Dissecting biological heterogeneity in major depressive disorder based on neuroimaging subtypes with multi-omics data

Dissecting biological heterogeneity in major depressive disorder based on neuroimaging subtypes with multi-omics data | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Dissecting biological heterogeneity in major depressive disorder based on neuroimaging subtypes with multi-omics data Fei Wang, Lili Tang, Rui Tang, Shuai Dong, Junjie Zheng, Pengfei Zhao, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4852981/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 04 Mar, 2025 Read the published version in Translational Psychiatry → Version 1 posted 12 You are reading this latest preprint version Abstract Background The heterogeneity of Major Depressive Disorder (MDD) has been increasingly recognized, challenging traditional symptom-based diagnostics and the development of mechanism-targeted therapies. This study aims to identify neuroimaging-based MDD subtypes and dissect their predominant biological characteristics using multi-omics data. Method A total of 807 participants were included in this study, comprising 327 individuals with MDD and 480 healthy controls (HC). The amplitude of low-frequency fluctuations (ALFF), a functional neuroimaging feature, was extracted for each participant and used to identify MDD subtypes through machine learning clustering. Multi-omics data, including profiles of genetic, epigenetics, metabolomics, and pro-inflammatory cytokines, were obtained. Comparative analyses of multi-omics data were conducted between each MDD subtype and HC to explore the molecular underpinnings involved in each subtype. Results We identified three neuroimaging-based MDD subtypes, each characterized by unique ALFF pattern alterations compared to HC. Multi-omics analysis showed a strong genetic predisposition for Subtype 1, primarily enriched in neuronal development and synaptic regulation pathways. This subtype also exhibited the most severe depressive symptoms and cognitive decline compared to the other subtypes. Subtype 2 is characterized by immuno-inflammation dysregulation, supported by elevated IL-1β levels, altered epigenetic inflammatory measures, and differential metabolites correlated with IL-1β levels. No significant biological markers were identified for Subtype 3. Conclusion Our results identify neuroimaging-based MDD subtypes and delineate the distinct biological features of each subtype. This provides a proof of concept for mechanism-targeted therapy in MDD, highlighting the importance of personalized treatment approaches based on neurobiological and molecular profiles. Biological sciences/Neuroscience/Molecular neuroscience Health sciences/Diseases/Psychiatric disorders/Depression Health sciences/Pathogenesis/Clinical genetics Biological sciences/Genetics/Clinical genetics Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Over the last few decades, the recognition of Major Depressive Disorder (MDD) as a condition of inherent heterogeneity has posed significant challenges to the conventional diagnostic framework that relies on symptom clusters 1 . This heterogeneity is characterized by a significant mismatch between symptom dimensions and their biological underpinnings, where identical symptoms may arise from different biological causes, and disparate symptoms may reflect similar biological changes 2 , 3 . Such discrepancies make it necessary to develop precise diagnostic and therapeutic strategies based on the disease's underlying mechanisms. Despite extensive efforts, the identification of reliable biomarkers for MDD diagnosis remains elusive. Furthermore, the prevalent trial-and-error approach in the current diagnostic system yields disappointing outcomes, with less than one-third of MDD patients experiencing remission after initial treatments 4 – 6 . About one-third of patients do not achieve clinical recovery even after undergoing 12-week courses of four distinct antidepressants over a year 7 . The intricate heterogeneity of MDD plays a significant role in these diagnostic and treatment dilemmas. Strategically redefining MDD into biologically homogeneous subtypes, guided by objective biomarkers, could pave the way for more accurate and effective precision diagnostics and treatments. With the rapid advancement in biotechnologies, researches into MDD adopt multiple biological characteristics, including neuroimaging 8 – 10 , electroencephalography (EEG) 11 , 12 , and multi-omics molecular profiles 13 – 15 , to delineate biological subtypes. Neuroimaging stands out as a pivotal intermediate phenotype bridging the gap between etiology and behavioral manifestations, 16 , 17 . Studies have employed functional, structural, and diffusion tensor imaging to categorize MDD patients into subgroups with homogenous neuroimaging patterns 18 , 19 , providing significant insights into the categorization and therapeutic management of MDD. The Amplitude of Low-Frequency Fluctuations (ALFF) is a neuroimaging metric for measuring local spontaneous neuronal activity during rest. It has been validated for high test-retest reliability, establishing its utility as a regional functional measure to discern individual differences 20 , 21 . In prior research, we observed alterations in the ALFF metric within specific brain areas, including the prefrontal cortex, occipital lobe, hippocampus, and amygdala of MDD patients, facilitating the delineation of subtypes in MDD 22 , 23 . Despite prior studies having classified MDD into relatively homogeneous subtypes using neuroimaging features, limited researches have verified their underlying biological mechanisms at the multi-omics level. Most studies focused solely on comparing their symptomatology, or indirect verification through publicly available databases. It is noteworthy that existing etiological hypotheses, such as the monoamine hypothesis 24 , polygenic hypothesis, immuno-inflammatory hypothesis 25 , mitochondrial dysfunction 26 , and metabolic dysregulation 27 , offer diverse insights into the biological underpinnings of MDD. However, none of these hypotheses can fully explain the pathological mechanism of MDD 28 , supporting the biological heterogeneity of MDD. Different subtypes may have stronger associations with particular etiological theories. Defining biologically homogeneous subtypes through neuroimaging characteristics and corroborating their biological relevance with omics data from the same neuroimaging cohort will help elucidate the leading etiological mechanisms of each subtype, offering vital directions for precision medicine of MDD. This study aims to identify subtypes of MDD based on ALFF and investigate the biological underpinnings using multi-omics data. In the present study, we recruited 327 MDD patients and 480 healthy controls (HC) and clustered MDD patients into three subtypes based on ALFF patterns. Subsequently, we conducted comparative analyses across various molecular levels, including pro-inflammatory cytokine, epigenetics, metabolomics, and genetics profiles, to delineate the predominant molecular alterations of each subtype. Materials and methods Participants A total of 807 participants were included in this study, including 327 MDD patients and 480 healthy controls (HC). Patients with MDD were recruited from both inpatient and outpatient services at the Department of Psychiatry within the First Affiliated Hospital of China Medical University and Shenyang Mental Health Center. HC were recruited through local community advertisements and had no personal or family history of psychiatric disorders. All participants were independently evaluated by two trained psychiatrists. For participants aged 18 years and older, the Structured Clinical Interview for Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) Axis I Disorders was employed. Participants under 18 years old were diagnosed using the Schedule for Affective Disorders and Schizophrenia for School-Age Children-Present and Lifetime version (K-SADS-PL). Exclusion criteria for participation included: 1) general contraindications for MRI, 2) a history of substance or alcohol abuse/dependence, 3) head trauma with loss of consciousness lasting at least 5 minutes or any neurological disorder, and 4) any major concurrent medical disorder. The study was approved by the Medical Research Ethics Committee of China Medical University in accordance with the Declaration of Helsinki. Written informed consent was received from each subject or their legal guardian for young participants under 18 years old. Clinical assessment The Hamilton Depression Scale (HAMD) and the Hamilton Anxiety Scale (HAMA) were used to evaluate the severity of depressive and anxiety symptoms in all participants. Cognitive function was assessed by a computerized version of the Wisconsin Card Sorting Test (WCST), which provides five indices as follows: correct responses (CR), completed categories (CC), total errors (TE), perseverative errors (PE), and non-perseverative errors (NPE). Neuroimaging data acquisition and preprocessing MRI scans were acquired with a 3.0T GE Sigma system (Sigma EXCITE HDx, GE Healthcare, USA) with a standard 8-channel head coil at the First Affiliated Hospital of the China Medical University, Shenyang, China. A restraining foam pad was utilized to minimize head movement. Participants received explicit instructions to remain in a relaxed state, keep their eyes closed, and refrain from any movement or falling asleep throughout the scanning. The functional images were performed using a gradient echo planar imaging (EPI) sequence, which was as following parameters: repetition time = 2000 ms, echo time = 30 ms, flip angle = 90°, field of view = 240 × 240 mm2, and matrix = 64 × 64. A total of 35 slices were acquired, each with a thickness/gap of 3 mm/0 mm. The scanning session lasted for 6min and 40s. The images underwent processing and analysis employing two toolkits: Statistical Parametric Mapping 8 (SPM8, http://www.fil.ion.ucl.ac.uk/spm ) and Data Processing Assistant for R-functional MRI-fMRI (DPARSF, http://www.restfmri.net/forum/DPARSF ). The specific parameters and processing of MRI data (Supplementary Methods S1) were consistent with our prior publications 29 – 31 . Identification of neuroimaging subtypes For neuroimaging data, a total of 47,636 voxel-level ALFF values were extracted for each participant using the AAL90 atlas. We utilized the singular value decomposition (SVD) algorithm to handle the high-dimensional ALFF data 32 , resulting in 30 dimensions. Subsequently, we employed the K-means algorithm 33 to cluster MDD patients into three subtypes based on the 30-dimensional data. The stability of clustering results was evaluated using the Consensus Clustering method 34 . These machine-learning algorithms were implemented using the scikit-learn library (version 0.22.2.post1) 35 . Measure of pro-inflammatory cytokines Levels of three common pro-inflammatory cytokines (TNF-α, IL-6, and IL-1β) were measured using the Human Premixed Multi-Analyte Kit (R&D Systems, Inc., Minneapolis, MN, United States) with the Human Magnetic Luminex Assay (Leptin [BR51]). Samples were magnetically labeled using a human, magnetic, premixed, microparticle cocktail of antibodies (Kit Lot Number L120614). Multi-omics data acquisition Peripheral blood samples were collected from participants for multi-omics profiling. Genomic, epigenomic, and metabolomic profiles were measured in this study. To obtain genomic data, we genotyped DNA samples following standard procedures, employing the Illumina Global Screening Chip-24 v1.0 BeadChip for Han Chinese populations. This BeadChip comprises 642,824 predetermined genetic variants and 53,411 custom-designed mutation sites. Epigenomic profiles were derived using the Illumina Infinium Methylation EPIC BeadChip 36 , facilitating the quantitative analysis of over 850,000 methylation sites across the genome with single-nucleotide precision. For the metabolomic profile, untargeted metabolomics and lipid-omics analyses were conducted on the Ultimate 3000 Ultra Performance Liquid Chromatography coupled with the Q Interactive Quadrupole-Orbitrap High Resolution Mass Spectrometer (UPLC-HRMS) 37 . Detailed information could be check in our prior publications 29 , 38 . Calculation of epigenetic inflammation score (EIS) We utilized the ChAMP R package 39 to load the raw DNA methylation data, and filter probes and samples by default parameters (Supplementary Methods S2). Then the probes were normalized and corrected for technical batch effects and cell heterogeneity. All DNA methylation levels were expressed as β values, ranging from 0 to 1, calculated as M/(M + U), where M is the signal from methylated beads, and U is the signal from unmethylated beads at the targeted CpG site. The calculation of epigenetic inflammation scores (EIS) is informed by methodologies established in prior research 40 . Specifically, we utilized findings from both the discovery and validation phases of a large-scale meta-analysis, involving 9 cohorts (n = 8,863) and 4 cohorts (n = 4,111) respectively, to select seven methylation sites significantly associated with plasma C-reactive protein (CRP) levels 40 . Due to the exclusion of one CpG site (cg06126421) from our dataset during quality control, our analysis proceeded with the six remaining CpG sites. The calculation involved multiplying the beta values of the six CpG sites by their respective regression weights and summing them to generate a composite score for each participant 41 . Given that all regression weights from the EWAS were negative, a lower EIS score is indicative of a higher state of inflammation, thus establishing a direct correlation between epigenetic markers and inflammation levels. Estimation of epigenetic immune cell proportions We utilized the reference-based GLINT method to assess the proportions of various immune cell types within each sample, relying on their methylation profiles 42 . This method estimates the proportion of monocytes, CD8 + cells, CD4 + cells, NK cells, B cells, neutrophils, and eosinophils according to the Houseman model, using a panel of 300 highly informative methylation sites in blood 43 and reference data collected from sorted blood cells 44 . It provides relative data of cell counts rather than absolute counts. Identification of differential metabolites (DMs) Following standard quality control procedures, a total of 669 metabolites were quantified. We applied linear regression modeling to identify the differential metabolites (DMs) associated with each subtype, incorporating 'group' as the independent variable and the relative abundance of each metabolite as the dependent variable, while adjusting for age, gender, BMI, and medication status as covariates. To account for multiple testing, we applied the Benjamini-Hochberg False Discovery Rate (FDR) correction, considering adjusted p-values values below 0.05 as statistically significant 45 . Calculation of polygenic risk scores (PRS) Quality control of genomic data and imputation were performed before the calculation of polygenic risk scores (PRS). The process included the exclusion of single nucleotide polymorphisms (SNPs) with a minor allele frequency (MAF) below 1%, a call rate less than 95%, or deviation from Hardy-Weinberg equilibrium (P-value < 10 − 5 ), and the exclusion of participants with more than 5% missing data, gender mismatches, or an identity-by-descent (IBD) score greater than 0.9. The refined genotype data were imputed using the GenoImpute engine 46 . The international MDD GWAS results published by the Psychiatric Genomics Consortium were used as discovery samples ( https://pgc.unc.edu/ ), and our imputed genotyping data were used as a target sample. In the paper by the Major depressive disorder Working Group of the Psychiatric Genomics Consortium, the specific genetic factors contributing to MDD were analysed in 135,458 people with MDD and 344,901 controls 47 . PRSs were generated using PRSice software ( www.PRSice.info ). P-value-informed clumping was performed with a cut-off of r 2 = 0.1 in a 250 kb window. Seven PRSs at different P-value thresholds (10 − 8 , 10 − 7 , 10 − 6 , 10 − 5 , 10 − 4 , 0.001, 0.01) were derived for each study participant. Statistical Analysis We performed a descriptive analysis to evaluate the demographic and clinical characteristics of the participants. The normality of continuous variables was determined using the Shapiro-Wilk test. These variables are presented as means ± standard deviations (SD), while categorical data are expressed as frequencies and percentages. To investigate differences among subtypes, ANOVA analyses were conducted for continuous variables, and chi-square tests were utilized for categorical variables. In the evaluation of clinical measures, encompassing both clinical symptoms and cognitive performance, we employed ANOVA to assess statistical differences among the identified subtypes and HC. This was followed by post-hoc comparisons using Tukey's Honestly Significant Difference test to discern specific pairwise differences between the groups. A threshold of p < 0.05 was set to determine statistical significance. To delineate the neuroimaging distinctions between subtypes and HC, we conducted a voxel-based two-sample t-test using the DPABI tool. Age and sex were included as covariates to mitigate potential biases, with statistical significance thresholds set at voxel P-value of < 0.001 and cluster P-value of < 0.05, corrected using the Gaussian random field (GRF) method. Differences in pro-inflammatory cytokines, EIS, and proportions of seven immune cell types among the subtypes and HC were evaluated using ANOVA analysis respectively. To identify specific pairwise group differences, a post-hoc analysis was performed using Tukey's Honestly Significant Difference test. A p-value of less than 0.05 was considered to denote a statistically significant difference. Pearson correlation analysis was used to investigate associations between DMs and pro-inflammatory cytokines. The association of PRS-MDD with each subtype was investigated using logistic regression embedded in the software PRSice, adjusted for the first 20 principal components, age, and gender. Nagelkerke’s pseudo-R2 was calculated to measure the proportion of variance explained. To explore the biological pathways associated with these genetic variants, we conducted gene ontology (GO) enrichment analysis on the genes mapped to the significant PRS. This analysis, performed using the "clusterProfiler" package in R 48 , covered three categories: cellular components (CC), molecular functions (MF), and biological processes (BP). To account for multiple testing, we applied