray matter volume/medial prefrontal cortex © The Author(s) 2021. Published by Oxford University Press on behalf of the Maryland Psychiatric Research Center. This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs licence (https:// creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial reproduction and distribution of the work, in any medium, provided the original work is not altered or transformed in any way, and that the work is properly cited. For commercial re-use, please contact journals.permissions@ oup.com 514 Neural Substrates of Psychotic Depression Introduction of clinical heterogeneity (eg, PMD or NPMD) on the GMV increase following ECT has not been investigated. The current diagnostic category of major depressive dis- In addition, previous studies have revealed larger GMV order (MDD) has been extensively criticized for the clinical increase in patients who remitted after ECT compared and biological heterogeneity within the same diagnosis.1,2 to those who did not remit,32,37 although the relationship Many efforts have been made to develop clinically and bi- between GMV increase following ECT and clinical out- ologically plausible subgroups within the broad concept come is still on debate.30,31 Since PMD is known to show of MDD. Historically, clinicians and researchers have superior ECT response,20 we hypothesized that PMD proposed that MDD with psychotic features (psychotic may show larger neuroplastic changes following ECT, major depression [PMD]) is a distinct clinical entity from and that this could be related to superior ECT response nonpsychotic major depression (NPMD).3–5 PMD is not compared to NPMD. a rare condition,6,7 and its prevalence increases across the The Global ECT-MRI Research Collaboration lifespan with higher rates in late-life depression (eg, up (GEMRIC) is a multisite consortium collecting clin- to 45% in older hospitalized patients).6,8 Compared with ical and neuroimaging data from patients who received NPMD, PMD has been associated with more severe cog- ECT.31 Neuroimaging data are acquired longitudinally nitive impairment,9 higher recurrence rate,10 higher sui- (eg, before and after ECT), which provides a unique op- cide risk,7,11 more psychosocial impairment,12 and lower portunity to make assumptions about the underlying neu- quality of life.13 Moreover, PMD has shown a diagnostic robiology of PMD and its stability during the treatment stability over the course of episodes compared to other course. Furthermore, we have an opportunity to examine suggested subtypes (eg, agitated/retarded depression),14 neurobiological characteristics that could explain the dif- and a stability of the clinical features of delusions and ferences in the clinical outcome of ECT between PMD hallucinations over time.5 From a therapeutic point of and NPMD.17–20 In addition, studying the extreme clinical view, PMD has shown poor response to placebo,15 and subset of depressed patients (eg, those who needed ECT) antidepressant/antipsychotic monotherapy,16 but fa- may reduce etiological heterogeneity and increase statis- vorable response to the combination of antipsychotics tical power.38,39 and antidepressants,16 and to electroconvulsive therapy The primary aim of the current study was to identify (ECT).17–19 Notably, ECT is particularly effective in late- regional GMV differences between PMD and NPMD in life PMD,20 and ECT is considered a first-line treatment cross-sectional whole-brain voxel-wise comparisons at for PMD in several clinical guidelines.21 Compared to this two time points (ie, before and after ECT). The secondary clinical and therapeutic evidence that supports the notion aim was to identify longitudinal differential effects of of PMD as a distinct clinical entity, little is known about ECT on GMV between PMD and NPMD, which could its underlying neurobiology. explain the superior ECT response in PMD reported in Previous structural MRI (magnetic resonance im- the literature. aging) studies that conducted a direct comparison be- tween PMD and NPMD reported inconsistent results: smaller prefrontal cortex,22–24 smaller brain stem,23 larger Methods lateral and third ventricle,23 larger posterior sulcus,25 and thinner cortex in multiple structural networks26 in PMD Participants compared with NPMD, whereas others did not find any All GEMRIC sites contributing data received approval differences.27–29 These inconsistencies may be due to the from their local ethics committees or institutional review methodological differences, including a variety of mag- board, and the centralized mega-analysis was approved netic field strengths (eg, from 0.5T to 3T), limited sample by the Regional Ethics Committee South-East in Norway size (eg, the average number of PMD patients in pre- (2018/769). Written informed consent was obtained from vious studies was 19.8), limited brain regions investigated all participants. using regions of interests (ROIs), or age difference across In the current study, we first selected 8 GEMRIC sites studies (supplementary table 1; supplementary material). collecting 3T MRI data from both PMD and NPMD to Gray matter volume (GMV) increase, especially in reduce between-group scanner variability. The first