Neurophysiological correlates of neuroanatomical dimensions in major depressive disorder: EEG markers of brain function and treatment outcome

Neurophysiological correlates of neuroanatomical dimensions in major depressive disorder: EEG markers of brain function and treatment outcome | 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 Research Article Neurophysiological correlates of neuroanatomical dimensions in major depressive disorder: EEG markers of brain function and treatment outcome Wenyi Xiao, Rachel D. Woodham, Mathilde Antoniades, Dhivya Srinivasan, and 45 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6658719/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Major depressive disorder (MDD) is a heterogeneous disorder with variable treatment responses and no established biomarkers for identification or predictors of treatment response. In the COORDINATE-MDD consortium, a data-driven classification identified two neuroanatomic-based dimensions: Dimension 1 (D1), with preserved grey and white matter volumes, and Dimension 2 (D2), with widespread reductions. Here, we investigated whether resting-state electroencephalography (EEG) features differ between these dimensions and whether such features predict treatment response. Participants were 237 MDD (155 women; mean age 37.47 ± 13.36 year) from two clinical trials: CAN-BIND (escitalopram) and EMBARC (randomized to sertraline or placebo). All were medication-free at baseline, in a current depressive episode of at least moderate severity. Resting-state, eyes-closed EEG was recorded at baseline. EEG features included spectral power, frontal alpha asymmetry (FAA), multiscale sample entropy (MSE), and inter-site phase clustering (ISPC). Analyses examined effects of neuroanatomical dimension (D1, D2) and clinical outcome (responder, non-responder), using ANCOVA and threshold-free cluster enhancement (TFCE). D1 participants showed greater absolute and log-transformed theta, alpha, and beta power, and lower relative delta power compared to D2, particularly in frontal and central regions. Among responders, D1 showed higher alpha and theta power and greater MSE at coarse time scales. Individuals with preserved neuroanatomy in D1 exhibit electrophysiological markers of more flexible and integrated brain function, possibly reflecting efficient top-down regulation and adaptive neural dynamics. In contrast, the D2 dimension, marked by lower complexity and elevated delta activity, may reflect disrupted network integrity, reduced cortical arousal, and impaired information processing. Psychiatry Major depressive disorder EEG neurophysiology neuroanatomical dimensions treatment response resting-state EEG spectral power frontal alpha asymmetry multiscale entropy Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Major depressive disorder (MDD) is common and is a major contributor to the global burden of disease (Yan et al., 2024 ). MDD is a clinically heterogeneous mental health condition, which is characterized by wide variability in symptom profiles, illness trajectories, and responses to treatment (Fried and Nesse, 2015 ; Fu et al., 2019 ). Current diagnostic systems are based on symptom checklists rather than underlying biological mechanisms and are not able to capture this complexity. Moreover, treatment decisions are largely by trial-and-error, and fewer than one-third of patients achieving remission after first-line antidepressant therapy (Pigott et al., 2023 ). There are currently no biomarkers capable of predicting treatment response at the individual level (Fu et al., 2023 ). We have sought to investigate the neurobiological heterogeneity of major depressive disorder (MDD) through data-driven machine learning analyses of structural MRI scans in the COORDINATE-MDD consortium (Fu et al., 2023 ). Based on neuroanatomical measures in MDD individuals, all medication-free, with first episode or recurrent MDD and in current depressive episode, the data-driven machine learning analysis revealed two dimensions (Fu et al., 2024 ). Dimension 1 (D1) was characterised by preserved grey and white matter volumes in which MDD individuals showed a significantly greater clinical response to treatment with selective serotonin reuptake inhibitors (SSRIs) antidepressant medication as compared to placebo. In contrast, Dimension 2 (D2) was characterised by widespread subtle reductions in grey and white matter in which MDD individuals showed limited clinical improvement to either SSRIs or placebo treatment. External validation in the UK Biobank confirmed the stability of these dimensions across clinical and general populations (Xiao