the Benjamini-Hochberg FDR correction, considering p-adjusted values below 0.05 as statistically significant 45 . We visually presented the top 10 significant pathways within each category to clearly represent the results. Results Characteristics of participants Demographic and clinical characteristics of 327 MDD patients and 480 HC are presented in Table 1 . There were significant differences between MDD patients and HC in age, gender, BMI, and education level. In clinical assessment, WCST scores (CC, CR, TE, PE, NPE), HAMD-17, and HAMA scores were significantly higher in MDD patients than in HC. Table 1 Demographic and clinical differences between MDD subtypes and HC. MDD patients (N = 327) Health controls (N = 480) P value # M-W U test Subtype 1 (N = 93) Subtype 2 (N = 93) Subtype 3 (N = 141) P value ## ANCOVA/χ2 F/χ2 Post hoc Turkey test Demographics Gender, female (%) 67.9 59.8 0.02 73.1 55.9 72.3 HC Age, mean (S.D.), years 27.6 (10.6) 30.9 (11.8) < 0.001 30.4 (10.6) 24.3 (9.7) 27.9 (10.6) 0.004 10.6 Subtype 2 < Subtype 1,HC; Subtype 3 < HC Education, mean (S.D.), years 12.3 (3.3) 14.1 (3.7) < 0.001 12.4 (3.2) 12.2 (2.9) 12.4 (3.6) < 0.001 15.5 Subtype 1, Subtype 2, Subtype 3 < HC BMI, mean (S.D.) 21.9 (3.8) 22.6 (3.9) 0.02 21.7 (3.6) 22.9 (4.1) 21.5 (3.6) 0.009 3.9 Subtype 3 Subtype 2 Duration, mean (S.D.), months 21.9 (41.8) — 20.3 (31.4) 24.4 (36.1) 22.6 (51.1) 0.838 0.2 Anti-depressants use Any (%) 23.5 0 47.3 63.4 61.0 0.050 3.0 SSRI (%) 13.3 0 30.1 35.5 32.6 0.543 0.6 TCA (%) 0 0 0 0 0 — — MAO (%) 0 0 0 0 0 — — Others (%) 17.6 0 36.6 44.1 47.5 0.484 0.7 Clinical symptoms HAMD total score, mean (S.D.) 18.5 (9.9) 1.3 (2.3) < 0.001 21.1 (10.1) 18.1 (10.0) 17.4 (9.6) HC, Subtype 2, Subtype 3; Subtype 2, Subtype 3 > HC HAMA total score, mean (S.D.) 16.3 (10.4) 1.2 (2.5) < 0.001 17.4 (10.3) 17.8 (10.2) 16.0 (11.1) HC Cognition performance (WCST) N = 253 N = 383 N = 62 N = 76 N = 115 CR score of WCST, mean (S.D.) 28.5 (10.9) 30.2 (12.3) 0.008 24.9 (10.8) 30.0 (10.1) 29.4 (11.2) 0.013 3.6 Subtype 1 < Subtype 2 CC score of WCST, mean (S.D.) 3.6 (1.9) 3.9 (2.3) 0.052 3.2 (1.9) 3.9 (1.9) 3.7 (2.0) 0.099 2.1 TE score of WCST, mean (S.D.) 19.5 (10.8) 17.8 (12.4) 0.01 23.1 (10.8) 17.9 (10.1) 18.6 (10.8) 0.012 3.7 Subtype 1 > Subtype 2 PE score of WCST, mean (S.D.) 7.1 (6.8) 6.4 (7.0) 0.014 9.4 (8.2) 6.5 (6.7) 6.3 (5.9) 0.013 3.6 Subtype 1 > Subtype 3 NPE score of WCST, mean (S.D.) 12.3 (6.7) 11.4 (7.2) 0.027 13.7 (6.3) 11.4 (5.9) 12.2 (7.2) 0.109 2.0 Data are presented as either number (%) or means (standard deviations). SZ, schizophrenia; BD, bipolar disorder; MDD, major depressive disorder; HC, healthy control; HAMD, Hamilton Depression Scale; HAMA, Hamilton Anxiety Scale; WCST, Wisconsin Card Sorting Test. FD, framewise-displacement. N/A, not available/not applicable. # The examination between the MDD patients and HC groups. ## The examination among the subtypes and HC groups. Identification of neuroimaging-based subtypes and their ALFF patterns 327 MDD patients were clustered into three subtypes based on ALFF data using the K-means algorithm: Subtype 1 (44%), Subtype 2 (28%), and Subtype 3 (28%). Each subtype exhibited unique functional imbalance patterns in ALFF. In Subtype 1, ALFF was significantly elevated in limbic areas including the hippocampus, cingulate, amygdala, and thalamus, but notably decreased in the primary cortices such as the occipital and postcentral cortex, in comparison to HC as depicted in Fig. 1 . Subtype 2 showed a marked increase in ALFF within the frontal cortex and a decrease in the primary sensory cortex relative to HC. Conversely, Subtype 3 demonstrated a significant reduction in ALFF in the frontal cortex and an increase in the primary sensory cortex compared to HC. Detailed information about these region-specific alterations, aligned with the Automated Anatomical Labeling (AAL) atlas, is available in the Supplementary Results S1. Demographic and clinical characteristics of the subtypes In the comparison of demographic characteristics across the three identified subtypes and HC, significant differences were observed in gender, age, education and BMI level among the four groups (Table 1 ). The results of subsequent post-hoc analyses are presented in Table 1 . Regarding clinical symptoms, Post-hoc comparisons revealed that Subtype 1 had significantly higher HAMD-17 scores than both Subtype 2 and Subtype 3. Subtype 1 also demonstrated more pronounced cognitive impairments in WCST than both Subtype 2 and Subtype 3. Specifically, Subtype 1 exhibited significantly lower CR scores and higher TE scores compared to Subtype 2. Subtype 1 had higher PE scores than Subtype 3. Detailed demographic and clinical statistical results are presented in Table 1 . Differences of Subtypes in molecular levels Significant alteration of pro-inflammatory cytokine in Subtype 2 We compared the differences in pro-inflammatory cytokines among three subtypes and HC using ANCOVA and post-hoc analysis. We found that IL-1β in Subtype 2 was more significantly elevated than HC (Fig. 2 ). TNF-alpha and IL-6 presented a trend of increased levels in Subtype 2, but did not reach statistical significance. This result indicated that Subtype 2 was in a higher inflammatory state than Subtype 1 and Subtype 2. Next, we further illuminate the inflammatory characteristics of Subtype 2 with multiple omics data. Significant changes of inflammation indicators in methylation level underlying Subtype 2 Epigenetic Inflammation Scores (EIS) are refined indicators of chronic low-grade inflammation 49 . Distinct from traditional markers like plasma CRP levels, which are prone to significant variations in response to acute inflammatory stimuli, EIS provides a steadier reflection of chronic inflammation 50 , 51 . The EIS, in particular, showcases enhanced test-retest reliability when compared to serum CRP levels, underscoring its value in offering a more reliable measure of inflammation's temporal stability 41 . Subtype 2 exhibited a significantly lower EIS compared to HC, indicating a higher state of inflammation (Fig. 3 A). Subtype 1 and Subtype 3 did not show significant differences in EIS, as illustrated in Fig. 3 A. In the analysis of proportions of seven immune cell types derived from methylation data, Subtype 2 demonstrated a significantly higher proportion of neutrophils relative to HC, while the other six immune cell types showed no statistical significance (Fig. 3 B- 3 H). Subtype 1 and Subtype 3 did not show significant differences in the proportions of the seven immune cell types. Neutrophils are responsible for the first line of the host immune response against invading pathogens. They employ various mechanisms, including chemotaxis, phagocytosis, the release of reactive oxygen species (ROS) and granular proteins, and the production and release of cytokines, all of which contribute to inflammation. These findings suggest that Subtype 2 is primarily associated with increased inflammation. Significant metabolomic alterations and the correlation with IL-1β level underlying Subtype 2 In the analysis of DMs, we observed significant metabolic disturbances in Subtype 2. Compared to HC, Subtype 2 exhibited 36 DMs, while Subtype 1 and Subtype 3 did not show any DMs (Fig. 4 A- 4 C). In 36 DMs, 9 fatty acids, 8 triglycerides, 1 peptide, and 1 microbiome metabolite showed higher abundance, while 7 organic acids, 6 amino acids, 1 carbohydrate, and 3 unclassified metabolites showed decreased abundance in Subtype 2 compared to HC (Fig. 4 D, 4 E). Pearson correlation analysis was used to further test the associations of DMs with IL-1β level, which was significantly higher in Subtype 2. 44% (16/36) of DMs significantly correlated with IL-1β levels in Subtype 2. Specifically, five of six amino acids - N-Acetylleucine, Acetyl-N-formyl-5-methoxykynurenamine, N-Acetylisoleucine, N-Lactoylvaline, and Methylglutaconic acid - were negatively correlated with plasma IL-1β. Similarly, six of the seven organic acids, including Fumaric acid, Isobutyric acid, Glutaric acid, Succinic acid semialdehyde, Dihydroxybutanoic acid, and Furoic acid, demonstrated significant negative associations with IL-1β (Fig. 4 F). Significant genetic predispositions underlying Subtype 1 PRS analysis revealed variation in genetic predispositions across the three MDD subtypes (Fig. 5 ). Only Subtype 1 demonstrated significant genetic predisposition associated with MDD at PTs of 0.01 and 0.001 after adjusting for multiple comparisons. Subtype 2 and Subtype 3 did not exhibit a genetic liability related to MDDs. Specifically, PRS-MDD at PT of 0.001 (P = 0.011, NSNPs = 1325) and 0.01 (P = 0.002, NSNPs = 4769) accounted for 5.9% and 8.1% of the phenotypic variance in Subtype 1, respectively. 4769 SNPs of PRS-MDD at the most significance threshold of PT = 0.01 were mapped to 2169 genes and found to be significantly enriched in pathways related to neuronal development and synaptic regulation. The top ten Gene Ontology (GO) terms within each category were presented in Fig. 5 , mainly involved in 'synapse organization', 'regulation of neuron projection development', 'axonogenesis', 'postsynaptic membrane', 'neuron to neuron synapse', and 'monoatomic ion gated channel activity'. Discussion In this research, the utilization of machine learning clustering techniques on functional neuroimaging features (ALFF) in MDD patients identified three distinct subtypes—Subtype 1, Subtype 2, and Subtype 3. Each subtype was characterized by unique ALFF patterns. These distinctions were further substantiated through comprehensive multi-omics biological profiling, enriching our understanding of the specific etiological factors underlying each subtype. Subtype 1 features an imbalance of brain activity between the limbic system and primary cortices, with increased ALFF in the hippocampus, cingulate, and amygdala, and decreased ALFF in the primary visual, sensory, and motor cortices. This subtype indicates a strong genetic predisposition toward MDD, primarily enriched in neuronal development and synaptic regulation pathways, accompanied by severe depressive symptoms and cognitive decline. Subtype 2 reveals a different pattern of ALFF imbalance, with increased activity in the higher-order cortices, notably the prefrontal cortex, and decreased activity in the primary cortices. Immune-inflammation dysregulation existed in Subtype 2, supported by multidimensional molecular evidence, including elevated IL-1β level, altered epigenetic inflammatory measures, and differential metabolites correlated with IL-1β level. Contrary to Subtype 2 in the ALFF pattern, Subtype 3 presents decreased activity in the prefrontal cortex and increased activity in primary cortices. There were no significant biological markers identified in Subtype 3. Our study showed that three subtypes may exhibit unique etiological mechanisms, with Subtype 1 predominantly influenced by genetic factors involved in neuronal development and synaptic regulation and Subtype 2 linked to immune-inflammatory processes. The variability of multidimensional features of subtypes reveals the complexity of MDD and highlights the potential of using neuroimaging-based subtypes to forge a path towards precision therapy in MDD. Interestingly, all three subtypes demonstrate alterations in the primary sensory cortices, emphasizing their significance in MDD. Traditionally, the higher-order cortices and limbic system, known for their roles in cognitive and emotional regulation, have been the focus of the studies of neural mechanisms of MDD 52 , 53 . The primary sensory cortices, tasked mainly with sensory input processing, have received less attention in MDD research. However, recent findings highlight the involvement of primary sensory cortices in higher-order functions, especially the visual cortex, suggesting its potential as a neuroregulatory target for alleviating depressive symptoms and improving cognitive function 54 . Our initial results from the MAM animal model, which revealed an imbalanced functional neuroimaging pattern between higher-order and primary cortices, showed that abnormal activity in the primary cortex has begun in adolescence, preceding abnormalities in the higher-order cortices that have appeared in early adulthood 55 . Studies by Jiang et al. on schizophrenia (SZ) patients examined the evolution of functional imaging features throughout the disease progression, revealing a gradual shift from the primary to higher-order cortices and subcortical areas as the disease advances 56 . These consistent observations indicate that changes in the primary cortex may be early indicators of psychiatric conditions like MDD and SZ, potentially making it a viable target for early intervention. Moreover, our prior animal study provided further support for these observations by demonstrating that targeted high-frequency repetitive Transcranial Magnetic Stimulation (rTMS) of the visual cortex during adolescence can reverse the abnormal functional connectivity in the higher-order cortex observed in early adulthood in a rat model of depression 57 . This outcome is further bolstered by human studies, where interventions targeting the visual cortex have proved effective for reversing abnormal neuroimaging and alleviating clinical symptoms 58 in mood disorders patients. Furthermore, one of our clinical trials showed that visual cortex stimulation significantly improved cognitive function in bipolar disorder (BD) patients 59 . These findings collectively suggest that primary cortices-targeted interventions may hold promising clinical utility. The multi-omics evidence consistently indicates that immune-inflammatory dysregulation primarily drives the pathogenesis of Subtype 2. We observed a significant elevation of the pro-inflammatory cytokine IL-1β and pronounced metabolic dysregulation in Subtype 2 compared to HC. Notably, the disturbances in fatty acid, triglyceride, and amino acid metabolism were significantly correlated with IL-1β level. The association between immune-inflammatory damage and MDD has been robustly supported by human and animal research 60 , 61 . Studies have identified increased levels of pro-inflammatory cytokines and acute response proteins in the plasma 62 , central nervous system 63 , and cerebrospinal fluid 64 of MDD patients, along with a rise in immune cells like neutrophils and monocytes 65 , 66 . Clinical observations have shown that cytokine therapy, such as using IFN-α for chronic hepatitis, can lead to depressive-like behaviors, including anhedonia, anorexia, and reduced libido 67 , 68 . Animal studies corroborate these findings, showing that inflammatory inducers like lipopolysaccharide (LPS) triggered depressive-like behaviors, such as decreased activity, appetite loss, and lowered sexual drive 69 . These findings suggest that at least a subset of MDD patients have immune-inflammatory damage. Clinical trials have demonstrated promising evidence that anti-inflammatory treatments, both standalone and adjunct, can ameliorate depressive symptoms 70 . Targeted anti-inflammatory therapy, considering the biological heterogeneity of MDD, is especially crucial for patients with immune-inflammatory profiles. The challenge lies in accurately identifying this subgroup. Recent research has extended the use of molecular markers from cardiovascular to psychiatric domains, primarily utilizing molecular markers in peripheral blood for clinical stratification, such as CRP > 3 as an indicator of chronic inflammation 6 . These approaches offer valuable insights, yet the variability and potential somatic condition confounders of traditional peripheral inflammation markers hinder their clinical applicability. Neuroimaging features, serving as a conduit between peripheral inflammation and the central nervous system 71 , 72 , emerge as promising markers for identifying inflammatory subtypes in psychiatric disorders. In our study, distinct subtypes of MDD were identified based on ALFF, of which Subtype 2 was characterized by immune-inflammatory dysregulation and anti-inflammatory therapies would offer a promising adjunctive treatment modality for this subgroup. Compared to Subtypes 1 and 2, our multi-omics analysis did not definitively reveal the pathogenic mechanisms of Subtype 3, which shows neither significant genetic risk factors nor notable changes in inflammatory markers. We speculate that the etiology and pathological mechanisms of Subtype 3 are more complex and heterogeneous. Limitation Several limitations should be considered with regard to interpreting the current findings. Firstly, our whole-genome genetics and epigenetics data are limited in sample size, thus lacking the statistical power to identify specific genes or pathways associated with each subtype. This limitation prompted us to complement our analysis with large-sample-based validated results and summarized measures from whole-genome data, such as PRS and EIS, given the well-established understanding of MDD as a polygenic disorder. Secondly, not all MDD patients are medication-free, although there is no significant variation in medication use status among the three subtypes. Notably, medication does not influence the genetic profile. To address potential confounding effects, we included medication status as a covariate in our analysis of epigenetics and metabolomics data. Lastly, our study is cross-sectional, precluding the determination of direct causal effects of molecular alterations on each subtype. Future longitudinal studies are essential to explore the presence of causal relationships. Conclusion This study leveraged functional imaging to decipher the biological heterogeneity of MDD and employed multi-omics data to identify the primary biological mechanisms underlying the homogenous imaging subtypes: Subtype 1 is driven by genetic anomalies; Subtype 2 by immune-inflammatory dysregulation; and Subtype 3 by a complex mix. Our results offer a proof of concept for mechanism-targeted therapy in MDD. Declarations Acknowledgements: Funding This work was supported by grants from NSFC-Guangdong Joint Fund [U20A6005 to Fei Wang], Jiangsu Provincial Key Research and Development Program [BE2021617 to Fei Wang], Strategic Topics Grant of University Grants Committee [STG1/M-501/23-N to FeiWang], Jiangsu Provincial Medical Innovation Team, Key Project supported by Medical Science and Technology Development Foundation, Jiangsu Commission of Health [ZD2021026 to Rongxin Zhu], Jiangsu Provincial Key Research and Development Program [BE2022160 to Rongxin Zhu], Natural Science Foundation of Jiangsu Province [BK20231126 to Rongxin Zhu], China Postdoctoral Science Foundation [2022M721681 to Junjie Zheng]. We would like to thank all participants who