au- the hippocampus and amygdala, has been consistently thor (A.T.) investigated the inclusion criteria in each site reported following ECT.30,31 Recent evidence has re- from already published papers19,32,40–45 and contacted each vealed the effect of ECT was prominent in one of the site’s principal investigator or GEMRIC members to con- most neuroplastic brain regions (ie, dentate gyrus in firm the inclusion of both psychotic and nonpsychotic the hippocampus),32,33 which is in line with neuroplastic depressed patients in each site. We included subjects aged hypothesis of ECT action.34 Several factors, including 60 years or older (ie, late-life depression) with a diagnosis age,30 the number of ECT,31 induced electric field,35 or of MDD with/without psychotic features. Individuals induced seizure duration,36 have been associated with with another major psychiatric disorder (eg, bipolar dis- this GMV increase following ECT. However, an effect order, schizoaffective disorder, or schizophrenia), those 515 A. Takamiya et al with neurological disorders (eg, Parkinson’s disease, de- First, a cross-sectional whole-brain voxel-wise GMV mentia), those with active physical illness, those who had comparison between PMD and NPMD at TP1 was con- contraindications to MRI, or those who received ECT at ducted. Age, sex, TIV, site, and HAM-D total scores least within the past 3 months were excluded. In all but were included as covariates. HAM-D total scores were 2 sites, concurrent psychotropic medications were kept included in the statistical model to identify regional during a course of ECT. GMV differences which relate to PMD independent of Clinical characteristics (eg, age, sex, depression se- the severity of depression. Threshold-free cluster en- verity, and concurrent psychotropic medications) and hancement (TFCE) was used for multiple comparison high-resolution T1-weighted images were used for the corrections,47 which was demonstrated as a sensitive and current analyses. Depression severity was assessed stable method for a group-level analysis of voxel-based by Montgomery-Asberg Depression Rating Scale morphometry data.48 The number of permutations was (MADRS) and/or 17-item Hamilton Depression Rating set to 20 000. The statistical significance threshold was set Scale (HAM-D). MADRS total scores were converted to at family-wise error (FWE)-corrected P < .05. The voxel- HAM-D by applying a validated equation.46 Response wise group comparisons were conducted including anti- was defined as a decrease of 50% or more, and remission depressants, antipsychotics, and lithium use as additional was defined as HAM-D total scores ≤7. covariates (categorical variable: yes = 1; no = 0) to ac- count for the potential effects of medications. The same MRI Acquisition and Image Processing voxel-wise GMV comparison between PMD and NPMD was conducted using data at TP2 to investigate whether Structural MRI data from 8 independent sites in the the regional GMV differences at TP1 would be also ob- GEMRIC database were analyzed. High-resolution served at TP2. T1-weighted images were acquired before (TP1) and after Second, we created a flexible factorial design that in- (TP2) ECT in all sites. Imaging parameters at each site cluded a between-subject factor group (PMD vs NPMD) were provided in supplementary table 2. All images were and a within-subject factor time (TP1 and TP2) to inves- processed using the default pipeline of the Computational tigate whether the longitudinal effect of ECT on GMV Anatomy Toolbox (CAT12, http://dbm.neuro.uni-jena. would differ between PMD and NPMD. A group-by- de/cat/), a toolbox for Statistical Parametric Mapping time interaction was examined to identify the longi- software (SPM12, http://www.fil.ion.ucl.ac.uk/spm). tudinal differential effects of ECT on GMV between Preprocessing included bias-correction, segmentation groups. A main effect of time was also conducted to into gray matter (GM), white matter (WM), and cerebro- replicate and confirm the already reported ECT-related spinal fluid (CSF), spatial normalization into MNI space GMV increase in widely distributed brain regions regard- using the DARTEL algorithm, and modulation. The iso- less of the presence of psychosis.31 In addition, paired t tropic voxel size of the resultant images was 1.5 mm. The tests were also conducted to investigate the longitudinal CAT12 longitudinal preprocessing pipeline including ad- effects of ECT on GMV in PMD and NPMD separately. ditional within-subject registration and bias-correction The statistical significance threshold was set at voxel-level before segmentation was used when investigating the lon- FWE-corrected P < .05. gitudinal effect of ECT on GMV. After preprocessing, a quality check was conducted by using the “Check sample Results homogeneity” module implemented in CAT12, which offers a straightforward visualization of the correlation Clinical demographics of the participants are presented and the Mahalanobis distance between the volumes using in table 1. Of the 108 patients with MDD, 56 patients a boxplot and correlation matrix, to identify outliers. were PMD, and 52 patients were NPMD. There were no All outliers were manually checked by 4 authors (A.T., significant differences in age (P = .42) nor gender dis- W.B., M.C., and G.W.) and low-quality images from 3 tribution (P = .76) between the 2 groups. PMD showed participants were excluded in this study. Then images higher HAM-D total scores than NPMD before ECT were smoothed with 10-mm full-width at half maximum (PMD: 30.5 ± 8.3; NPMD: 24.6 ± 5.2; P = .001). PMD (FWHM) Gaussian kernel. Total intracranial volume showed higher response rate than NPMD (92.8% vs (TIV) was calculated using CAT12. 