et al., 2025 ). D2 was aligned with an “immuno-metabolic” MDD profile which is characterised by systemic inflammation and metabolic dysregulation, as observed in the Netherlands Study of Depression and Anxiety (NESDA) (Lamers et al., 2020 ). D2 was further associated with cognitive impairments, greater childhood adversity, and greater rates of self-harm and suicidality, suggesting a biologically distinct, metabolically and immunologically dysregulated MDD dimension (Xiao et al., 2025 ). Complementary to structural neuroimaging markers, electrophysiological measures, such as electroencephalography (EEG) provide a temporally sensitive assessment of brain function. Quantitative EEG (QEEG) involves spectral analysis of resting-state brain activity, capturing dynamic neural oscillations, and has shown potential in identifying neurophysiological predictors of treatment response to antidepressant medication as well as to repetitive transcranial magnetic stimulation (rTMS) treatments (Watts et al. 2022 ). ​Schwartzmann et al. ( 2023 ) developed and validated a multivariate model based on resting-state EEG features which predicted treatment response to SSRI antidepressant in MDD. The model achieved a balanced accuracy of 64.2% in internal validation in the national Canadian Biomarker Integration Network in Depression (CAN-BIND) cohort and 63.7% in external validation in the Establishing Moderators and Biosignatures of Antidepressant Response for Clinical Care (EMBARC) multi-site clinical trial cohort. Notably, when applied to the EMBARC placebo treatment group, the model's balanced accuracy dropped to 48.7%, indicating its specificity to pharmacological treatment response. More advanced computational techniques provide deeper analysis of neural dynamics. Measures such as frontal alpha asymmetry (FAA), multiscale entropy (MSE) and inter-site phase clustering (ISPC) capture additional features of brain function, including asymmetry, complexity and connectivity. FAA quantifies the difference in alpha-band power between the left and right frontal cortices and has long been proposed as a marker of affective style. Greater left-sided activity has often been associated with approach-related motivation and positive affect and has been examined as a potential predictor of antidepressant treatment outcomes, though findings have been mixed (Fitzgerald, 2024 ). Pizzagalli et al. ( 2018 ) reported that increased rostral anterior cingulate cortex theta activity is prognostic marker of treatment outcome in EMBARC which has been consistently observed in MDD treatment studies. MSE quantifies the complexity of neural signals across multiple time scales. Higher MSE values at coarse scales and lower MSE at fine scales may reflect greater adaptability and efficiency of brain function (Sporns et al., 2000 ). Higher MSE at coarse scales has been associated with better treatment outcomes to antidepressant medications in MDD (Jaworska et al., 2018) and reductions in MSE at fine time scales following seizure therapy have been associated with symptom improvement (Farzan et al., 2017 ). Zhdanov et al. ( 2020 ) identified MSE as among the most informative features for predicting treatment response to escitalopram medication in MDD. ISPC captures the consistency of phase relationships between signals from different brain regions and is commonly used as an index of functional connectivity, with stronger phase synchronisation potentially reflecting greater coordination within neural networks (Cohen, 2014 ). Increased connectivity at baseline has been associated with better clinical outcomes to antidepressant medication (Olbrich et al., 2014 ). In the present study, we sought to examine the electrophysiological correlates of the two neuroanatomical dimensions (D1 and D2) identified in the COORDINATE-MDD consortium (Fu et al., 2023 , 2024 ; Xiao et al., 2025 ). We sought to determine whether these structurally defined dimensions exhibit distinct resting-state EEG signatures, which could provide complementary, temporally sensitive markers of underlying neural function. EEG data were eyes-closed resting-state recordings obtained from the Canadian Biomarker Integration Network in Depression (CAN-BIND) (MacQueen et al., 2019 ) and the multi-site clinical trial Establishing Moderators and Biosignatures of Antidepressant Response in Clinical Care (EMBARC) (Trivedi et al., 2016 ) datasets, both of which are part of the COORDINATE-MDD consortium. We further assessed whether the EEG differences were associated with antidepressant treatment response and whether effects vary by treatment type (SSRI and placebo). Specifically, we