took part in this study. Declaration of interest ： None References Buch, A. M. & Liston, C. Dissecting diagnostic heterogeneity in depression by integrating neuroimaging and genetics. Neuropsychopharmacology 46, 156–175 (2021). https://doi.org:10.1038/s41386-020-00789-3 Lynall, M. E. & McIntosh, A. M. The Heterogeneity of Depression. Am J Psychiatry 180, 703–704 (2023). https://doi.org:10.1176/appi.ajp.20230574 Hasler, G. PATHOPHYSIOLOGY OF DEPRESSION: DO WE HAVE ANY SOLID EVIDENCE OF INTEREST TO CLINICIANS? World Psychiatry 9, 155–161 (2010). https://doi.org:10.1002/j.2051-5545.2010.tb00298.x Warden, D., Rush, A. J., Trivedi, M. H., Fava, M. & Wisniewski, S. R. The STAR*D Project results: a comprehensive review of findings. Curr Psychiatry Rep 9, 449–459 (2007). https://doi.org:10.1007/s11920-007-0061-3 Rush, A. J. et al. Acute and longer-term outcomes in depressed outpatients requiring one or several treatment steps: a STAR*D report. Am J Psychiatry 163, 1905–1917 (2006). https://doi.org:10.1176/ajp.2006.163.11.1905 Drevets, W. C., Wittenberg, G. M., Bullmore, E. T. & Manji, H. K. Immune targets for therapeutic development in depression: towards precision medicine. Nat Rev Drug Discov 21, 224–244 (2022). https://doi.org:10.1038/s41573-021-00368-1 Kennard, B. D. et al. Remission and recovery in the Treatment for Adolescents with Depression Study (TADS): acute and long-term outcomes. J Am Acad Child Adolesc Psychiatry 48, 186–195 (2009). https://doi.org:10.1097/CHI.0b013e31819176f9 Drysdale, A. T. et al. Erratum: Resting-state connectivity biomarkers define neurophysiological subtypes of depression. Nat Med 23, 264 (2017). https://doi.org:10.1038/nm0217-264d Lynch, C. J., Gunning, F. M. & Liston, C. Causes and Consequences of Diagnostic Heterogeneity in Depression: Paths to Discovering Novel Biological Depression Subtypes. Biol Psychiatry 88, 83–94 (2020). https://doi.org:10.1016/j.biopsych.2020.01.012 Beijers, L., Wardenaar, K. J., van Loo, H. M. & Schoevers, R. A. Data-driven biological subtypes of depression: systematic review of biological approaches to depression subtyping. Mol Psychiatry 24, 888–900 (2019). https://doi.org:10.1038/s41380-019-0385-5 Zhang, Y. et al. Identification of psychiatric disorder subtypes from functional connectivity patterns in resting-state electroencephalography. Nat Biomed Eng 5, 309–323 (2021). https://doi.org:10.1038/s41551-020-00614-8 Fagiolini, A. & Kupfer, D. J. Is treatment-resistant depression a unique subtype of depression? Biol Psychiatry 53, 640–648 (2003). https://doi.org:10.1016/s0006-3223(02)01670-0 Haroon, E. et al. Increased inflammation and brain glutamate define a subtype of depression with decreased regional homogeneity, impaired network integrity, and anhedonia. Transl Psychiatry 8, 189 (2018). https://doi.org:10.1038/s41398-018-0241-4 Nguyen, T. D. et al. Genetic heterogeneity and subtypes of major depression. Mol Psychiatry 27, 1667–1675 (2022). https://doi.org:10.1038/s41380-021-01413-6 Yu, C., Arcos-Burgos, M., Licinio, J. & Wong, M. L. A latent genetic subtype of major depression identified by whole-exome genotyping data in a Mexican-American cohort. Transl Psychiatry 7, e1134 (2017). https://doi.org:10.1038/tp.2017.102 Meyer-Lindenberg, A. & Weinberger, D. R. Intermediate phenotypes and genetic mechanisms of psychiatric disorders. Nat Rev Neurosci 7, 818–827 (2006). https://doi.org:10.1038/nrn1993 Pearlson, G. D. Etiologic, phenomenologic, and endophenotypic overlap of schizophrenia and bipolar disorder. Annu Rev Clin Psychol 11, 251–281 (2015). https://doi.org:10.1146/annurev-clinpsy-032814-112915 Chen, D. et al. Neurophysiological stratification of major depressive disorder by distinct trajectories. Nature Mental Health 1, 863–875 (2023). https://doi.org:10.1038/s44220-023-00139-4 Sun, X. et al. Mapping Neurophysiological Subtypes of Major Depressive Disorder Using Normative Models of the Functional Connectome. Biol Psychiatry 94, 936–947 (2023). https://doi.org:10.1016/j.biopsych.2023.05.021 Küblböck, M. et al. Stability of low-frequency fluctuation amplitudes in prolonged resting-state fMRI. Neuroimage 103, 249–257 (2014). https://doi.org:10.1016/j.neuroimage.2014.09.038 Liu, J. et al. Alterations in amplitude of low frequency fluctuation in treatment-naïve major depressive disorder measured with resting-state fMRI. Hum Brain Mapp 35, 4979–4988 (2014). https://doi.org:10.1002/hbm.22526 Gong, J. et al. Common and distinct patterns of intrinsic brain activity alterations in major depression and bipolar disorder: voxel-based meta-analysis. Transl Psychiatry 10, 353 (2020). https://doi.org:10.1038/s41398-020-01036-5 Teng, C. et al. Abnormal resting state activity of left middle occipital gyrus and its functional connectivity in female patients with major depressive disorder. BMC Psychiatry 18, 370 (2018). https://doi.org:10.1186/s12888-018-1955-9 Meyer, J. H. et al. Elevated monoamine oxidase a levels in the brain: an explanation for the monoamine imbalance of major depression. Arch Gen Psychiatry 63, 1209–1216 (2006). https://doi.org:10.1001/archpsyc.63.11.1209 Kiecolt-Glaser, J. K., Derry, H. M. & Fagundes, C. P. Inflammation: depression fans the flames and feasts on the heat. Am J Psychiatry 172, 1075–1091 (2015). https://doi.org:10.1176/appi.ajp.2015.15020152 Scaini, G. et al. Dysregulation of mitochondrial dynamics, mitophagy and apoptosis in major depressive disorder: Does inflammation play a role? Mol Psychiatry 27, 1095–1102 (2022). https://doi.org:10.1038/s41380-021-01312-w Amin, N. et al. Interplay of Metabolome and Gut Microbiome in Individuals With Major Depressive Disorder vs Control Individuals. JAMA Psychiatry 80, 597–609 (2023). https://doi.org:10.1001/jamapsychiatry.2023.0685 Hasler, G. Pathophysiology of depression: do we have any solid evidence of interest to clinicians? World Psychiatry 9, 155–161 (2010). https://doi.org:10.1002/j.2051-5545.2010.tb00298.x Chang, M. et al. Identifying and validating subtypes within major psychiatric disorders based on frontal-posterior functional imbalance via deep learning. Mol Psychiatry 26, 2991–3002 (2021). https://doi.org:10.1038/s41380-020-00892-3 Guo, H. et al. Brain Functional and Structural Alterations in Women With Bipolar Disorder and Suicidality. Front Psychiatry 12, 630849 (2021). https://doi.org:10.3389/fpsyt.2021.630849 Duan, J. et al. Neurodevelopmental trajectories, polygenic risk, and lipometabolism in vulnerability and resilience to schizophrenia. BMC Psychiatry 23, 153 (2023). https://doi.org:10.1186/s12888-023-04597-z Worsley, K. J., Chen, J. I., Lerch, J. & Evans, A. C. Comparing functional connectivity via thresholding correlations and singular value decomposition. Philos Trans R Soc Lond B Biol Sci 360, 913–920 (2005). https://doi.org:10.1098/rstb.2005.1637 MacQueen, J. Monti, S., Tamayo, P., Mesirov, J. & Golub, T. Consensus Clustering: A Resampling-Based Method for Class Discovery and Visualization of Gene Expression Microarray Data. Machine Learning 52, 91–118 (2003). https://doi.org:10.1023/A:1023949509487 Pedregosa, F. et al. Scikit-learn: Machine Learning in Python. ArXiv abs/1201.0490 (2011). Pidsley, R. et al. Critical evaluation of the Illumina MethylationEPIC BeadChip microarray for whole-genome DNA methylation profiling. Genome Biol 17, 208 (2016). https://doi.org:10.1186/s13059-016-1066-1 Khan, N., Bano, A., Rahman, M. A., Rathinasabapathi, B. & Babar, M. A. UPLC-HRMS-based untargeted metabolic profiling reveals changes in chickpea (Cicer arietinum) metabolome following long-term drought stress. Plant Cell Environ 42, 115–132 (2019). https://doi.org:10.1111/pce.13195 Zheng, J. et al. Integrative omics analysis reveals epigenomic and transcriptomic signatures underlying brain structural deficits in major depressive disorder. Transl Psychiatry 14, 17 (2024). https://doi.org:10.1038/s41398-023-02724-8 Tian, Y. et al. ChAMP: updated methylation analysis pipeline for Illumina BeadChips. Bioinformatics 33, 3982–3984 (2017). https://doi.org:10.1093/bioinformatics/btx513 Ligthart, S. et al. DNA methylation signatures of chronic low-grade inflammation are associated with complex diseases. Genome Biol 17, 255 (2016). https://doi.org:10.1186/s13059-016-1119-5 Stevenson, A. J. et al. Characterisation of an inflammation-related epigenetic score and its association with cognitive ability. Clin Epigenetics 12, 113 (2020). https://doi.org:10.1186/s13148-020-00903-8 Rahmani, E. et al. GLINT: a user-friendly toolset for the analysis of high-throughput DNA-methylation array data. Bioinformatics 33, 1870–1872 (2017). https://doi.org:10.1093/bioinformatics/btx059 Koestler, D. C. et al. Improving cell mixture deconvolution by identifying optimal DNA methylation libraries (IDOL). BMC Bioinformatics 17, 120 (2016). https://doi.org:10.1186/s12859-016-0943-7 Reinius, L. E. et al. Differential DNA methylation in purified human blood cells: implications for cell lineage and studies on disease susceptibility. PLoS One 7, e41361 (2012). https://doi.org:10.1371/journal.pone.0041361 Benjamini, Y. & Hochberg, Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. Journal of the Royal Statistical Society: Series B (Methodological) 57, 289–300 (2018). https://doi.org:10.1111/j.2517-6161.1995.tb02031.x Wang, Y., Lu, J., Yu, J., Gibbs, R. A. & Yu, F. An integrative variant analysis pipeline for accurate genotype/haplotype inference in population NGS data. Genome Res 23, 833–842 (2013). https://doi.org:10.1101/gr.146084.112 Wray, N. R. et al. Genome-wide association analyses identify 44 risk variants and refine the genetic architecture of major depression. Nat Genet 50, 668–681 (2018). https://doi.org:10.1038/s41588-018-0090-3 Yu, G., Wang, L. G., Han, Y. & He, Q. Y. clusterProfiler: an R package for comparing biological themes among gene clusters. Omics 16, 284–287 (2012). https://doi.org:10.1089/omi.2011.0118 Green, C. et al. Structural brain correlates of serum and epigenetic markers of inflammation in major depressive disorder. Brain Behav Immun 92, 39–48 (2021). https://doi.org:10.1016/j.bbi.2020.11.024 Barker, E. D. et al. Inflammation-related epigenetic risk and child and adolescent mental health: A prospective study from pregnancy to middle adolescence. Dev Psychopathol 30, 1145–1156 (2018). https://doi.org:10.1017/s0954579418000330 Talens, R. P. et al. Variation, patterns, and temporal stability of DNA methylation: considerations for epigenetic epidemiology. Faseb j 24, 3135–3144 (2010). https://doi.org:10.1096/fj.09-150490 Northoff, G., Wiebking, C., Feinberg, T. & Panksepp, J. The 'resting-state hypothesis' of major depressive disorder-a translational subcortical-cortical framework for a system disorder. Neurosci Biobehav Rev 35, 1929–1945 (2011). https://doi.org:10.1016/j.neubiorev.2010.12.007 Phillips, M. L. et al. Identifying predictors, moderators, and mediators of antidepressant response in major depressive disorder: neuroimaging approaches. Am J Psychiatry 172, 124–138 (2015). https://doi.org:10.1176/appi.ajp.2014.14010076 Wu, F., Lu, Q., Kong, Y. & Zhang, Z. A Comprehensive Overview of the Role of Visual Cortex Malfunction in Depressive Disorders: Opportunities and Challenges. Neurosci Bull 39, 1426–1438 (2023). https://doi.org:10.1007/s12264-023-01052-7 Sun, D. et al. Frontal-posterior functional imbalance and aberrant function developmental patterns in schizophrenia. Transl Psychiatry 11, 495 (2021). https://doi.org:10.1038/s41398-021-01617-y Jiang, S. et al. Progressive trajectories of schizophrenia across symptoms, genes, and the brain. BMC Med 21, 237 (2023). https://doi.org:10.1186/s12916-023-02935-2 Guo, H. et al. Early-Stage Repetitive Transcranial Magnetic Stimulation Altered Posterior-Anterior Cerebrum Effective Connectivity in Methylazoxymethanol Acetate Rats. Front Neurosci 15, 652715 (2021). https://doi.org:10.3389/fnins.2021.652715 Liu, J. et al. Visual cortex repetitive transcranial magnetic stimulation (rTMS) reversing neurodevelopmental impairments in adolescents with major psychiatric disorders (MPDs): A cross-species translational study. CNS Neurosci Ther 30, e14427 (2024). https://doi.org:10.1111/cns.14427 Wang, D. et al. Targeted visual cortex stimulation (TVCS): a novel neuro-navigated repetitive transcranial magnetic stimulation mode for improving cognitive function in bipolar disorder. Transl Psychiatry 13, 193 (2023). https://doi.org:10.1038/s41398-023-02498-z Berk, M. et al. So depression is an inflammatory disease, but where does the inflammation come from? BMC Med 11, 200 (2013). https://doi.org:10.1186/1741-7015-11-200 Wohleb, E. S., Franklin, T., Iwata, M. & Duman, R. S. Integrating neuroimmune systems in the neurobiology of depression. Nat Rev Neurosci 17, 497–511 (2016). https://doi.org:10.1038/nrn.2016.69 Dowlati, Y. et al. A meta-analysis of cytokines in major depression. Biol Psychiatry 67, 446–457 (2010). https://doi.org:10.1016/j.biopsych.2009.09.033 Kim, Y. K., Na, K. S., Myint, A. M. & Leonard, B. E. The role of pro-inflammatory cytokines in neuroinflammation, neurogenesis and the neuroendocrine system in major depression. Prog Neuropsychopharmacol Biol Psychiatry 64, 277–284 (2016). https://doi.org:10.1016/j.pnpbp.2015.06.008 Enache, D., Pariante, C. M. & Mondelli, V. Markers of central inflammation in major depressive disorder: A systematic review and meta-analysis of studies examining cerebrospinal fluid, positron emission tomography and post-mortem brain tissue. Brain Behav Immun 81, 24–40 (2019). https://doi.org:10.1016/j.bbi.2019.06.015 Singh, D. et al. Changes in leukocytes and CRP in different stages of major depression. J Neuroinflammation 19, 74 (2022). https://doi.org:10.1186/s12974-022-02429-7 Sørensen, N. V. et al. Immune cell composition in unipolar depression: a comprehensive systematic review and meta-analysis. Mol Psychiatry 28, 391–401 (2023). https://doi.org:10.1038/s41380-022-01905-z Capuron, L. & Dantzer, R. Cytokines and depression: the need for a new paradigm. Brain Behav Immun 17 Suppl 1, S119–124 (2003). https://doi.org:10.1016/s0889-1591(02)00078-8 Schaefer, M. et al. Prevention of interferon-alpha associated depression in psychiatric risk patients with chronic hepatitis C. J Hepatol 42, 793–798 (2005). https://doi.org:10.1016/j.jhep.2005.01.020 Dunn, A. J., Swiergiel, A. H. & de Beaurepaire, R. Cytokines as mediators of depression: what can we learn from animal studies? Neurosci Biobehav Rev 29, 891–909 (2005). https://doi.org:10.1016/j.neubiorev.2005.03.023 Su, W. J., Hu, T. & Jiang, C. L. Cool the Inflamed Brain: A Novel Anti-inflammatory Strategy for the Treatment of Major Depressive Disorder. Curr Neuropharmacol 22, 810–842 (2024). https://doi.org:10.2174/1570159x21666230809112028 Lin, K. et al. Inflammation, brain structure and cognition interrelations among individuals with differential risks for bipolar disorder. Brain Behav Immun 83, 192–199 (2020). https://doi.org:10.1016/j.bbi.2019.10.010 Goldsmith, D. R., Bekhbat, M., Mehta, N. D. & Felger, J. C. Inflammation-Related Functional and Structural Dysconnectivity as a Pathway to Psychopathology. Biol Psychiatry 93, 405–418 (2023). https://doi.org:10.1016/j.biopsych.2022.11.003 Additional Declarations The authors have declared there is NO conflict of interest to disclose Supplementary Files Supplementary.docx Cite Share Download PDF Status: Published Journal Publication published 04 Mar, 2025 Read the published version in Translational Psychiatry → Version 1 posted Editorial decision: revise 18 Sep, 2024 Review # 3 received at journal 03 Sep, 2024 Review # 2 received at journal 02 Sep, 2024 Review # 1 received at journal 01 Sep, 2024 Reviewer # 3 agreed at journal 22 Aug, 2024 Reviewer # 2 agreed at journal 21 Aug, 2024 Reviewer # 1 agreed at journal 21 Aug, 2024 Reviewers invited by journal 21 Aug, 2024 Submission checks completed at journal 08 Aug, 2024 First submitted to journal 07 Aug, 2024 Unknown event 05 Aug, 2024 Editor assigned by journal 03 Aug, 2024 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4852981","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":343372778,"identity":"daf5b1ae-58fb-4baa-ab98-f9eea9594601","order_by":0,"name":"Fei Wang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA2ElEQVRIiWNgGAWjYBAC9gbGBjCDn4GB8QCQhnLxAJ4DUDWSQIpYLVCGwQGitbA3NzDz1Nyx23wj+cGBDww2shsOENLCcxCo5diz5G030gwOzmBIMyaoxV4isYE5h+1wstmNBIPDPAyHEwnbIv8QqOXf4WTjGekfDv9h+E+EFgnGBubctsN2BhI5BoeBIUCEFp7EhsN/+w4nSJx5U3CwxyDZeCZBLezHHz6c8e2wPX97+sYHPyrsZPsIaQEBkJrEBoEEIGVAhHIYsGfgJ8b0UTAKRsEoGJEAABHpTC3jbuwCAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0002-7650-6345","institution":"The Affiliated Brain Hospital of Nanjing Medical University","correspondingAuthor":true,"prefix":"","firstName":"Fei","middleName":"","lastName":"Wang","suffix":""},{"id":343372779,"identity":"6da18ae6-0533-4a78-9df3-b53d6d6c0f75","order_by":1,"name":"Lili Tang","email":"","orcid":"","institution":"The Affiliated Brain Hospital of Nanjing Medical University","correspondingAuthor":false,"prefix":"","firstName":"Lili","middleName":"","lastName":"Tang","suffix":""},{"id":343372780,"identity":"84e367b4-039f-4fa9-99d0-68f8dec29e24","order_by":2,"name":"Rui Tang","email":"","orcid":"","institution":"The Affiliated Brain Hospital of Nanjing Medical University","correspondingAuthor":false,"prefix":"","firstName":"Rui","middleName":"","lastName":"Tang","suffix":""},{"id":343372781,"identity":"ec1bc4fe-c543-426e-991a-b60d5c25f142","order_by":3,"name":"Shuai Dong","email":"","orcid":"","institution":"The Affiliated Brain Hospital of Nanjing Medical University","correspondingAuthor":false,"prefix":"","firstName":"Shuai","middleName":"","lastName":"Dong","suffix":""},{"id":343372782,"identity":"7fc1a802-cc19-4b85-8f2e-c989bf0c8489","order_by":4,"name":"Junjie Zheng","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Junjie","middleName":"","lastName":"Zheng","suffix":""},{"id":343372783,"identity":"30087dcf-fb18-4b8d-8d0d-c27b81a300a5","order_by":5,"name":"Pengfei Zhao","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Pengfei","middleName":"","lastName":"Zhao","suffix":""},{"id":343372784,"identity":"314b3e79-40d3-484d-8157-19ec9351c5b0","order_by":6,"name":"Xizhe Zhang","email":"","orcid":"","institution":"Nanjing Medical University","correspondingAuthor":false,"prefix":"","firstName":"Xizhe","middleName":"","lastName":"Zhang","suffix":""},{"id":343372785,"identity":"5ce90d21-69ec-49ef-9f60-b71ca664bdb3","order_by":7,"name":"Rongxin Zhu","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Rongxin","middleName":"","lastName":"Zhu","suffix":""},{"id":343372786,"identity":"71f4ebb2-fd4e-4e39-83f2-73f963122c86","order_by":8,"name":"Yanqing Tang","email":"","orcid":"https://orcid.org/0000-0002-1331-4936","institution":"Shengjing Hospital of China Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yanqing","middleName":"","lastName":"Tang","suffix":""},{"id":343372787,"identity":"083cbeef-dd25-4bb3-aa1d-4ada3a8cc657","order_by":9,"name":"Xiaohong Gong","email":"","orcid":"https://orcid.org/0000-0002-9730-8815","institution":"Fudan University","correspondingAuthor":false,"prefix":"","firstName":"Xiaohong","middleName":"","lastName":"Gong","suffix":""}],"badges":[],"createdAt":"2024-08-03 11:05:17","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4852981/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4852981/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41398-025-03286-7","type":"published","date":"2025-03-04T05:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":66654226,"identity":"e9b4309b-7179-4bc9-80cf-fa06abfa450f","added_by":"auto","created_at":"2024-10-15 07:59:13","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":26188011,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDistinct functional alterations in three MDD subtypes.