80.8%), although this difference did not reach statistical significance, and showed similar remission rate as NPMD Statistical Analysis (73.2% vs 71.2%). Descriptive statistics were used to characterize the study participants. Distributions of all variables were inspected Cross-Sectional Regional GMV Differences Between using histograms and Shapiro-Wilk tests. A whole-brain PMD and NPMD at Two Time Points GMV analysis was conducted using SPM12. Only voxel In the cross-sectional whole-brain voxel-wise analysis, values greater than 0.2 were analyzed using absolute PMD showed lower GMV compared with NPMD at threshold masking option in SPM. TP1 in the medial segment of the superior frontal gyrus 516 Neural Substrates of Psychotic Depression Table 1. Clinical Characteristics of Participants showed lower GMV in the right medial prefrontal cortex (MPFC) at TP2 (figure 1; supplementary table 4). P After including medication use as additional covariates, PMD NPMD value* a new region (eg, left SFG medial segment) was identified Number 56 52 as a significant larger brain region in PMD compared Age, years .42 to NPMD, whereas some brain regions in the left hem- Mean (SD) 71.7 (7.7) 70.2 (6.7) isphere (eg, left precentral gyrus, left posterior cingulate Median (IQR) 70.0 (65.8–77.0) 69.0 (65.0–75.3) gyrus, and left MFG) were no longer significant (supple- Female 35 32 .76 mentary figure 1; supplementary table 5). At TP2, PMD (62.5%) (65.4%) HAM-D before .001 showed lower GMV in the right MPFC (supplementary ECT table 6). Mean (SD) 30.5 (8.3) 24.6 (5.2) Median (IQR) 31.1 (24.2–37.6) 24.8 (20.6–27.7) HAM-D after ECT .38 Longitudinal Effect of ECT on GMV Between Groups Mean (SD) 5.1 (6.5) 6.1 (7.3) There was no significant group (PMD vs NPMD) by time Median (IQR) 3.0 (1.0–7.9) 4.0 (0.8–9.0) (TP1 vs TP2) interaction, whereas there was a significant %change of .17 HAM-D scores effect of time on GMV in widely distributed brain re- Mean (SD) 82.4 (22.8) 74.3 (31.7) gions with the largest effect on the right medial temporal Median (IQR) 89.3 (75.4–97.5) 82.7 (67.1–96.8) lobe (MTL) (supplementary figure 2). The results of the Response rate 51/55 42/52 .10 separate paired t tests in PMD and NPMD are presented (92.8%) (80.8%) in figure 2 and show that the main effect of time is present Remission rate 41/56 37/52 .81 (73.2%) (71.2%) in both groups (supplementary tables 7 and 8). All identi- Number of ECT .10 fied regions showed GMV increase following ECT. session Mean (SD) 12.2 (4.5) 11.2 (5.0) Discussion Median (IQR) 12.0 (9.0–14.3) 10.5 (8.8–13.0) Electrode place- .93 To the best of our knowledge, this is the first study to in- ment vestigate GMV differences between PMD and NPMD at RUL 31 (55.4%) 30 (57.7%) BT 15 (26.8%) 12 (23.1%) two time points. We found that PMD showed lower GMV RUL to BT 10 (17.9%) 10 (19.2%) in the MPFC compared to NPMD both before and after Medications ECT in this largest cohort ever, suggesting that lower Antidepressants 30 (53.6%) 33 (63.5%) .30 GMV in the MPFC is one of the neural substrates of Antipsychotics 22 (39.3%) 20 (38.5%) .93 PMD. We also found that ECT increased GMV in widely Lithium 1 (1.8%) 3 (5.8%) .30 Benzodiazepine 10 (17.9%) 10 (19.2%) 1.0 distributed brain regions in both PMD and NPMD, but the longitudinal effect of ECT on GMV did not differ be- Abbreviations: BT, bitemporal; HAM-D, Hamilton Depression tween PMD and NPMD. Rating Scale; IQR, interquartile range; NPMD, nonpsychotic major depression; PMD, psychotic major depression; RUL, right Lower GMV in the MPFC in PMD unilateral. *Differences in participants’ characteristics between groups Previous studies suggest cortisol dysfunction in PMD. were examined using independent t test (HAM-D before ECT) For example, there are evidence of excessive activity of or Mann-Whitney U test (age, HAM-D after ECT, %change of the hypothalamic-pituitary-adrenal (HPA) axis,49 in- HAM-D scores, Number of ECT sessions) for continuous vari- ables, and χ 2 analysis (Response rate, Remission rate, Antidepres- creased urinary cortisol levels,50 elevated serum adreno- sants, Antipsychotics) or Fisher exact test (Electrode placement, corticotropic hormone,51 increased evening cortisol levels, Lithium) for categorical variables. Each test was selected in ac- and high rates of non-suppression of cortisol on the dex- cordance with the distribution of each variable. amethasone suppression test52 in patients with PMD. Genetic variation of the glucocorticoid receptor and (SFG), left middle frontal gyrus (MFG), right amygdala/ corticotropin-releasing hormone receptor 1 has also been hippocampus, right lingual gyrus, medial segment of the associated with psychotic depression.49 Preclinical