investigate spectral power, FAA, MSE, and ISPC to identify EEG correlates of structural brain differences and clinical outcomes. Methods Datasets Data were integrated from two independent clinical trials: CAN-BIND (Kennedy et al., 2019 ) and EMBARC (Trivedi et al., 2016 ), within the COORDINATE-MDD consortium (Fu et al., 2023 ). Ethical approval was granted by the respective Research Ethics Boards. All participants had provided written informed consent. CAN-BIND is a multi-site prospective treatment study with four recruitment sites in Canada (Kennedy et al., 2019 ). Inclusion criteria included major depressive disorder (MDD) diagnosis, according to DSM-IV-TR criteria, using the Mini-International Neuropsychiatric Interview (MINI), in current depressive episode with Montgomery-Åsberg Depression Rating Scale (MADRS) score of a minimum of 24 and being medication-free. From CAN-BIND, EEG data were available in 138 MDD participants (82 women; mean age 36.24 years, SD 12.80 years) and 54 healthy control participants (35 women; mean age 33.32 years, SD 11.01 years), acquired at baseline before treatment initiation. All MDD participants received treatment with SSRI antidepressant, escitalopram, for 8 weeks. Treatment response was defined as a > = 50% reduction in HAMD equivalent scores from baseline to week 8. EMBARC is a mulit-site prospective, double-blind, randomised controlled trial with four recruitment sites in USA (Trivedi et al., 2016 ). Inclusion criteria included MDD diagnosis based on Structured Clinical Interview for DSM-IV (SCID), being in a current depressive episode of at least moderate severity defined as a 17-item Hamilton Depression Rating Scale (HAMD-17) score of a minimum of 14, and being medication-free. From EMBARC, EEG data were available in 281 MDD participants (185 women; mean age 37.18 years, SD 14.78 years) and 39 healthy control participants (24 women; mean age 37.21 years, SD 13.29 years), acquired at baseline before treatment initiation. In the initial treatment phase, MDD participants were randomised to receive either SSRI, sertraline, or placebo medication for 8 weeks. Treatment response was defined as achieving a > = 50% reduction in HAMD-17 score from baseline to week 8. The combined dataset initially consisted of 250 MDD participants (CAN-BIND: 54; EMBARC: 196) who had both structural MRI dimension classification (D1, D2) and EEG data. During EEG preprocessing, 13 participant data were excluded due to data quality issues, resulting in a final sample of 237 MDD participants in the present analysis: D1: 116 MDD (70 women; mean age 37.58 years, SD 13.25 years; mean years of education 15.21 years, SD 2.49 years), and D2: 121 MDD (85 women; mean age 37.36 years, SD 13.51 years; mean years of education 14.60 years, SD 2.62 years)). Within CAN-BIND, D1: 15 MDD (9 women); D2: 39 MDD (29 women), and within EMBARC, D1: 101 MDD (61 women); D2: 82 MDD (56 women) (Table 1 ). Table 1 Demographic characteristics of participants CAN-BIND EMBARC Demographic feature Dimension 1 Dimension 2 Dimension 1 Dimension 2 Number (women) 15 (9) 39 (29) 101(61) 82 (56) Age (years) 34.2 (14.11) 38.72 (12.96) 38.08 (13.12) 36.72 (13.8) Years of Education 13.33 (1.54) 13.79 (2.61) 15.5 (2.49) 14.98 (2.55) Depressive severity (baseline) MADRS 29.08(5.66) 30.39(5.28) NA NA HAMD-17 NA NA 19.42 (3.82) 19.68 (3.63) Treatment Antidepressant medication 15 39 54 33 Placebo NA NA 47 49 Values represent means with standard deviations in parentheses. Age and years of education are in years. Depression severity was assessed using the Montgomery–Åsberg Depression Rating Scale (MADRS) in CAN-BIND and the 17-item Hamilton Depression Rating Scale (HAMD-17) in EMBARC. HAMD-17 scores reported for CAN-BIND are estimated from MADRS using equipercentile linking (Leucht et al., 2018 ) to allow comparison across studies. In CAN-BIND, all participants received escitalopram (ESC). In EMBARC, participants were randomised to receive either sertraline (SER) or placebo (PLA). Abbreviations: ESC = escitalopram; SER = sertraline; PLA = placebo; HAMD = Hamilton Depression Rating Scale; MADRS = Montgomery–Åsberg Depression Rating Scale. Mean baseline depressive severity scores were CAN-BIND D1 (MADRS mean = 28.10, SD = 5.22), CAN-BIND D2 (MADRS mean = 30.18, SD = 5.45), EMBARC D1 (HAMD-17 mean = 19.42, SD = 3.82), and EMBARC D2 (HAMD-17 mean = 19.68, SD = 3.63). Following treatment: CAN-BIND D1 (MADRS mean = 17.53, SD = 5.36), CAN-BIND D2 (MADRS mean = 17.67, SD = 5.89), EMBARC D1 (HAMD-17 mean = 11.63, SD = 7.50) and EMBARC D2 (HAMD-17 mean = 11.50, SD = 6.29). In CAN-BIND, all MDD participants