\u003c/strong\u003e Brain charts display significantly differed brain regions in amplitude of low-frequency fluctuations (ALFF) in Subtypes 1 (A), Subtype 2 (B), and Subtype 2 (C) compared to HC. The significance level was set to p\u0026lt;0.05 (with Gaussian random field correction). L, left. R, right. The color bar represents the t-value. The elevated bar indicates a significant increase in ALFF of brain regions, and the lower bar indicates a significant decrease in ALFF in this subtype compared to HC. Red represents increased and blue decreased ALFF.\u003c/p\u003e","description":"","filename":"Figure1.tif.png","url":"https://assets-eu.researchsquare.com/files/rs-4852981/v1/b258901bcc488b9a85722534.png"},{"id":66654221,"identity":"548b7348-c9cc-4c91-bb0e-68133ccd728c","added_by":"auto","created_at":"2024-10-15 07:59:13","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1518391,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePro-inflammatory cytokines changes in three MDD subtypes. \u003c/strong\u003eViolin plots of the changes of IL-6 (A), IL-1 beta (B), and TNF-alpha (C) between each subtype and HC. *P\u0026lt;0.05\u003c/p\u003e","description":"","filename":"Figure2.tif.png","url":"https://assets-eu.researchsquare.com/files/rs-4852981/v1/856fb568a29eb11ee39fee6e.png"},{"id":66655732,"identity":"72b8f79a-058c-4bb0-ac62-7f8c1e18f45d","added_by":"auto","created_at":"2024-10-15 08:15:13","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1215402,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDifferences in EIS and Epigenetic immune cell proportions in three MDD subtypes.\u003c/strong\u003e (A) The differences in EIS across three MDD subtypes and HC. (B-H) The differences in CD4+ (B), CD8+ (C), monocytes (D), B cells (E), NK cells (F), neutrophils (G), and eosinophils (H) cell proportions across three MDD subtypes and HC. A statistically significant distinction in EIS and neutrophil cell proportions was discerned between Subtype 2 and the HC. EIS, epigenetic inflammation score. *P\u0026lt;0.05\u003c/p\u003e","description":"","filename":"Figure3.tif.png","url":"https://assets-eu.researchsquare.com/files/rs-4852981/v1/07143397e6892d53604ffc71.png"},{"id":66654220,"identity":"57a46635-7d48-456f-abf9-4a45a2940c04","added_by":"auto","created_at":"2024-10-15 07:59:13","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":6586780,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDifferential metabolites (DMs)in three MDD subtypes and their association with IL-1 beta.\u003c/strong\u003e (A-C) Volcano plots of DMs related to MDD subtypes. (D) Nightingale Rose Chart of 8 DMs classes in Subtype 2, the different colors indicate different metabolite classes. (E) Line chart of DMs within 8 classes in Subtype 2 and HC groups. (F) Lollipop chart of the correlation coefficient between IL-1 beta and 36 DMs in Subtype 2. The x-axis denotes the 36 DMs in Subtype 2, and y-axis represents the correlation coefficient. Calculated using Pearson correlation analysis. *P\u0026lt;0.05 **P\u0026lt;0.01 ***P\u0026lt;0.001\u003c/p\u003e","description":"","filename":"Figure4.tif.png","url":"https://assets-eu.researchsquare.com/files/rs-4852981/v1/25b169f3ef3bf19e322dcf12.png"},{"id":66654684,"identity":"048909bf-f4fc-423b-9c8f-2950eac8fa13","added_by":"auto","created_at":"2024-10-15 08:07:13","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":5492109,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDifferences in genetic predispositions in three MDD subtypes.\u003c/strong\u003e (A-C) Differential genetic predispositions of three MDD subtypes. The x-axis denotes the p-value threshold, y-axis represents PRS model fit: R2 (Nagelkerke’s). The bars represent PRS calculated for MDD at seven different p-value thresholds. * Indicates statistical significance with P adj \u0026lt; 0.05, corrected for FDR multiple comparisons. (D) Top 10 enriched GO categories for genes involved in the significant p-value threshold (PT = 0.01).\u003c/p\u003e","description":"","filename":"Figure5.tif.png","url":"https://assets-eu.researchsquare.com/files/rs-4852981/v1/222466b329b17d45c793558a.png"},{"id":77678533,"identity":"1d5d5c89-97a5-4fb6-819e-3f6c3f727d6e","added_by":"auto","created_at":"2025-03-04 08:14:53","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":36979945,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4852981/v1/c491ec29-8cae-4584-8eee-33e7ae168676.pdf"},{"id":66654223,"identity":"c80a72e7-b12c-4621-88e6-60095b3e11cb","added_by":"auto","created_at":"2024-10-15 07:59:13","extension":"docx","order_by":8,"title":"","display":"","copyAsset":false,"role":"supplement","size":86262,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementary.docx","url":"https://assets-eu.researchsquare.com/files/rs-4852981/v1/2a8849d360a949ec324fb836.docx"}],"financialInterests":"The authors have declared there is \u003cb\u003eNO\u003c/b\u003e conflict of interest to disclose","formattedTitle":"Dissecting biological heterogeneity in major depressive disorder based on neuroimaging subtypes with multi-omics data","fulltext":[{"header":"Introduction","content":"\u003cp\u003eOver the last few decades, the recognition of Major Depressive Disorder (MDD) as a condition of inherent heterogeneity has posed significant challenges to the conventional diagnostic framework that relies on symptom clusters\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e. This heterogeneity is characterized by a significant mismatch between symptom dimensions and their biological underpinnings, where identical symptoms may arise from different biological causes, and disparate symptoms may reflect similar biological changes\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e,\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e. Such discrepancies make it necessary to develop precise diagnostic and therapeutic strategies based on the disease's underlying mechanisms. Despite extensive efforts, the identification of reliable biomarkers for MDD diagnosis remains elusive. Furthermore, the prevalent trial-and-error approach in the current diagnostic system yields disappointing outcomes, with less than one-third of MDD patients experiencing remission after initial treatments\u003csup\u003e\u003cspan additionalcitationids=\"CR5\" citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e. About one-third of patients do not achieve clinical recovery even after undergoing 12-week courses of four distinct antidepressants over a year\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. The intricate heterogeneity of MDD plays a significant role in these diagnostic and treatment dilemmas. Strategically redefining MDD into biologically homogeneous subtypes, guided by objective biomarkers, could pave the way for more accurate and effective precision diagnostics and treatments.\u003c/p\u003e \u003cp\u003eWith the rapid advancement in biotechnologies, researches into MDD adopt multiple biological characteristics, including neuroimaging\u003csup\u003e\u003cspan additionalcitationids=\"CR9\" citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e, electroencephalography (EEG)\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e,\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e, and multi-omics molecular profiles\u003csup\u003e\u003cspan additionalcitationids=\"CR14\" citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e, to delineate biological subtypes. Neuroimaging stands out as a pivotal intermediate phenotype bridging the gap between etiology and behavioral manifestations, \u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e,\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e. Studies have employed functional, structural, and diffusion tensor imaging to categorize MDD patients into subgroups with homogenous neuroimaging patterns\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e,\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e, providing significant insights into the categorization and therapeutic management of MDD. The Amplitude of Low-Frequency Fluctuations (ALFF) is a neuroimaging metric for measuring local spontaneous neuronal activity during rest. It has been validated for high test-retest reliability, establishing its utility as a regional functional measure to discern individual differences\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e,\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e. In prior research, we observed alterations in the ALFF metric within specific brain areas, including the prefrontal cortex, occipital lobe, hippocampus, and amygdala of MDD patients, facilitating the delineation of subtypes in MDD\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e,\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eDespite prior studies having classified MDD into relatively homogeneous subtypes using neuroimaging features, limited researches have verified their underlying biological mechanisms at the multi-omics level. Most studies focused solely on comparing their symptomatology, or indirect verification through publicly available databases. It is noteworthy that existing etiological hypotheses, such as the monoamine hypothesis\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e, polygenic hypothesis, immuno-inflammatory hypothesis\u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e, mitochondrial dysfunction\u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e, and metabolic dysregulation\u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e, offer diverse insights into the biological underpinnings of MDD. However, none of these hypotheses can fully explain the pathological mechanism of MDD\u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e, supporting the biological heterogeneity of MDD. Different subtypes may have stronger associations with particular etiological theories. Defining biologically homogeneous subtypes through neuroimaging characteristics and corroborating their biological relevance with omics data from the same neuroimaging cohort will help elucidate the leading etiological mechanisms of each subtype, offering vital directions for precision medicine of MDD.\u003c/p\u003e \u003cp\u003eThis study aims to identify subtypes of MDD based on ALFF and investigate the biological underpinnings using multi-omics data. In the present study, we recruited 327 MDD patients and 480 healthy controls (HC) and clustered MDD patients into three subtypes based on ALFF patterns. Subsequently, we conducted comparative analyses across various molecular levels, including pro-inflammatory cytokine, epigenetics, metabolomics, and genetics profiles, to delineate the predominant molecular alterations of each subtype.\u003c/p\u003e"},{"header":"Materials and methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eParticipants\u003c/h2\u003e \u003cp\u003eA total of 807 participants were included in this study, including 327 MDD patients and 480 healthy controls (HC). Patients with MDD were recruited from both inpatient and outpatient services at the Department of Psychiatry within the First Affiliated Hospital of China Medical University and Shenyang Mental Health Center. HC were recruited through local community advertisements and had no personal or family history of psychiatric disorders. All participants were independently evaluated by two trained psychiatrists. For participants aged 18 years and older, the Structured Clinical Interview for Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) Axis I Disorders was employed. Participants under 18 years old were diagnosed using the Schedule for Affective Disorders and Schizophrenia for School-Age Children-Present and Lifetime version (K-SADS-PL). Exclusion criteria for participation included: 1) general contraindications for MRI, 2) a history of substance or alcohol abuse/dependence, 3) head trauma with loss of consciousness lasting at least 5 minutes or any neurological disorder, and 4) any major concurrent medical disorder. The study was approved by the Medical Research Ethics Committee of China Medical University in accordance with the Declaration of Helsinki. Written informed consent was received from each subject or their legal guardian for young participants under 18 years old.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eClinical assessment\u003c/h3\u003e\n\u003cp\u003eThe Hamilton Depression Scale (HAMD) and the Hamilton Anxiety Scale (HAMA) were used to evaluate the severity of depressive and anxiety symptoms in all participants. Cognitive function was assessed by a computerized version of the Wisconsin Card Sorting Test (WCST), which provides five indices as follows: correct responses (CR), completed categories (CC), total errors (TE), perseverative errors (PE), and non-perseverative errors (NPE).\u003c/p\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eNeuroimaging data acquisition and preprocessing\u003c/h2\u003e \u003cp\u003eMRI scans were acquired with a 3.0T GE Sigma system (Sigma EXCITE HDx, GE Healthcare, USA) with a standard 8-channel head coil at the First Affiliated Hospital of the China Medical University, Shenyang, China. A restraining foam pad was utilized to minimize head movement. Participants received explicit instructions to remain in a relaxed state, keep their eyes closed, and refrain from any movement or falling asleep throughout the scanning. The functional images were performed using a gradient echo planar imaging (EPI) sequence, which was as following parameters: repetition time\u0026thinsp;=\u0026thinsp;2000 ms, echo time\u0026thinsp;=\u0026thinsp;30 ms, flip angle\u0026thinsp;=\u0026thinsp;90\u0026deg;, field of view\u0026thinsp;=\u0026thinsp;240 \u0026times; 240 mm2, and matrix\u0026thinsp;=\u0026thinsp;64 \u0026times; 64. A total of 35 slices were acquired, each with a thickness/gap of 3 mm/0 mm. The scanning session lasted for 6min and 40s. The images underwent processing and analysis employing two toolkits: Statistical Parametric Mapping 8 (SPM8, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.fil.ion.ucl.ac.uk/spm\u003c/span\u003e\u003cspan address=\"http://www.fil.ion.ucl.ac.uk/spm\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) and Data Processing Assistant for R-functional MRI-fMRI (DPARSF, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.restfmri.net/forum/DPARSF\u003c/span\u003e\u003cspan address=\"http://www.restfmri.net/forum/DPARSF\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). The specific parameters and processing of MRI data (Supplementary Methods S1) were consistent with our prior publications \u003csup\u003e\u003cspan additionalcitationids=\"CR30\" citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eIdentification of neuroimaging subtypes\u003c/h2\u003e \u003cp\u003eFor neuroimaging data, a total of 47,636 voxel-level ALFF values were extracted for each participant using the AAL90 atlas. We utilized the singular value decomposition (SVD) algorithm to handle the high-dimensional ALFF data\u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e, resulting in 30 dimensions. Subsequently, we employed the K-means algorithm\u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e to cluster MDD patients into three subtypes based on the 30-dimensional data. The stability of clustering results was evaluated using the Consensus Clustering method\u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e. These machine-learning algorithms were implemented using the scikit-learn library (version 0.22.2.post1)\u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eMeasure of pro-inflammatory cytokines\u003c/h2\u003e \u003cp\u003eLevels of three common pro-inflammatory cytokines (TNF-α, IL-6, and IL-1β) were measured using the Human Premixed Multi-Analyte Kit (R\u0026amp;D Systems, Inc., Minneapolis, MN, United States) with the Human Magnetic Luminex Assay (Leptin [BR51]). Samples were magnetically labeled using a human, magnetic, premixed, microparticle cocktail of antibodies (Kit Lot Number L120614).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eMulti-omics data acquisition\u003c/h2\u003e \u003cp\u003ePeripheral blood samples were collected from participants for multi-omics profiling. Genomic, epigenomic, and metabolomic profiles were measured in this study. To obtain genomic data, we genotyped DNA samples following standard procedures, employing the Illumina Global Screening Chip-24 v1.0 BeadChip for Han Chinese populations. This BeadChip comprises 642,824 predetermined genetic variants and 53,411 custom-designed mutation sites. Epigenomic profiles were derived using the Illumina Infinium Methylation EPIC BeadChip\u003csup\u003e\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e, facilitating the quantitative analysis of over 850,000 methylation sites across the genome with single-nucleotide precision. For the metabolomic profile, untargeted metabolomics and lipid-omics analyses were conducted on the Ultimate 3000 Ultra Performance Liquid Chromatography coupled with the Q Interactive Quadrupole-Orbitrap High Resolution Mass Spectrometer (UPLC-HRMS)\u003csup\u003e\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e. Detailed information could be check in our prior publications\u003csup\u003e\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e,\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eCalculation of epigenetic inflammation score (EIS)\u003c/h2\u003e \u003cp\u003eWe utilized the ChAMP R package\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e to load the raw DNA methylation data, and filter probes and samples by default parameters (Supplementary Methods S2). Then the probes were normalized and corrected for technical batch effects and cell heterogeneity. All DNA methylation levels were expressed as β values, ranging from 0 to 1, calculated as M/(M\u0026thinsp;+\u0026thinsp;U), where M is the signal from methylated beads, and U is the signal from unmethylated beads at the targeted CpG site.