studies posterior cingulate gyrus, and praecuneus (figure 1; sup- have reported corticosterone-induced dendritic atrophy plementary table 3). We did not identify brain regions in the hippocampus53 and MPFC.54 The relationship be- with greater volume in PMD than NPMD. We conducted tween elevated cortisol levels and prefrontal,55 temporal an additional analysis excluding each site which provided and parietal atrophy56 have been reported in healthy in- less than 10 subjects to reduce potential site/scanner ef- dividuals. A meta-analysis reported the relationship be- fects, and found similar results, showing lower GMV tween increased cortisol levels and hippocampal atrophy in PMD compared to NPMD in the fronto-temporo- in patients with late-life depression.57 This cortisol-GMV parietal brain regions at TP1 (data not shown). PMD relationship was also reported in subjects at high risk of 517 A. Takamiya et al Fig. 1. Cross-sectional whole-brain voxel-wise GMV comparisons between PMD and NPMD at two time points. PMD showed lower GMV in the left middle frontal gyrus, medial prefrontal cortex (MPFC), praecuneus, right amygdala/hippocampus, and right lingual gyrus at TP1 (before ECT). PMD showed lower GMV in the MPFC at TP2 (after ECT). There were no brain regions that were larger in PMD than NPMD. Significance threshold was set at family-wise error-corrected P < .05 determined by threshold-free cluster enhancement. The color bars represent –log(P) (ie, 1.3 is equivalent to “P = .05”). Abbreviations: ECT, electroconvulsive therapy; GMV, gray matter volume; NPMD, nonpsychotic major depression; PMD, psychotic major depression. psychosis,58 and this interaction between cortisol and dopamine activity in the MPFC,59 which may be asso- GMV reduction may be associated with vulnerability to ciated with psychotic symptoms. Future studies should emerging psychosis. Although speculative, the underlying focus on the complex relationships among cortisol levels, mechanisms of lower prefrontal and hippocampal GMV dopamine, GMV reduction, and psychotic symptoms in PMD might be related to the dysregulated cortisol in PMD to elucidate the underlying biology of affective system. Moreover, stress and/or glucocorticoids increase psychosis. 518 Neural Substrates of Psychotic Depression Fig. 2. Longitudinal effects of ECT on GMV. Results of paired t tests for data at two time points (before and after ECT) in PMD and NPMD. GMV increased in widely distributed brain regions following ECT in both PMD and NPMD. Significance threshold was set at family-wise error-corrected P < .05 determined by threshold-free cluster enhancement. The color bars represent –log(P) (ie, 1.3 is equivalent to “P = .05”). Red represents regional GMV increase following ECT in PMD, and green represents regional GMV increase following ECT in NPMD. Yellow represents overlapped brain regions. Although there seems to be regional GMV increase specific to PMD or NPMD in the figure, there was no significant group-by-time interaction in a flexible factorial model. Abbreviations: ECT, electroconvulsive therapy; GMV, gray matter volume; NPMD, nonpsychotic major depression; PMD, psychotic major depression. Prefrontal Cortex and Psychosis guilt was more frequent in PMD compared to schizo- The development of delusion is hypothesized to be re- phrenia or other psychotic disorders.61 The latter factor, lated to the content of delusional belief and the failure to in other words, the dysfunction of the “belief evaluation reject the delusional belief despite all the evidence against system” may be common to the development of delu- it.60 The former can be different across individual patients sional thinking. In general, the prefrontal cortex is related or across psychotic disorders. For instance, delusion of to the integration of sensory inputs, spatial organization, 519 A. Takamiya et al and coordination of sensory and emotional efference.62 One interpretation is that ECT-related GMV change The lateral PFC is considered to have a role in evaluating may not be related to clinical improvement as previously information from other cortical areas, and rejecting those shown in the meta-30 and mega-analysis.31 Future inves- interpretations that are unreal.61 The medial PFC is as- tigations would benefit from multi-modal data analysis sociated with self-reflection, past or future events, meta- to elucidate the underlying neural mechanisms relating cognition, and hypothetical scenarios.63,64 We found lower to superior response to ECT in PMD. Another explana- GMV in the medial and lateral PFC in PMD, suggesting tion for our result might be related to older participants that these prefrontal GMV reductions may be associated in this study. Older age is associated with superior ECT with dysfunction of the belief evaluation system and with response,20 and indeed, 80.8% of NPMD in this study the onset of affective psychosis. Prefrontal GMV reduc- met the response criteria. There was no statistically sig- tion has been reported in patients with schizophrenia,65 nificant difference in response nor remission rate between patients with bipolar disorder with delusion,66 and pa- PMD and NPMD in our cohort, although PMD showed tients with bipolar disorder with a history of psychosis.67 numerically higher response rate (eg, 92.8% vs 80.8%). Our finding of the prefrontal GMV reduction in PMD Since HAM-D does not account for detailed symptom compared with NPMD may be a common neural sub- domains in PMD, the more suitable rating scale, such as strate of psychosis across different disorders. Considering the Psychotic Depression Assessment Scale,74 should be the differences in clinical course5,10,12–14 and the treatment used in future studies. strategy between PMD and NPMD21 as well as our re- sults showing different neurobiology between PMD and Strength and Limitations NPMD, PMD could be considered as a different clinical entity from NPMD. First, our study design and participants should be em- phasized. Because the GEMRIC has collected clinical and MRI data from depressed patients who received Default Mode Network and PMD ECT, we had this unique opportunity to investigate the In addition to the MPFC, PMD showed lower GMV in GMV differences between PMD and NPMD at two time the praecuneus; our results could be interpreted as lower points and to investigate the longitudinal effect of ECT GMV in the critical hubs of the default mode network on GMV in this largest cohort ever. By this design, we (DMN) in PMD. The DMN is associated with self-refer- provide evidence of a GMV difference in the MPFC ential thinking,68 autobiographical memory retrieval,69,70 between PMD and NPMD which was stable during a and depressive rumination.71 Notably, the frequency of course of ECT, although ECT has a widespread effect on ruminative thinking was higher in PMD than NPMD.3 the GMV. Moreover, this is the largest multisite study to An abnormal DMN-related functional connectivity pat- conduct a direct comparison between PMD and NPMD tern has been reported in PMD during a remitted state (supplementary table 1). Because of the multisite na- compared to healthy subjects.72 Our results provide evi- ture, there might be a scanner effect and/or unpredict- dence that structural abnormality of the DMN regions able confounders between sites. However, we selected the may be related to PMD in addition to functional abnor- sites which included both PMD and NPMD from the mality reported so far. However, functional abnormality GEMRIC database to reduce the effect of scanner dif- of the DMN has also been reported in depression regard- ferences when comparing 2 groups, and we included site less of the presence of psychosis.71 A previous fMRI study information as a covariate. Second, because all patients did not find DMN abnormalities but found abnormal in our cohort were referred to ECT, the participants in functional connectivity of the frontoparietal network this study may be more homogeneous than other studies compared to NPMD, although this was not replicated in regarding the clinical characteristics (eg, severity and another cohort.73 Further study is needed to investigate treatment resistance) and regarding the etiology of de- how structural abnormality of the DMN would correlate pression.39 In addition, we focused on late-life depression with functional abnormality and its specificity to PMD. to reduce age-related clinical and biological variability, which could lead to a more homogeneous cohort. In con- trast, this limited inclusion could limit the generalizability ECT for PMD of our results. However, our cohort is not uncommon ECT is a recommended treatment for PMD,21 and supe- because PMD is prevalent in hospitalized patients with rior ECT response in PMD compared with NPMD has late-life depression,8 and ECT is a recommended treat- been consistently reported.17–20 In this study, there was no ment for PMD in several guidelines.21 Moreover, focusing significant group-by-time interaction in the longitudinal on extreme subset of the clinical spectrum is one reason- GMV analysis; the longitudinal effect of ECT on GMV able approach to uncover the underlying neurobiology did not differ between PMD and NPMD. Our result of depression.39 Third, PMD showed higher symptom suggests that GMV change with ECT may not explain severity than NPMD at TP1, which might affect GMV PMD’s superior ECT response reported in the literature. at baseline. However, we included HAM-D total scores 520 Neural Substrates of Psychotic Depression as a covariate in the statistical model, and GMV reduc- A.T., M.C., W.B., and G.V.W. analyzed data. A.T. and tion in the MPFC was also identified at TP2, when the T.K. interpreted the data. A.T. wrote the final manuscript differences in symptom severity disappeared. Fourth, al- draft. All authors contributed data, as well as critical re- though we had not collected uniform longitudinal cogni- vision of the manuscript. All authors approved the final tive assessments, there is a possibility that our results of manuscript. the lower GMV in PMD might be associated with cogni- tive dysfunction. We now collect prospectively uniform cognitive assessments in the GEMRIC, and we can in- References vestigate differences in cognitive function between PMD 1. Ostergaard SD, Jensen SO, Bech P. The heterogeneity of and NPMD, and their association with GMV and/or the depressive syndrome: when numbers get serious. Acta longitudinal changes during ECT course in future study. Psychiatr Scand. 