\u003c/p\u003e \u003cp\u003eThe calculation of epigenetic inflammation scores (EIS) is informed by methodologies established in prior research\u003csup\u003e\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e. Specifically, we utilized findings from both the discovery and validation phases of a large-scale meta-analysis, involving 9 cohorts (n\u0026thinsp;=\u0026thinsp;8,863) and 4 cohorts (n\u0026thinsp;=\u0026thinsp;4,111) respectively, to select seven methylation sites significantly associated with plasma C-reactive protein (CRP) levels\u003csup\u003e\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e. Due to the exclusion of one CpG site (cg06126421) from our dataset during quality control, our analysis proceeded with the six remaining CpG sites. The calculation involved multiplying the beta values of the six CpG sites by their respective regression weights and summing them to generate a composite score for each participant\u003csup\u003e\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e. Given that all regression weights from the EWAS were negative, a lower EIS score is indicative of a higher state of inflammation, thus establishing a direct correlation between epigenetic markers and inflammation levels.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003eEstimation of epigenetic immune cell proportions\u003c/h2\u003e \u003cp\u003eWe utilized the reference-based GLINT method to assess the proportions of various immune cell types within each sample, relying on their methylation profiles\u003csup\u003e\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e\u003c/sup\u003e. This method estimates the proportion of monocytes, CD8\u0026thinsp;+\u0026thinsp;cells, CD4\u0026thinsp;+\u0026thinsp;cells, NK cells, B cells, neutrophils, and eosinophils according to the Houseman model, using a panel of 300 highly informative methylation sites in blood\u003csup\u003e\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e and reference data collected from sorted blood cells\u003csup\u003e\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e\u003c/sup\u003e. It provides relative data of cell counts rather than absolute counts.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eIdentification of differential metabolites (DMs)\u003c/h2\u003e \u003cp\u003eFollowing standard quality control procedures, a total of 669 metabolites were quantified. We applied linear regression modeling to identify the differential metabolites (DMs) associated with each subtype, incorporating 'group' as the independent variable and the relative abundance of each metabolite as the dependent variable, while adjusting for age, gender, BMI, and medication status as covariates. To account for multiple testing, we applied the Benjamini-Hochberg False Discovery Rate (FDR) correction, considering adjusted p-values values below 0.05 as statistically significant\u003csup\u003e\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eCalculation of polygenic risk scores (PRS)\u003c/h2\u003e \u003cp\u003eQuality control of genomic data and imputation were performed before the calculation of polygenic risk scores (PRS). The process included the exclusion of single nucleotide polymorphisms (SNPs) with a minor allele frequency (MAF) below 1%, a call rate less than 95%, or deviation from Hardy-Weinberg equilibrium (P-value\u0026thinsp;\u0026lt;\u0026thinsp;10\u003csup\u003e\u0026minus;\u0026thinsp;5\u003c/sup\u003e), and the exclusion of participants with more than 5% missing data, gender mismatches, or an identity-by-descent (IBD) score greater than 0.9. The refined genotype data were imputed using the GenoImpute engine\u003csup\u003e\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e\u003c/sup\u003e. The international MDD GWAS results published by the Psychiatric Genomics Consortium were used as discovery samples (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://pgc.unc.edu/\u003c/span\u003e\u003cspan address=\"https://pgc.unc.edu/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), and our imputed genotyping data were used as a target sample. In the paper by the Major depressive disorder Working Group of the Psychiatric Genomics Consortium, the specific genetic factors contributing to MDD were analysed in 135,458 people with MDD and 344,901 controls\u003csup\u003e\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e. PRSs were generated using PRSice software (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e\u003ca href=\"http://www.fil.ion.ucl.ac.uk/spm\" target=\"_blank\"\u003ewww.PRSice.info\u003c/a\u003e\u003c/span\u003e\u003cspan address=\"http://www.PRSice.info\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). P-value-informed clumping was performed with a cut-off of r 2\u0026thinsp;=\u0026thinsp;0.1 in a 250 kb window. Seven PRSs at different P-value thresholds (10\u003csup\u003e\u0026minus;\u0026thinsp;8\u003c/sup\u003e, 10\u003csup\u003e\u0026minus;\u0026thinsp;7\u003c/sup\u003e, 10\u003csup\u003e\u0026minus;\u0026thinsp;6\u003c/sup\u003e, 10\u003csup\u003e\u0026minus;\u0026thinsp;5\u003c/sup\u003e, 10\u003csup\u003e\u0026minus;\u0026thinsp;4\u003c/sup\u003e, 0.001, 0.01) were derived for each study participant.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eStatistical Analysis\u003c/h2\u003e \u003cp\u003eWe performed a descriptive analysis to evaluate the demographic and clinical characteristics of the participants. The normality of continuous variables was determined using the Shapiro-Wilk test. These variables are presented as means\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviations (SD), while categorical data are expressed as frequencies and percentages. To investigate differences among subtypes, ANOVA analyses were conducted for continuous variables, and chi-square tests were utilized for categorical variables. In the evaluation of clinical measures, encompassing both clinical symptoms and cognitive performance, we employed ANOVA to assess statistical differences among the identified subtypes and HC. This was followed by post-hoc comparisons using Tukey's Honestly Significant Difference test to discern specific pairwise differences between the groups. A threshold of p\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was set to determine statistical significance.\u003c/p\u003e \u003cp\u003eTo delineate the neuroimaging distinctions between subtypes and HC, we conducted a voxel-based two-sample t-test using the DPABI tool. Age and sex were included as covariates to mitigate potential biases, with statistical significance thresholds set at voxel P-value of \u0026lt;\u0026thinsp;0.001 and cluster P-value of \u0026lt;\u0026thinsp;0.05, corrected using the Gaussian random field (GRF) method.\u003c/p\u003e \u003cp\u003eDifferences in pro-inflammatory cytokines, EIS, and proportions of seven immune cell types among the subtypes and HC were evaluated using ANOVA analysis respectively. To identify specific pairwise group differences, a post-hoc analysis was performed using Tukey's Honestly Significant Difference test. A p-value of less than 0.05 was considered to denote a statistically significant difference.\u003c/p\u003e \u003cp\u003ePearson correlation analysis was used to investigate associations between DMs and pro-inflammatory cytokines.\u003c/p\u003e \u003cp\u003eThe association of PRS-MDD with each subtype was investigated using logistic regression embedded in the software PRSice, adjusted for the first 20 principal components, age, and gender. Nagelkerke\u0026rsquo;s pseudo-R2 was calculated to measure the proportion of variance explained. To explore the biological pathways associated with these genetic variants, we conducted gene ontology (GO) enrichment analysis on the genes mapped to the significant PRS. This analysis, performed using the \"clusterProfiler\" package in R\u003csup\u003e\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e\u003c/sup\u003e, covered three categories: cellular components (CC), molecular functions (MF), and biological processes (BP). To account for multiple testing, we applied the Benjamini-Hochberg FDR correction, considering p-adjusted values below 0.05 as statistically significant\u003csup\u003e\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e\u003c/sup\u003e. We visually presented the top 10 significant pathways within each category to clearly represent the results.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eCharacteristics of participants\u003c/h2\u003e \u003cp\u003eDemographic and clinical characteristics of 327 MDD patients and 480 HC are presented in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. There were significant differences between MDD patients and HC in age, gender, BMI, and education level. In clinical assessment, WCST scores (CC, CR, TE, PE, NPE), HAMD-17, and HAMA scores were significantly higher in MDD patients than in HC.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eDemographic and clinical differences between MDD subtypes and HC.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"10\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMDD patients\u003c/p\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;327)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHealth controls\u003c/p\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;480)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eP value\u003csup\u003e#\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eM-W U test\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSubtype 1\u003c/p\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;93)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eSubtype 2\u003c/p\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;93)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eSubtype 3\u003c/p\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;141)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eP value\u003csup\u003e##\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eANCOVA/χ2\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003eF/χ2\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e \u003cp\u003ePost hoc\u003c/p\u003e \u003cp\u003eTurkey test\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDemographics\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGender, female (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e67.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e59.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e73.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e55.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e72.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e4.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003eSubtype 3\u0026thinsp;\u0026gt;\u0026thinsp;HC\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge, mean (S.D.), years\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e27.6 (10.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e30.9 (11.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e30.4 (10.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e24.3 (9.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e27.9 (10.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.004\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e10.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003eSubtype 2\u0026thinsp;\u0026lt;\u0026thinsp;Subtype 1,HC; Subtype 3\u0026thinsp;\u0026lt;\u0026thinsp;HC\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEducation, mean (S.D.), years\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e12.3 (3.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e14.1 (3.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e12.4 (3.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e12.2 (2.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e12.4 (3.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e15.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003eSubtype 1, Subtype 2, Subtype 3\u0026thinsp;\u0026lt;\u0026thinsp;HC\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBMI, mean (S.D.)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e21.9 (3.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e22.6 (3.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e21.7 (3.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e22.9 (4.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e21.5 (3.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.009\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e3.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003eSubtype 3\u0026thinsp;\u0026lt;\u0026thinsp;Subtype 2, HC\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eClinical characteristics\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge of onset, mean (S.D.), years\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e24.9 (10.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026mdash;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e27.9 (10.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e22.7 (9.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e24.9 (10.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.009\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e4.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003eSubtype 1\u0026thinsp;\u0026gt;\u0026thinsp;Subtype 2\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDuration, mean (S.D.), months\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e21.9 (41.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026mdash;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e20.3 (31.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e24.4 (36.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e22.6 (51.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.838\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAnti-depressants use\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAny (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e23.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e47.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e63.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e61.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.050\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e3.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSSRI (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e13.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e30.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e35.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e32.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.543\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTCA (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026mdash;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u0026mdash;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMAO (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026mdash;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u0026mdash;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOthers (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e17.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e36.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e44.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e47.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.484\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eClinical symptoms\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHAMD total score, mean (S.D.)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e18.5 (9.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.3 (2.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e21.1 (10.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e18.1 (10.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e17.4 (9.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e443.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003eSubtype 1\u0026thinsp;\u0026gt;\u0026thinsp;HC, Subtype 2, Subtype 3; Subtype 2, Subtype 3\u0026thinsp;\u0026gt;\u0026thinsp;HC\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHAMA total score, mean (S.D.)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e16.3 (10.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.2 (2.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e17.4 (10.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e17.8 (10.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e16.0 (11.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e293.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003eSubtype 1, Subtype 2, Subtype 3\u0026thinsp;\u0026gt;\u0026thinsp;HC\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCognition performance (WCST)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eN\u0026thinsp;=\u0026thinsp;253\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003eN\u0026thinsp;=\u0026thinsp;383\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003eN\u0026thinsp;=\u0026thinsp;62\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003eN\u0026thinsp;=\u0026thinsp;76\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003eN\u0026thinsp;=\u0026thinsp;115\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCR score of WCST, mean (S.D.)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e28.5 (10.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e30.2 (12.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.008\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e24.9 (10.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e30.0 (10.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e29.4 (11.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.013\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e3.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003eSubtype 1\u0026thinsp;\u0026lt;\u0026thinsp;Subtype 2\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCC score of WCST, mean (S.D.)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3.6 (1.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3.9 (2.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.052\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3.2 (1.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e3.9 (1.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e3.7 (2.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.099\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e2.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTE score of WCST, mean (S.D.)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e19.5 (10.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e17.8 (12.