2011;124(6):495–496. Fifth, although we accounted for the effect of medication 2. Lynch CJ, Gunning FM, Liston C. Causes and conse- on brain volume, we only included it as a binary variable quences of diagnostic heterogeneity in depression: paths to discovering novel biological depression subtypes. Biol in our statistical models. The cumulative effect of anti- Psychiatry. 2020;88(1):83–94. psychotic medication75 might affect brain volumes and it 3. Charney DS, Nelson JC. Delusional and nondelusional uni- should be investigated in future studies. polar depression: further evidence for distinct subtypes. Am J In conclusion, lower GMV in the MPFC was consist- Psychiatry. 1981;138(3):328–333. ently identified in PMD both before and after ECT, sug- 4. Schatzberg AF, Rothschild AJ. Psychotic (delusional) major gesting this may be one of the trait-like neural substrates depression: should it be included as a distinct syndrome in of PMD. The longitudinal effect of ECT on brain struc- DSM-IV? Am J Psychiatry. 1992;149(6):733–745. ture did not differ between PMD and NPMD, suggesting 5. Nelson JC, Bickford D, Delucchi K, Fiedorowicz JG, Coryell WH. Risk of psychosis in recurrent episodes of GMV changes following ECT may not explain superior psychotic and nonpsychotic major depressive disorder: ECT response in PMD. Future research should focus a systematic review and meta-analysis. Am J Psychiatry. on brain functional changes to elucidate the underlying 2018;175(9):897–904. neural mechanisms of ECT action. 6. Ohayon MM, Schatzberg AF. Prevalence of depressive epi- sodes with psychotic features in the general population. Am J Psychiatry. 2002;159(11):1855–1861. Supplementary Material 7. Dold M, Bartova L, Kautzky A, et al. Psychotic features in patients with major depressive disorder: a report from the Supplementary material is available at Schizophrenia European Group for the Study of Resistant Depression. J Bulletin. Clin Psychiatry. 2019;80:17m12090. 8. Meyers BS, Greenberg R. Late-life delusional depression. J Affect Disord. 1986;11(2):133–137. Funding 9. Gomez RG, Fleming SH, Keller J, et al. The neuropsycho- logical profile of psychotic major depression and its relation This work was supported by Keio University Medical to cortisol. Biol Psychiatry. 2006;60(5):472–478. Science Fund (A.T., grant number 99-095-0007), AMED 10. Goldberg JF, Harrow M. Consistency of remission and out- (A.T., grant number JP20dm0307102h0003), Research come in bipolar and unipolar mood disorders: a 10-year pro- Foundation Flanders (FWO) grant (L.E., F.B., M.V., spective follow-up. J Affect Disord. 2004;81(2):123–131. grant number G0C0319N), KU Leuven Fund (L.M., F.B., 11. Gournellis R, Tournikioti K, Touloumi G, et al. Psychotic M.V., grant number C24/18/095), the Sequoia Fund for (delusional) depression and suicidal attempts: a system- Research on Ageing and Mental Health (M.V.), National atic review and meta-analysis. Acta Psychiatr Scand. 2018;137(1):18–29. Institute of Health (C.A., grant number MH125126, 12. Coryell W, Leon A, Winokur G, et al. Importance of psych- MH111826), Sara Borrell postdoctoral fellowships from otic features to long-term course in major depressive disorder. the Carlos III Health Institute (M.C., grant number Am J Psychiatry. 1996;153(4):483–489. CD20/00189), Lundbeck Foundation (M.B.J., A.J., 13. Cramer V, Torgersen S, Kringlen E. Mood disorders and O.B.P.), and Western Norway Regional Health Authority quality of life. A community study. Nord J Psychiatry. (L.O., grant number 911986 and 912238). 2010;64(1):58–62. 14. Coryell W, Winokur G, Shea T, Maser JD, Endicott J, Akiskal HS. The long-term stability of depressive subtypes. Conflict of Interest Am J Psychiatry. 1994;151(2):199–204. 15. Spiker DG, Kupfer DJ. Placebo response rates in psychotic and The authors report no conflict of interest. nonpsychotic depression. J Affect Disord. 1988;14(1):21–23. 16. Wijkstra J, Lijmer J, Burger H, Cipiriani A, Geddes J, Nolen WA. Pharmacological treatment for psychotic depres- Author Contribution sion. Cochrane Database Syst Rev. 2015;7:CD004044. 17. Petrides G, Fink M, Husain MM, et al. ECT remission rates L.O. and H.B. conducted consortium coordination. in psychotic versus nonpsychotic depressed patients: a report A.T. wrote the first draft and coordinated the work. from CORE. J ECT. 2001;17(4):244–253. 521 A. Takamiya et al 18. Birkenhäger TK, Pluijms EM, Lucius SA. ECT response 35. Argyelan M, Oltedal L, Deng ZD, et al. Electric field causes in delusional versus non-delusional depressed inpatients. J volumetric changes in the human brain. eLife. 2019;8:e49115. Affect Disord. 2003;74(2):191–195. 36. Takamiya A, Bouckaert F, Laroy M, et al. Biophysical mech- 19. Dols A, Bouckaert F, Sienaert P, et al. Early- and late-onset anisms of electroconvulsive therapy-induced volume ex- depression in late life: a prospective study on clinical and pansion in the medial temporal lobe: a longitudinal in vivo structural brain characteristics and response to electrocon- human imaging study. Brain Stimul. 2021;14:1038–1047. vulsive therapy. Am J Geriatr Psychiatry. 2017;25(2):178–189. 