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e23.1 (10.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e17.9 (10.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e18.6 (10.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.012\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e3.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003eSubtype 1\u0026thinsp;\u0026gt;\u0026thinsp;Subtype 2\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePE score of WCST, mean (S.D.)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e7.1 (6.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e6.4 (7.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.014\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e9.4 (8.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e6.5 (6.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e6.3 (5.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.013\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e3.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003eSubtype 1\u0026thinsp;\u0026gt;\u0026thinsp;Subtype 3\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNPE score of WCST, mean (S.D.)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e12.3 (6.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e11.4 (7.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.027\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e13.7 (6.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e11.4 (5.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e12.2 (7.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.109\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e2.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"10\"\u003eData are presented as either number (%) or means (standard deviations). SZ, schizophrenia; BD, bipolar disorder; MDD, major depressive disorder; HC, healthy control; HAMD, Hamilton Depression Scale; HAMA, Hamilton Anxiety Scale; WCST, Wisconsin Card Sorting Test. FD, framewise-displacement. N/A, not available/not applicable.\u003c/td\u003e\u003c/tr\u003e \u003ctr\u003e\u003ctd colspan=\"10\"\u003e# The examination between the MDD patients and HC groups.\u003c/td\u003e\u003c/tr\u003e \u003ctr\u003e\u003ctd colspan=\"10\"\u003e## The examination among the subtypes and HC groups.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eIdentification of neuroimaging-based subtypes and their ALFF patterns\u003c/h2\u003e \u003cp\u003e327 MDD patients were clustered into three subtypes based on ALFF data using the K-means algorithm: Subtype 1 (44%), Subtype 2 (28%), and Subtype 3 (28%). Each subtype exhibited unique functional imbalance patterns in ALFF. In Subtype 1, ALFF was significantly elevated in limbic areas including the hippocampus, cingulate, amygdala, and thalamus, but notably decreased in the primary cortices such as the occipital and postcentral cortex, in comparison to HC as depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. Subtype 2 showed a marked increase in ALFF within the frontal cortex and a decrease in the primary sensory cortex relative to HC. Conversely, Subtype 3 demonstrated a significant reduction in ALFF in the frontal cortex and an increase in the primary sensory cortex compared to HC. Detailed information about these region-specific alterations, aligned with the Automated Anatomical Labeling (AAL) atlas, is available in the Supplementary Results S1.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eDemographic and clinical characteristics of the subtypes\u003c/h2\u003e \u003cp\u003eIn the comparison of demographic characteristics across the three identified subtypes and HC, significant differences were observed in gender, age, education and BMI level among the four groups (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The results of subsequent post-hoc analyses are presented in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. Regarding clinical symptoms, Post-hoc comparisons revealed that Subtype 1 had significantly higher HAMD-17 scores than both Subtype 2 and Subtype 3. Subtype 1 also demonstrated more pronounced cognitive impairments in WCST than both Subtype 2 and Subtype 3. Specifically, Subtype 1 exhibited significantly lower CR scores and higher TE scores compared to Subtype 2. Subtype 1 had higher PE scores than Subtype 3. Detailed demographic and clinical statistical results are presented in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eDifferences of Subtypes in molecular levels\u003c/h2\u003e \u003cdiv id=\"Sec19\" class=\"Section3\"\u003e \u003ch2\u003eSignificant alteration of pro-inflammatory cytokine in Subtype 2\u003c/h2\u003e \u003cp\u003eWe compared the differences in pro-inflammatory cytokines among three subtypes and HC using ANCOVA and post-hoc analysis. We found that IL-1β in Subtype 2 was more significantly elevated than HC (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). TNF-alpha and IL-6 presented a trend of increased levels in Subtype 2, but did not reach statistical significance. This result indicated that Subtype 2 was in a higher inflammatory state than Subtype 1 and Subtype 2. Next, we further illuminate the inflammatory characteristics of Subtype 2 with multiple omics data.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003eSignificant changes of inflammation indicators in methylation level underlying Subtype 2\u003c/h2\u003e \u003cp\u003eEpigenetic Inflammation Scores (EIS) are refined indicators of chronic low-grade inflammation\u003csup\u003e\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e\u003c/sup\u003e. Distinct from traditional markers like plasma CRP levels, which are prone to significant variations in response to acute inflammatory stimuli, EIS provides a steadier reflection of chronic inflammation\u003csup\u003e\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e,\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e\u003c/sup\u003e. The EIS, in particular, showcases enhanced test-retest reliability when compared to serum CRP levels, underscoring its value in offering a more reliable measure of inflammation's temporal stability\u003csup\u003e\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e. Subtype 2 exhibited a significantly lower EIS compared to HC, indicating a higher state of inflammation (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA). Subtype 1 and Subtype 3 did not show significant differences in EIS, as illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eIn the analysis of proportions of seven immune cell types derived from methylation data, Subtype 2 demonstrated a significantly higher proportion of neutrophils relative to HC, while the other six immune cell types showed no statistical significance (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB-\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eH). Subtype 1 and Subtype 3 did not show significant differences in the proportions of the seven immune cell types. Neutrophils are responsible for the first line of the host immune response against invading pathogens. They employ various mechanisms, including chemotaxis, phagocytosis, the release of reactive oxygen species (ROS) and granular proteins, and the production and release of cytokines, all of which contribute to inflammation. These findings suggest that Subtype 2 is primarily associated with increased inflammation.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003eSignificant metabolomic alterations and the correlation with IL-1β level underlying Subtype 2\u003c/h2\u003e \u003cp\u003eIn the analysis of DMs, we observed significant metabolic disturbances in Subtype 2. Compared to HC, Subtype 2 exhibited 36 DMs, while Subtype 1 and Subtype 3 did not show any DMs (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA- \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC). In 36 DMs, 9 fatty acids, 8 triglycerides, 1 peptide, and 1 microbiome metabolite showed higher abundance, while 7 organic acids, 6 amino acids, 1 carbohydrate, and 3 unclassified metabolites showed decreased abundance in Subtype 2 compared to HC (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eD, \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eE).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003ePearson correlation analysis was used to further test the associations of DMs with IL-1β level, which was significantly higher in Subtype 2. 44% (16/36) of DMs significantly correlated with IL-1β levels in Subtype 2. Specifically, five of six amino acids - N-Acetylleucine, Acetyl-N-formyl-5-methoxykynurenamine, N-Acetylisoleucine, N-Lactoylvaline, and Methylglutaconic acid - were negatively correlated with plasma IL-1β. Similarly, six of the seven organic acids, including Fumaric acid, Isobutyric acid, Glutaric acid, Succinic acid semialdehyde, Dihydroxybutanoic acid, and Furoic acid, demonstrated significant negative associations with IL-1β (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eF).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003eSignificant genetic predispositions underlying Subtype 1\u003c/h2\u003e \u003cp\u003ePRS analysis revealed variation in genetic predispositions across the three MDD subtypes (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e). Only Subtype 1 demonstrated significant genetic predisposition associated with MDD at PTs of 0.01 and 0.001 after adjusting for multiple comparisons. Subtype 2 and Subtype 3 did not exhibit a genetic liability related to MDDs. Specifically, PRS-MDD at PT of 0.001 (P\u0026thinsp;=\u0026thinsp;0.011, NSNPs\u0026thinsp;=\u0026thinsp;1325) and 0.01 (P\u0026thinsp;=\u0026thinsp;0.002, NSNPs\u0026thinsp;=\u0026thinsp;4769) accounted for 5.9% and 8.1% of the phenotypic variance in Subtype 1, respectively. 4769 SNPs of PRS-MDD at the most significance threshold of PT\u0026thinsp;=\u0026thinsp;0.01 were mapped to 2169 genes and found to be significantly enriched in pathways related to neuronal development and synaptic regulation. The top ten Gene Ontology (GO) terms within each category were presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e, mainly involved in 'synapse organization', 'regulation of neuron projection development', 'axonogenesis', 'postsynaptic membrane', 'neuron to neuron synapse', and 'monoatomic ion gated channel activity'.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this research, the utilization of machine learning clustering techniques on functional neuroimaging features (ALFF) in MDD patients identified three distinct subtypes\u0026mdash;Subtype 1, Subtype 2, and Subtype 3. Each subtype was characterized by unique ALFF patterns. These distinctions were further substantiated through comprehensive multi-omics biological profiling, enriching our understanding of the specific etiological factors underlying each subtype. Subtype 1 features an imbalance of brain activity between the limbic system and primary cortices, with increased ALFF in the hippocampus, cingulate, and amygdala, and decreased ALFF in the primary visual, sensory, and motor cortices. This subtype indicates a strong genetic predisposition toward MDD, primarily enriched in neuronal development and synaptic regulation pathways, accompanied by severe depressive symptoms and cognitive decline. Subtype 2 reveals a different pattern of ALFF imbalance, with increased activity in the higher-order cortices, notably the prefrontal cortex, and decreased activity in the primary cortices. Immune-inflammation dysregulation existed in Subtype 2, supported by multidimensional molecular evidence, including elevated IL-1β level, altered epigenetic inflammatory measures, and differential metabolites correlated with IL-1β level. Contrary to Subtype 2 in the ALFF pattern, Subtype 3 presents decreased activity in the prefrontal cortex and increased activity in primary cortices. There were no significant biological markers identified in Subtype 3. Our study showed that three subtypes may exhibit unique etiological mechanisms, with Subtype 1 predominantly influenced by genetic factors involved in neuronal development and synaptic regulation and Subtype 2 linked to immune-inflammatory processes. The variability of multidimensional features of subtypes reveals the complexity of MDD and highlights the potential of using neuroimaging-based subtypes to forge a path towards precision therapy in MDD.\u003c/p\u003e \u003cp\u003eInterestingly, all three subtypes demonstrate alterations in the primary sensory cortices, emphasizing their significance in MDD. Traditionally, the higher-order cortices and limbic system, known for their roles in cognitive and emotional regulation, have been the focus of the studies of neural mechanisms of MDD\u003csup\u003e\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e,\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e\u003c/sup\u003e. The primary sensory cortices, tasked mainly with sensory input processing, have received less attention in MDD research. However, recent findings highlight the involvement of primary sensory cortices in higher-order functions, especially the visual cortex, suggesting its potential as a neuroregulatory target for alleviating depressive symptoms and improving cognitive function\u003csup\u003e\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e\u003c/sup\u003e. Our initial results from the MAM animal model, which revealed an imbalanced functional neuroimaging pattern between higher-order and primary cortices, showed that abnormal activity in the primary cortex has begun in adolescence, preceding abnormalities in the higher-order cortices that have appeared in early adulthood\u003csup\u003e\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e\u003c/sup\u003e. Studies by Jiang et al. on schizophrenia (SZ) patients examined the evolution of functional imaging features throughout the disease progression, revealing a gradual shift from the primary to higher-order cortices and subcortical areas as the disease advances\u003csup\u003e\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e\u003c/sup\u003e. These consistent observations indicate that changes in the primary cortex may be early indicators of psychiatric conditions like MDD and SZ, potentially making it a viable target for early intervention. Moreover, our prior animal study provided further support for these observations by demonstrating that targeted high-frequency repetitive Transcranial Magnetic Stimulation (rTMS) of the visual cortex during adolescence can reverse the abnormal functional connectivity in the higher-order cortex observed in early adulthood in a rat model of depression\u003csup\u003e\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e\u003c/sup\u003e. This outcome is further bolstered by human studies, where interventions targeting the visual cortex have proved effective for reversing abnormal neuroimaging and alleviating clinical symptoms\u003csup\u003e\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e\u003c/sup\u003e in mood disorders patients. Furthermore, one of our clinical trials showed that visual cortex stimulation significantly improved cognitive function in bipolar disorder (BD) patients\u003csup\u003e\u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e\u003c/sup\u003e. These findings collectively suggest that primary cortices-targeted interventions may hold promising clinical utility.\u003c/p\u003e \u003cp\u003eThe multi-omics evidence consistently indicates that immune-inflammatory dysregulation primarily drives the pathogenesis of Subtype 2. We observed a significant elevation of the pro-inflammatory cytokine IL-1β and pronounced metabolic dysregulation in Subtype 2 compared to HC. Notably, the disturbances in fatty acid, triglyceride, and amino acid metabolism were significantly correlated with IL-1β level. The association between immune-inflammatory damage and MDD has been robustly supported by human and animal research\u003csup\u003e\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e,\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e\u003c/sup\u003e. Studies have identified increased levels of pro-inflammatory cytokines and acute response proteins in the plasma\u003csup\u003e\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e\u003c/sup\u003e, central nervous system\u003csup\u003e\u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e\u003c/sup\u003e, and cerebrospinal fluid\u003csup\u003e\u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e\u003c/sup\u003e of MDD patients, along with a rise in immune cells like neutrophils and monocytes\u003csup\u003e\u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e,\u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e66\u003c/span\u003e\u003c/sup\u003e. Clinical observations have shown that cytokine therapy, such as using IFN-α for chronic hepatitis, can lead to depressive-like behaviors, including anhedonia, anorexia, and reduced libido\u003csup\u003e\u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e67\u003c/span\u003e,\u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e68\u003c/span\u003e\u003c/sup\u003e. Animal studies corroborate these findings, showing that inflammatory inducers like lipopolysaccharide (LPS) triggered depressive-like behaviors, such as decreased activity, appetite loss, and lowered sexual drive\u003csup\u003e\u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e69\u003c/span\u003e\u003c/sup\u003e. These findings suggest that at least a subset of MDD patients have immune-inflammatory damage.\u003c/p\u003e \u003cp\u003eClinical trials have demonstrated promising evidence that anti-inflammatory treatments, both standalone and adjunct, can ameliorate depressive symptoms\u003csup\u003e\u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e70\u003c/span\u003e\u003c/sup\u003e. Targeted anti-inflammatory therapy, considering the biological heterogeneity of MDD, is especially crucial for patients with immune-inflammatory profiles. The challenge lies in accurately identifying this subgroup. Recent research has extended the use of molecular markers from cardiovascular to psychiatric domains, primarily utilizing molecular markers in peripheral blood for clinical stratification, such as CRP\u0026thinsp;\u0026gt;\u0026thinsp;3 as an indicator of chronic inflammation\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e. These approaches offer valuable insights, yet the variability and potential somatic condition confounders of traditional peripheral inflammation markers hinder their clinical applicability. Neuroimaging features, serving as a conduit between peripheral inflammation and the central nervous system\u003csup\u003e\u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e71\u003c/span\u003e,\u003cspan citationid=\"CR72\" class=\"CitationRef\"\u003e72\u003c/span\u003e\u003c/sup\u003e, emerge as promising markers for identifying inflammatory subtypes in psychiatric disorders. In our study, distinct subtypes of MDD were identified based on ALFF, of which Subtype 2 was characterized by immune-inflammatory dysregulation and anti-inflammatory therapies would offer a promising adjunctive treatment modality for this subgroup.