37. Takamiya A, Kishimoto T, Hirano J, Kikuchi T, Yamagata B, 20. van Diermen L, van den Ameele S, Kamperman AM, et al. Mimura M. Association of electroconvulsive therapy- Prediction of electroconvulsive therapy response and remis- induced structural plasticity with clinical remission. Prog sion in major depression: meta-analysis. Br J Psychiatry. Neuropsychopharmacol Biol Psychiatry. 2021;110:110286. 2018;212(2):71–80. 38. Soda T, McLoughlin DM, Clark SR, et al. International 21. Leadholm AK, Rothschild AJ, Nolen WA, Bech P, Munk- Consortium on the Genetics of Electroconvulsive Therapy Jørgensen P, Ostergaard SD. The treatment of psychotic and Severe Depressive Disorders (Gen-ECT-ic). Eur Arch depression: is there consensus among guidelines and psychi- Psychiatry Clin Neurosci. 2020;270(7):921–932. atrists? J Affect Disord. 2013;145(2):214–220. 39. Clements CC, Karlsson R, Lu Y, et al. Genome-wide asso- 22. Kim DK, Kim BL, Sohn SE, et al. Candidate neuroanatomic ciation study of patients with a severe major depressive epi- substrates of psychosis in old-aged depression. Prog sode treated with electroconvulsive therapy. Mol Psychiatry. Neuropsychopharmacol Biol Psychiatry. 1999;23(5):793–807. 2021;26(6):2429–2439. 23. Simpson S, Baldwin RC, Jackson A, Burns A. The differen- 40. Nordanskog P, Dahlstrand U, Larsson MR, Larsson EM, tiation of DSM-III-R psychotic depression in later life from Knutsson L, Johanson A. Increase in hippocampal volume nonpsychotic depression: comparisons of brain changes after electroconvulsive therapy in patients with depression: measured by multispectral analysis of magnetic resonance a volumetric magnetic resonance imaging study. J ECT. brain images, neuropsychological findings, and clinical fea- 2010;26(1):62–67. tures. Biol Psychiatry. 1999;45(2):193–204. 41. Abbott CC, Jones T, Lemke NT, et al. Hippocampal struc- 24. Oudega ML, van Exel E, Stek ML, et al. The structure of the tural and functional changes associated with electroconvul- geriatric depressed brain and response to electroconvulsive sive therapy response. Transl Psychiatry. 2014;4:e483. therapy. Psychiatry Res. 2014;222(1–2):1–9. 42. Jorgensen A, Magnusson P, Hanson LG, et al. Regional brain 25. Salokangas RKR, Cannon T, Van Erp T, et al. Structural mag- volumes, diffusivity, and metabolite changes after electrocon- netic resonance imaging in patients with first-episode schizo- vulsive therapy for severe depression. Acta Psychiatr Scand. phrenia, psychotic and severe non-psychotic depression and 2016;133(2):154–164. healthy control: results of the Schizophrenia and Affective 43. Bouckaert F, Dols A, Emsell L, et al. Relationship between Psychoses (SAP) projects. Br J Psychiatry. 2002;181:s58–s65. hippocampal volume, serum BDNF, and depression severity 26. Neufeld NH, Kaczkurkin AN, Sotiras A, et al. Structural following electroconvulsive therapy in late-life depression. brain networks in remitted psychotic depression. Neuropsychopharmacology. 2016;41(11):2741–2748. Neuropsychopharmacology. 2020;45(7):1223–1231. 44. Cano M, Martínez-Zalacaín I, Bernabéu-Sanz Á, et al. Brain 27. Keller J, Shen L, Gomez RG, et al. Hippocampal and volumetric and metabolic correlates of electroconvulsive amygdalar volumes in psychotic and nonpsychotic unipolar therapy for treatment-resistant depression: a longitudinal depression. Am J Psychiatry. 2008;165(7):872–880. neuroimaging study. Transl Psychiatry. 2017;7(2):e1023. 28. Vassilopoulou K, Papathanasiou M, Michopoulos I, et al. A 45. Yrondi A, Nemmi F, Billoux S, et al. Grey matter changes magnetic resonance imaging study of hippocampal, amyg- in treatment-resistant depression during electroconvulsive dala and subgenual prefrontal cortex volumes in major de- therapy. J Affect Disord. 2019;258:42–49. pression subtypes: melancholic versus psychotic depression. 46. Heo M, Murphy CF, Meyers BS. Relationship between the J Affect Disord. 2013;146(2):197–204. Hamilton Depression Rating Scale and the Montgomery- 29. Cohen JD, Nichols T, Keller J, Gomez RG, Schatzberg AF, Asberg Depression Rating Scale in depressed elderly: a meta- Reiss AL. Insular cortex abnormalities in psychotic major analysis. Am J Geriatr Psychiatry. 2007;15(10):899–905. depression: relationship to gender and psychotic symptoms. 47. Smith SM, Nichols TE. Threshold-free cluster enhance- Neurosci Res. 2013;75(4):331–339. ment: addressing problems of smoothing, threshold de- 30. Takamiya A, Chung JK, Liang KC, Graff-Guerrero A, pendence and localisation in cluster inference. Neuroimage. Mimura M, Kishimoto T. Effect of electroconvulsive therapy 2009;44(1):83–98. on hippocampal and amygdala volumes: systematic review 48. Li H, Nickerson LD, Nichols TE, Gao JH. Comparison of and meta-analysis. Br J Psychiatry. 2018;212(1):19–26. a non-stationary voxelation-corrected cluster-size test with 31. Ousdal OT, Argyelan M, Narr KL, et al.; GEMRIC. Brain TFCE for group-level MRI inference. Hum Brain Mapp. changes induced by electroconvulsive therapy are broadly dis- 2017;38(3):1269–1280. tributed. Biol Psychiatry. 2020;87(5):451–461. 