\u003c/p\u003e \u003cp\u003eCompared to Subtypes 1 and 2, our multi-omics analysis did not definitively reveal the pathogenic mechanisms of Subtype 3, which shows neither significant genetic risk factors nor notable changes in inflammatory markers. We speculate that the etiology and pathological mechanisms of Subtype 3 are more complex and heterogeneous.\u003c/p\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003eLimitation\u003c/h2\u003e \u003cp\u003eSeveral limitations should be considered with regard to interpreting the current findings. Firstly, our whole-genome genetics and epigenetics data are limited in sample size, thus lacking the statistical power to identify specific genes or pathways associated with each subtype. This limitation prompted us to complement our analysis with large-sample-based validated results and summarized measures from whole-genome data, such as PRS and EIS, given the well-established understanding of MDD as a polygenic disorder. Secondly, not all MDD patients are medication-free, although there is no significant variation in medication use status among the three subtypes. Notably, medication does not influence the genetic profile. To address potential confounding effects, we included medication status as a covariate in our analysis of epigenetics and metabolomics data. Lastly, our study is cross-sectional, precluding the determination of direct causal effects of molecular alterations on each subtype. Future longitudinal studies are essential to explore the presence of causal relationships.\u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis study leveraged functional imaging to decipher the biological heterogeneity of MDD and employed multi-omics data to identify the primary biological mechanisms underlying the homogenous imaging subtypes: Subtype 1 is driven by genetic anomalies; Subtype 2 by immune-inflammatory dysregulation; and Subtype 3 by a complex mix. Our results offer a proof of concept for mechanism-targeted therapy in MDD.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by grants from NSFC-Guangdong Joint Fund [U20A6005 to Fei Wang], Jiangsu Provincial Key Research and Development Program [BE2021617 to Fei Wang],\u0026nbsp;Strategic Topics Grant of University Grants Committee [STG1/M-501/23-N to FeiWang],\u0026nbsp;Jiangsu Provincial Medical Innovation Team,\u0026nbsp;Key Project supported by Medical Science and Technology Development Foundation, Jiangsu Commission of Health [ZD2021026 to\u0026nbsp;Rongxin Zhu],\u0026nbsp;Jiangsu Provincial Key Research and Development Program [BE2022160 to\u0026nbsp;Rongxin Zhu],\u0026nbsp;Natural Science Foundation of Jiangsu Province [BK20231126 to\u0026nbsp;Rongxin Zhu],\u0026nbsp;China Postdoctoral Science Foundation [2022M721681 to Junjie Zheng]. We would like to thank all participants who took part in this study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDeclaration of interest\u003c/strong\u003e\u003cstrong\u003e：\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNone\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cbr\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eBuch, A. M. \u0026amp; Liston, C. Dissecting diagnostic heterogeneity in depression by integrating neuroimaging and genetics. Neuropsychopharmacology 46, 156\u0026ndash;175 (2021). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/s41386-020-00789-3\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/s41386-020-00789-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLynall, M. E. \u0026amp; McIntosh, A. M. The Heterogeneity of Depression. Am J Psychiatry 180, 703\u0026ndash;704 (2023). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1176/appi.ajp.20230574\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1176/appi.ajp.20230574\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHasler, G. PATHOPHYSIOLOGY OF DEPRESSION: DO WE HAVE ANY SOLID EVIDENCE OF INTEREST TO CLINICIANS? World Psychiatry 9, 155\u0026ndash;161 (2010). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1002/j.2051-5545.2010.tb00298.x\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1002/j.2051-5545.2010.tb00298.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWarden, D., Rush, A. J., Trivedi, M. H., Fava, M. \u0026amp; Wisniewski, S. R. The STAR*D Project results: a comprehensive review of findings. Curr Psychiatry Rep 9, 449\u0026ndash;459 (2007). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1007/s11920-007-0061-3\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1007/s11920-007-0061-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRush, A. J. \u003cem\u003eet al.\u003c/em\u003e Acute and longer-term outcomes in depressed outpatients requiring one or several treatment steps: a STAR*D report. Am J Psychiatry 163, 1905\u0026ndash;1917 (2006). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1176/ajp.2006.163.11.1905\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1176/ajp.2006.163.11.1905\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDrevets, W. C., Wittenberg, G. M., Bullmore, E. T. \u0026amp; Manji, H. K. Immune targets for therapeutic development in depression: towards precision medicine. Nat Rev Drug Discov 21, 224\u0026ndash;244 (2022). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/s41573-021-00368-1\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/s41573-021-00368-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKennard, B. D. \u003cem\u003eet al.\u003c/em\u003e Remission and recovery in the Treatment for Adolescents with Depression Study (TADS): acute and long-term outcomes. J Am Acad Child Adolesc Psychiatry 48, 186\u0026ndash;195 (2009). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1097/CHI.0b013e31819176f9\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1097/CHI.0b013e31819176f9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDrysdale, A. T. \u003cem\u003eet al.\u003c/em\u003e Erratum: Resting-state connectivity biomarkers define neurophysiological subtypes of depression. Nat Med 23, 264 (2017). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/nm0217-264d\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/nm0217-264d\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLynch, C. J., Gunning, F. M. \u0026amp; Liston, C. Causes and Consequences of Diagnostic Heterogeneity in Depression: Paths to Discovering Novel Biological Depression Subtypes. Biol Psychiatry 88, 83\u0026ndash;94 (2020). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1016/j.biopsych.2020.01.012\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1016/j.biopsych.2020.01.012\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBeijers, L., Wardenaar, K. J., van Loo, H. M. \u0026amp; Schoevers, R. A. Data-driven biological subtypes of depression: systematic review of biological approaches to depression subtyping. Mol Psychiatry 24, 888\u0026ndash;900 (2019). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/s41380-019-0385-5\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/s41380-019-0385-5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang, Y. \u003cem\u003eet al.\u003c/em\u003e Identification of psychiatric disorder subtypes from functional connectivity patterns in resting-state electroencephalography. Nat Biomed Eng 5, 309\u0026ndash;323 (2021). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/s41551-020-00614-8\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/s41551-020-00614-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFagiolini, A. \u0026amp; Kupfer, D. J. Is treatment-resistant depression a unique subtype of depression? Biol Psychiatry 53, 640\u0026ndash;648 (2003). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1016/s0006-3223(02)01670-0\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1016/s0006-3223(02)01670-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHaroon, E. \u003cem\u003eet al.\u003c/em\u003e Increased inflammation and brain glutamate define a subtype of depression with decreased regional homogeneity, impaired network integrity, and anhedonia. Transl Psychiatry 8, 189 (2018). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/s41398-018-0241-4\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/s41398-018-0241-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNguyen, T. D. \u003cem\u003eet al.\u003c/em\u003e Genetic heterogeneity and subtypes of major depression. Mol Psychiatry 27, 1667\u0026ndash;1675 (2022). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/s41380-021-01413-6\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/s41380-021-01413-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYu, C., Arcos-Burgos, M., Licinio, J. \u0026amp; Wong, M. L. A latent genetic subtype of major depression identified by whole-exome genotyping data in a Mexican-American cohort. Transl Psychiatry 7, e1134 (2017). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/tp.2017.102\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/tp.2017.102\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMeyer-Lindenberg, A. \u0026amp; Weinberger, D. R. Intermediate phenotypes and genetic mechanisms of psychiatric disorders. Nat Rev Neurosci 7, 818\u0026ndash;827 (2006). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/nrn1993\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/nrn1993\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePearlson, G. D. Etiologic, phenomenologic, and endophenotypic overlap of schizophrenia and bipolar disorder. Annu Rev Clin Psychol 11, 251\u0026ndash;281 (2015). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1146/annurev-clinpsy-032814-112915\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1146/annurev-clinpsy-032814-112915\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen, D. \u003cem\u003eet al.\u003c/em\u003e Neurophysiological stratification of major depressive disorder by distinct trajectories. Nature Mental Health 1, 863\u0026ndash;875 (2023). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/s44220-023-00139-4\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/s44220-023-00139-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSun, X. \u003cem\u003eet al.\u003c/em\u003e Mapping Neurophysiological Subtypes of Major Depressive Disorder Using Normative Models of the Functional Connectome. Biol Psychiatry 94, 936\u0026ndash;947 (2023). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1016/j.biopsych.2023.05.021\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1016/j.biopsych.2023.05.021\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eK\u0026uuml;blb\u0026ouml;ck, M. \u003cem\u003eet al.\u003c/em\u003e Stability of low-frequency fluctuation amplitudes in prolonged resting-state fMRI. Neuroimage 103, 249\u0026ndash;257 (2014). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1016/j.neuroimage.2014.09.038\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1016/j.neuroimage.2014.09.038\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu, J. \u003cem\u003eet al.\u003c/em\u003e Alterations in amplitude of low frequency fluctuation in treatment-na\u0026iuml;ve major depressive disorder measured with resting-state fMRI. Hum Brain Mapp 35, 4979\u0026ndash;4988 (2014). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1002/hbm.22526\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1002/hbm.22526\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGong, J. \u003cem\u003eet al.\u003c/em\u003e Common and distinct patterns of intrinsic brain activity alterations in major depression and bipolar disorder: voxel-based meta-analysis. Transl Psychiatry 10, 353 (2020). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/s41398-020-01036-5\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/s41398-020-01036-5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTeng, C. \u003cem\u003eet al.\u003c/em\u003e Abnormal resting state activity of left middle occipital gyrus and its functional connectivity in female patients with major depressive disorder. BMC Psychiatry 18, 370 (2018). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1186/s12888-018-1955-9\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1186/s12888-018-1955-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMeyer, J. H. \u003cem\u003eet al.\u003c/em\u003e Elevated monoamine oxidase a levels in the brain: an explanation for the monoamine imbalance of major depression. Arch Gen Psychiatry 63, 1209\u0026ndash;1216 (2006). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1001/archpsyc.63.11.1209\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1001/archpsyc.63.11.1209\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKiecolt-Glaser, J. K., Derry, H. M. \u0026amp; Fagundes, C. P. Inflammation: depression fans the flames and feasts on the heat. Am J Psychiatry 172, 1075\u0026ndash;1091 (2015). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1176/appi.ajp.2015.15020152\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1176/appi.ajp.2015.15020152\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eScaini, G. \u003cem\u003eet al.\u003c/em\u003e Dysregulation of mitochondrial dynamics, mitophagy and apoptosis in major depressive disorder: Does inflammation play a role? Mol Psychiatry 27, 1095\u0026ndash;1102 (2022). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/s41380-021-01312-w\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/s41380-021-01312-w\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAmin, N. \u003cem\u003eet al.\u003c/em\u003e Interplay of Metabolome and Gut Microbiome in Individuals With Major Depressive Disorder vs Control Individuals. JAMA Psychiatry 80, 597\u0026ndash;609 (2023). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1001/jamapsychiatry.2023.0685\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1001/jamapsychiatry.2023.0685\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHasler, G. Pathophysiology of depression: do we have any solid evidence of interest to clinicians? World Psychiatry 9, 155\u0026ndash;161 (2010). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1002/j.2051-5545.2010.tb00298.x\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1002/j.2051-5545.2010.tb00298.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChang, M. \u003cem\u003eet al.\u003c/em\u003e Identifying and validating subtypes within major psychiatric disorders based on frontal-posterior functional imbalance via deep learning. Mol Psychiatry 26, 2991\u0026ndash;3002 (2021). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/s41380-020-00892-3\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/s41380-020-00892-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuo, H. \u003cem\u003eet al.\u003c/em\u003e Brain Functional and Structural Alterations in Women With Bipolar Disorder and Suicidality. Front Psychiatry 12, 630849 (2021). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.3389/fpsyt.2021.630849\u003c/span\u003e\u003cspan address=\"https://doi.org:10.3389/fpsyt.2021.630849\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDuan, J. \u003cem\u003eet al.\u003c/em\u003e Neurodevelopmental trajectories, polygenic risk, and lipometabolism in vulnerability and resilience to schizophrenia. BMC Psychiatry 23, 153 (2023). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1186/s12888-023-04597-z\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1186/s12888-023-04597-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWorsley, K. J., Chen, J. I., Lerch, J. \u0026amp; Evans, A. C. Comparing functional connectivity via thresholding correlations and singular value decomposition. Philos Trans R Soc Lond B Biol Sci 360, 913\u0026ndash;920 (2005). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1098/rstb.2005.1637\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1098/rstb.2005.1637\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMacQueen, J.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMonti, S., Tamayo, P., Mesirov, J. \u0026amp; Golub, T. Consensus Clustering: A Resampling-Based Method for Class Discovery and Visualization of Gene Expression Microarray Data. Machine Learning 52, 91\u0026ndash;118 (2003). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1023/A:1023949509487\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1023/A:1023949509487\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePedregosa, F. \u003cem\u003eet al.\u003c/em\u003e Scikit-learn: Machine Learning in Python. ArXiv abs/1201.0490 (2011).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePidsley, R. \u003cem\u003eet al.\u003c/em\u003e Critical evaluation of the Illumina MethylationEPIC BeadChip microarray for whole-genome DNA methylation profiling. Genome Biol 17, 208 (2016). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1186/s13059-016-1066-1\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1186/s13059-016-1066-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKhan, N., Bano, A., Rahman, M. A., Rathinasabapathi, B. \u0026amp; Babar, M. A. UPLC-HRMS-based untargeted metabolic profiling reveals changes in chickpea (Cicer arietinum) metabolome following long-term drought stress. Plant Cell Environ 42, 115\u0026ndash;132 (2019). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1111/pce.13195\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1111/pce.13195\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZheng, J. \u003cem\u003eet al.\u003c/em\u003e Integrative omics analysis reveals epigenomic and transcriptomic signatures underlying brain structural deficits in major depressive disorder. Transl Psychiatry 14, 17 (2024). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/s41398-023-02724-8\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/s41398-023-02724-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTian, Y. \u003cem\u003eet al.\u003c/em\u003e ChAMP: updated methylation analysis pipeline for Illumina BeadChips. Bioinformatics 33, 3982\u0026ndash;3984 (2017). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1093/bioinformatics/btx513\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1093/bioinformatics/btx513\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLigthart, S. \u003cem\u003eet al.\u003c/em\u003e DNA methylation signatures of chronic low-grade inflammation are associated with complex diseases. Genome Biol 17, 255 (2016). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1186/s13059-016-1119-5\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1186/s13059-016-1119-5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStevenson, A. J. \u003cem\u003eet al.\u003c/em\u003e Characterisation of an inflammation-related epigenetic score and its association with cognitive ability. Clin Epigenetics 12, 113 (2020). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1186/s13148-020-00903-8\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1186/s13148-020-00903-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRahmani, E. \u003cem\u003eet al.\u003c/em\u003e GLINT: a user-friendly toolset for the analysis of high-throughput DNA-methylation array data. Bioinformatics 33, 1870\u0026ndash;1872 (2017). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1093/bioinformatics/btx059\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1093/bioinformatics/btx059\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKoestler, D. C. \u003cem\u003eet al.\u003c/em\u003e Improving cell mixture deconvolution by identifying optimal DNA methylation libraries (IDOL). BMC Bioinformatics 17, 120 (2016). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1186/s12859-016-0943-7\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1186/s12859-016-0943-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eReinius, L. E. \u003cem\u003eet al.\u003c/em\u003e Differential DNA methylation in purified human blood cells: implications for cell lineage and studies on disease susceptibility. PLoS One 7, e41361 (2012). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1371/journal.pone.0041361\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1371/journal.pone.0041361\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBenjamini, Y. \u0026amp; Hochberg, Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. Journal of the Royal Statistical Society: Series B (Methodological) 57, 289\u0026ndash;300 (2018). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1111/j.2517-6161.1995.tb02031.x\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1111/j.2517-6161.1995.tb02031.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, Y., Lu, J., Yu, J., Gibbs, R. A. \u0026amp; Yu, F. An integrative variant analysis pipeline for accurate genotype/haplotype inference in population NGS data. Genome Res 23, 833\u0026ndash;842 (2013). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1101/gr.146084.112\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1101/gr.146084.112\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWray, N. R. \u003cem\u003eet al.\u003c/em\u003e Genome-wide association analyses identify 44 risk variants and refine the genetic architecture of major depression. Nat Genet 50, 668\u0026ndash;681 (2018). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/s41588-018-0090-3\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/s41588-018-0090-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYu, G., Wang, L. G., Han, Y. \u0026amp; He, Q. Y. clusterProfiler: an R package for comparing biological themes among gene clusters. Omics 16, 284\u0026ndash;287 (2012). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1089/omi.2011.0118\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1089/omi.2011.0118\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGreen, C. \u003cem\u003eet al.\u003c/em\u003e Structural brain correlates of serum and epigenetic markers of inflammation in major depressive disorder. Brain Behav Immun 92, 39\u0026ndash;48 (2021). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1016/j.bbi.2020.11.024\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1016/j.bbi.2020.11.024\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBarker, E. D. \u003cem\u003eet al.\u003c/em\u003e Inflammation-related epigenetic risk and child and adolescent mental health: A prospective study from pregnancy to middle adolescence. Dev Psychopathol 30, 1145\u0026ndash;1156 (2018). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1017/s0954579418000330\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1017/s0954579418000330\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTalens, R. P. \u003cem\u003eet al.\u003c/em\u003e Variation, patterns, and temporal stability of DNA methylation: considerations for epigenetic epidemiology. Faseb j 24, 3135\u0026ndash;3144 (2010). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1096/fj.09-150490\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1096/fj.09-150490\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNorthoff, G., Wiebking, C., Feinberg, T. \u0026amp; Panksepp, J. The 'resting-state hypothesis' of major depressive disorder-a translational subcortical-cortical framework for a system disorder. Neurosci Biobehav Rev 35, 1929\u0026ndash;1945 (2011). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1016/j.neubiorev.2010.12.007\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1016/j.neubiorev.2010.12.007\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePhillips, M. L. \u003cem\u003eet al.\u003c/em\u003e Identifying predictors, moderators, and mediators of antidepressant response in major depressive disorder: neuroimaging approaches. Am J Psychiatry 172, 124\u0026ndash;138 (2015). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1176/appi.ajp.2014.14010076\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1176/appi.ajp.2014.14010076\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu, F., Lu, Q., Kong, Y. \u0026amp; Zhang, Z. A Comprehensive Overview of the Role of Visual Cortex Malfunction in Depressive Disorders: Opportunities and Challenges. Neurosci Bull 39, 1426\u0026ndash;1438 (2023). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1007/s12264-023-01052-7\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1007/s12264-023-01052-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSun, D. \u003cem\u003eet al.\u003c/em\u003e Frontal-posterior functional imbalance and aberrant function developmental patterns in schizophrenia. Transl Psychiatry 11, 495 (2021). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/s41398-021-01617-y\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/s41398-021-01617-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJiang, S. \u003cem\u003eet al.\u003c/em\u003e Progressive trajectories of schizophrenia across symptoms, genes, and the brain. BMC Med 21, 237 (2023). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1186/s12916-023-02935-2\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1186/s12916-023-02935-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuo, H. \u003cem\u003eet al.\u003c/em\u003e Early-Stage Repetitive Transcranial Magnetic Stimulation Altered Posterior-Anterior Cerebrum Effective Connectivity in Methylazoxymethanol Acetate Rats. Front Neurosci 15, 652715 (2021). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.3389/fnins.2021.652715\u003c/span\u003e\u003cspan address=\"https://doi.org:10.3389/fnins.2021.652715\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu, J. \u003cem\u003eet al.\u003c/em\u003e Visual cortex repetitive transcranial magnetic stimulation (rTMS) reversing neurodevelopmental impairments in adolescents with major psychiatric disorders (MPDs): A cross-species translational study. CNS Neurosci Ther 30, e14427 (2024). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1111/cns.14427\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1111/cns.14427\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, D. \u003cem\u003eet al.\u003c/em\u003e Targeted visual cortex stimulation (TVCS): a novel neuro-navigated repetitive transcranial magnetic stimulation mode for improving cognitive function in bipolar disorder. Transl Psychiatry 13, 193 (2023). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/s41398-023-02498-z\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/s41398-023-02498-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBerk, M. \u003cem\u003eet al.\u003c/em\u003e So depression is an inflammatory disease, but where does the inflammation come from? BMC Med 11, 200 (2013). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1186/1741-7015-11-200\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1186/1741-7015-11-200\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWohleb, E. S., Franklin, T., Iwata, M. \u0026amp; Duman, R. S. Integrating neuroimmune systems in the neurobiology of depression. Nat Rev Neurosci 17, 497\u0026ndash;511 (2016). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/nrn.2016.69\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/nrn.2016.69\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDowlati, Y. \u003cem\u003eet al.\u003c/em\u003e A meta-analysis of cytokines in major depression. Biol Psychiatry 67, 446\u0026ndash;457 (2010). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1016/j.biopsych.2009.09.033\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1016/j.biopsych.2009.09.033\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKim, Y. K., Na, K. S., Myint, A. M. \u0026amp; Leonard, B. E. The role of pro-inflammatory cytokines in neuroinflammation, neurogenesis and the neuroendocrine system in major depression. Prog Neuropsychopharmacol Biol Psychiatry 64, 277\u0026ndash;284 (2016). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1016/j.pnpbp.2015.06.008\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1016/j.pnpbp.2015.06.008\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEnache, D., Pariante, C. M. \u0026amp; Mondelli, V. Markers of central inflammation in major depressive disorder: A systematic review and meta-analysis of studies examining cerebrospinal fluid, positron emission tomography and post-mortem brain tissue. Brain Behav Immun 81, 24\u0026ndash;40 (2019). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1016/j.bbi.2019.06.015\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1016/j.bbi.2019.06.015\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSingh, D. \u003cem\u003eet al.\u003c/em\u003e Changes in leukocytes and CRP in different stages of major depression. J Neuroinflammation 19, 74 (2022). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1186/s12974-022-02429-7\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1186/s12974-022-02429-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eS\u0026oslash;rensen, N. V. \u003cem\u003eet al.\u003c/em\u003e Immune cell composition in unipolar depression: a comprehensive systematic review and meta-analysis. Mol Psychiatry 28, 391\u0026ndash;401 (2023). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1038/s41380-022-01905-z\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1038/s41380-022-01905-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCapuron, L. \u0026amp; Dantzer, R. Cytokines and depression: the need for a new paradigm. Brain Behav Immun 17 Suppl 1, S119\u0026ndash;124 (2003). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1016/s0889-1591(02)00078-8\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1016/s0889-1591(02)00078-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchaefer, M. \u003cem\u003eet al.\u003c/em\u003e Prevention of interferon-alpha associated depression in psychiatric risk patients with chronic hepatitis C. J Hepatol 42, 793\u0026ndash;798 (2005). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1016/j.jhep.2005.01.020\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1016/j.jhep.2005.01.020\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDunn, A. J., Swiergiel, A. H. \u0026amp; de Beaurepaire, R. Cytokines as mediators of depression: what can we learn from animal studies? Neurosci Biobehav Rev 29, 891\u0026ndash;909 (2005). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1016/j.neubiorev.2005.03.023\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1016/j.neubiorev.2005.03.023\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSu, W. J., Hu, T. \u0026amp; Jiang, C. L. Cool the Inflamed Brain: A Novel Anti-inflammatory Strategy for the Treatment of Major Depressive Disorder. Curr Neuropharmacol 22, 810\u0026ndash;842 (2024). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.2174/1570159x21666230809112028\u003c/span\u003e\u003cspan address=\"https://doi.org:10.2174/1570159x21666230809112028\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLin, K. \u003cem\u003eet al.\u003c/em\u003e Inflammation, brain structure and cognition interrelations among individuals with differential risks for bipolar disorder. Brain Behav Immun 83, 192\u0026ndash;199 (2020). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1016/j.bbi.2019.10.010\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1016/j.bbi.2019.10.010\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGoldsmith, D. R., Bekhbat, M., Mehta, N. D. \u0026amp; Felger, J. C. Inflammation-Related Functional and Structural Dysconnectivity as a Pathway to Psychopathology. Biol Psychiatry 93, 405\u0026ndash;418 (2023). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org:10.1016/j.biopsych.2022.11.003\u003c/span\u003e\u003cspan address=\"https://doi.org:10.1016/j.biopsych.2022.11.003\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"translational-psychiatry","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"tp","sideBox":"Learn more about [Translational Psychiatry](http://www.nature.com/tp/)","snPcode":"41398","submissionUrl":"https://mts-tp.nature.com/cgi-bin/main.plex","title":"Translational Psychiatry","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-4852981/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4852981/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eThe heterogeneity of Major Depressive Disorder (MDD) has been increasingly recognized, challenging traditional symptom-based diagnostics and the development of mechanism-targeted therapies. This study aims to identify neuroimaging-based MDD subtypes and dissect their predominant biological characteristics using multi-omics data.\u003c/p\u003e\u003ch2\u003eMethod\u003c/h2\u003e \u003cp\u003eA total of 807 participants were included in this study, comprising 327 individuals with MDD and 480 healthy controls (HC). The amplitude of low-frequency fluctuations (ALFF), a functional neuroimaging feature, was extracted for each participant and used to identify MDD subtypes through machine learning clustering. Multi-omics data, including profiles of genetic, epigenetics, metabolomics, and pro-inflammatory cytokines, were obtained. Comparative analyses of multi-omics data were conducted between each MDD subtype and HC to explore the molecular underpinnings involved in each subtype.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eWe identified three neuroimaging-based MDD subtypes, each characterized by unique ALFF pattern alterations compared to HC. Multi-omics analysis showed a strong genetic predisposition for Subtype 1, primarily enriched in neuronal development and synaptic regulation pathways. This subtype also exhibited the most severe depressive symptoms and cognitive decline compared to the other subtypes. Subtype 2 is characterized by immuno-inflammation dysregulation, supported by elevated IL-1β levels, altered epigenetic inflammatory measures, and differential metabolites correlated with IL-1β levels. No significant biological markers were identified for Subtype 3.\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eOur results identify neuroimaging-based MDD subtypes and delineate the distinct biological features of each subtype. This provides a proof of concept for mechanism-targeted therapy in MDD, highlighting the importance of personalized treatment approaches based on neurobiological and molecular profiles.\u003c/p\u003e","manuscriptTitle":"Dissecting biological heterogeneity in major depressive disorder based on neuroimaging subtypes with multi-omics data","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-10-15 07:59:08","doi":"10.21203/rs.3.rs-4852981/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"revise","date":"2024-09-18T14:18:46+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"This content is not available.","date":"2024-09-04T00:26:36+00:00","index":3,"fulltext":"This content is not available."},{"type":"editorInvitedReview","content":"This content is not available.","date":"2024-09-02T07:26:14+00:00","index":2,"fulltext":"This content is not available."},{"type":"editorInvitedReview","content":"This content is not available.","date":"2024-09-01T09:34:57+00:00","index":1,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2024-08-22T09:28:07+00:00","index":3,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2024-08-22T01:32:51+00:00","index":2,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2024-08-21T21:27:55+00:00","index":1,"fulltext":"This content is not available."},{"type":"reviewersInvited","content":"","date":"2024-08-21T19:52:15+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-08-08T11:08:44+00:00","index":"","fulltext":""},{"type":"submitted","content":"Translational Psychiatry","date":"2024-08-08T03:38:23+00:00","index":"","fulltext":""},{"type":"checksFailed","content":"","date":"2024-08-05T10:28:47+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-08-03T11:03:11+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"translational-psychiatry","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"tp","sideBox":"Learn more about [Translational Psychiatry](http://www.nature.com/tp/)","snPcode":"41398","submissionUrl":"https://mts-tp.nature.com/cgi-bin/main.plex","title":"Translational Psychiatry","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"f4c5d7a6-f262-4175-b759-a2211d74ed4a","owner":[],"postedDate":"October 15th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":36384358,"name":"Biological sciences/Neuroscience/Molecular neuroscience"},{"id":36384359,"name":"Health sciences/Diseases/Psychiatric disorders/Depression"},{"id":36384360,"name":"Health sciences/Pathogenesis/Clinical genetics"},{"id":36384361,"name":"Biological sciences/Genetics/Clinical genetics"}],"tags":[],"updatedAt":"2025-03-04T08:14:25+00:00","versionOfRecord":{"articleIdentity":"rs-4852981","link":"https://doi.org/10.1038/s41398-025-03286-7","journal":{"identity":"translational-psychiatry","isVorOnly":false,"title":"Translational Psychiatry"},"publishedOn":"2025-03-04 05:00:00","publishedOnDateReadable":"March 4th, 2025"},"versionCreatedAt":"2024-10-15 07:59:08","video":"","vorDoi":"10.1038/s41398-025-03286-7","vorDoiUrl":"https://doi.org/10.1038/s41398-025-03286-7","workflowStages":[]},"version":"v1","identity":"rs-4852981","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4852981","identity":"rs-4852981","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}