49. Schatzberg AF, Keller J, Tennakoon L, et al. HPA axis genetic 32. Takamiya A, Plitman E, Chung JK, et al. Acute and long-term variation, cortisol and psychosis in major depression. Mol effects of electroconvulsive therapy on human dentate gyrus. Psychiatry. 2014;19(2):220–227. Neuropsychopharmacology. 2019;44(10):1805–1811. 50. Anton RF. Urinary free cortisol in psychotic depression. Biol 33. Gryglewski G, Lanzenberger R, Silberbauer LR, et al. Meta- Psychiatry. 1987;22(1):24–34. analysis of brain structural changes after electroconvulsive 51. Posener JA, DeBattista C, Williams GH, Chmura Kraemer H, therapy in depression. Brain Stimul. 2021;14(4):927–937. Kalehzan BM, Schatzberg AF. 24-Hour monitoring of 34. Bouckaert F, Sienaert P, Obbels J, et al. ECT: its brain enabling cortisol and corticotropin secretion in psychotic and effects: a review of electroconvulsive therapy-induced struc- nonpsychotic major depression. Arch Gen Psychiatry. tural brain plasticity. J ECT. 2014;30(2):143–151. 2000;57(8):755–760. 522 Neural Substrates of Psychotic Depression 52. Nelson JC, Davis JM. DST studies in psychotic depression: a 65. Fornito A, Yücel M, Patti J, Wood SJ, Pantelis C. Mapping meta-analysis. Am J Psychiatry. 1997;154(11):1497–1503. grey matter reductions in schizophrenia: an anatomical like- 53. Woolley CS, Gould E, McEwen BS. Exposure to excess gluco- lihood estimation analysis of voxel-based morphometry corticoids alters dendritic morphology of adult hippocampal studies. Schizophr Res. 2009;108(1–3):104–113. pyramidal neurons. Brain Res. 1990;531(1–2):225–231. 66. Radaelli D, Poletti S, Gorni I, et al. Neural correlates of delu- 54. Wellman CL. Dendritic reorganization in pyramidal neurons sion in bipolar depression. Psychiatry Res. 2014;221(1):1–5. in medial prefrontal cortex after chronic corticosterone ad- 67. Ekman CJ, Petrovic P, Johansson AG, Sellgren C, Ingvar M, ministration. J Neurobiol. 2001;49(3):245–253. Landén M. A history of psychosis in bipolar disorder is as- 55. Stomby A, Boraxbekk CJ, Lundquist A, et al. Higher diurnal sociated with gray matter volume reduction. Schizophr Bull. salivary cortisol levels are related to smaller prefrontal cortex 2017;43(1):99–107. surface area in elderly men and women. Eur J Endocrinol. 68. Gusnard DA, Akbudak E, Shulman GL, Raichle ME. Medial 2016;175(2):117–126. prefrontal cortex and self-referential mental activity: relation 56. Lebedeva A, Sundström A, Lindgren L, et al. Longitudinal to a default mode of brain function. Proc Natl Acad Sci U S relationships among depressive symptoms, cortisol, and A. 2001;98(7):4259–4264. brain atrophy in the neocortex and the hippocampus. Acta 69. Kim H. A dual-subsystem model of the brain’s default net- Psychiatr Scand. 2018;137(6):491–502. work: self-referential processing, memory retrieval pro- 57. Geerlings MI, Gerritsen L. Late-life depression, hippocampal cesses, and autobiographical memory retrieval. Neuroimage. volumes, and hypothalamic-pituitary-adrenal axis regula- 2012;61(4):966–977. tion: a systematic review and meta-analysis. Biol Psychiatry. 70. Zhu X, Wang X, Xiao J, et al. Evidence of a dissociation pat- 2017;82(5):339–350. tern in resting-state default mode network connectivity in 58. Valli I, Crossley NA, Day F, et al. HPA-axis function and first-episode, treatment-naive major depression patients. Biol grey matter volume reductions: imaging the diathesis-stress Psychiatry. 2012;71(7):611–617. model in individuals at ultra-high risk of psychosis. Transl 71. Hamilton JP, Farmer M, Fogelman P, Gotlib IH. Depressive Psychiatry. 2016;6:e797. rumination, the default-mode network, and the dark matter 59. Vaessen T, Hernaus D, Myin-Germeys I, van Amelsvoort T. of clinical neuroscience. Biol Psychiatry. 2015;78(4):224–230. The dopaminergic response to acute stress in health and psy- 72. Neufeld NH, Mulsant BH, Dickie EW, et al. Resting state chopathology: a systematic review. Neurosci Biobehav Rev. functional connectivity in patients with remitted psychotic 2015;56:241–251. depression: a multi-centre STOP-PD study. EBioMedicine. 60. Coltheart M. The neuropsychology of delusions. Ann N Y 2018;36:446–453. Acad Sci. 2010;1191:16–26. 73. Oudega ML, van der Werf YD, Dols A, et al. Exploring 61. Picardi A, Fonzi L, Pallagrosi M, Gigantesco A, Biondi M. resting state connectivity in patients with psychotic depres- Delusional themes across affective and non-affective psych- sion. PLoS One. 2019;14(1):e0209908. oses. Front Psychiatry. 2018;9:132. 74. Østergaard SD, Rothschild AJ, Flint AJ, et al. Rating 62. Miller EK, Cohen JD. An integrative theory of prefrontal scales measuring the severity of psychotic depression. Acta cortex function. Annu Rev Neurosci. 2001;24:167–202. Psychiatr Scand. 2015;132(5):335–344. 63. Nejad AB, Fossati P, Lemogne C. Self-referential processing, 75. Chopra S, Fornito A, Francey SM, et al. Differentiating rumination, and cortical midline structures in major depres- the effect of antipsychotic medication and illness on brain sion. Front Hum Neurosci. 2013;7:666. volume reductions in first-episode psychosis: a longitudinal, 64. Fleming SM. Relating introspective accuracy to individual randomised, triple-blind, placebo-controlled MRI study. differences in brain structure. Science. 2012;336:670. Neuropsychopharmacology. 2021;46(8):1494–1501. 523