received SSRI medication (escitalopram). In EMBARC, participants were randomized to either SSRI medication (D1: 54 participants; D2: 33 participants) or placebo (D1: 47 participants; D2: 49 participants). Treatment response was attained following the course of treatments in, D1: 50 MDD (34 women) and D2: 48 MDD (34 women), and persistent depressive symptoms were observed in, D1: 66 MDD (36 women) and D2: 73 MDD (51 women) (Table 2 ). Table 2 Demographic characteristics by Dimension (D1, D2) and treatment response status Dimension 1 Dimension 2 Treatment response Persistent symptoms Treatment response Persistent symptoms Number (women) 50 (34) 66 (36) 48 (34) 73 (51) Age (years) 35.02 (12.04) 39.52 (13.87) 37.42 (14.42) 37.33 (12.98) Year of Education 15.32 (2.32) 15.14 (2.63) 14.43 (2.81) 14.71 (2.51) HAMD baseline 19.74 (4.00) 19.40 (3.64) 20.71 (3.79) 20.47 (4.19) HAMD post-treatment 4.90 (2.93) 17.05 (4.90) 5.79 (3.62) 16.29 (4.60) Treatment EMBARC placebo 16 31 19 30 EMBARC medication 29 25 14 19 CAN-BIND medication 5 10 15 24 Values represent means with standard deviations in parentheses. Age and years of education are in years. Depression severity was assessed using the Montgomery–Åsberg Depression Rating Scale (MADRS) in CAN-BIND and the 17-item Hamilton Depression Rating Scale (HAMD-17) in EMBARC. HAMD-17 scores for CAN-BIND were estimated from MADRS using equipercentile linking (Leucht et al., 2018 ) for harmonisation. Original MADRS baseline and post-treatment values for CAN-BIND are in Table 1 . In CAN-BIND, participants received escitalopram (ESC). In EMBARC, participants were randomised to sertraline (SER) or placebo (PLA). Abbreviations: ESC = escitalopram; SER = sertraline; PLA = placebo; HAMD = Hamilton Depression Rating Scale; MADRS = Montgomery–Åsberg Depression Rating Scale. EEG data recording In CAN-BIND, participants underwent an 8-minute resting-state EEG recording with their eyes closed at baseline (Farzan et al., 2017 ; Lam et al., 2016 ). In EMBARC, resting-state EEG was acquired in four 2-minute blocks at baseline, consisting of two eye-open segments and two eye-closed segments (Trivedi et al., 2016 ). Data from the eyes-closed recordings only were included to ensure consistency across datasets. EEG data preprocessing EEG recordings were preprocessed to minimize artifacts and standardize the dataset for analysis. Data were first segmented into 2-second epochs to facilitate artifact detection and spectral analysis. To mitigate edge effects from filtering, each epoch was forward and backward reflected before applying standard EEG preprocessing procedures. EEG data were then high-pass filtered at 0.5 Hz using the MATLAB highpass function to remove slow drifts, low-pass filtered at 200 Hz using the lowpass function to remove high-frequency noise, and notch filtered at either 60 Hz or 50 Hz using designNotchPeakIIR and filtfilt to eliminate line noise. Specifically, a 60 Hz notch filter was applied to EEG data collected from both the EMBARC (United States) and CAN-BIND (Canada) datasets, consistent with the local electrical frequency. To ensure data quality, a voltage deflection threshold approach was applied for identifying and rejecting noisy epochs. A moving 80 ms window was used to compute the difference between the maximum and minimum voltage within each epoch. Epochs were flagged as potentially noisy if any channel exhibited a voltage deflection exceeding 100 µV. After noisy epochs were identified, hannels exhibiting high-amplitude artifacts, defined as exceeding 100 µV threshold in more than 50% trials, were identified and removed. Artifact rejection was then applied to the remaining data to exclude trials containing residual noise, having more than 25% of remaining channels exceeding 100 µV noise threshold. At the participant level, additional data quality criteria were applied to ensure reliable EEG signals base on the following criteria: original EEG recording contained fewer than 20 usable EEG channels or if more than 50% of channels were excluded due to excessive noise contamination. Based on these criteria, from the initial sample of 250 MDD, 13 MDD participant data were excluded, and the final dataset consisted of 237 MDD participants (Tables 1 – 2 ). To ensure consistency between datasets, EMBARC EEG data were re-referenced to an average reference, and 58 common electrodes from the eyes-closed condition were retained to match the CAN-BIND dataset. Subsequently, all EEG recordings were re-referenced to the average reference and spatially interpolated to a standardized 32-channel montage to enable harmonized analyses. Analysis of primary EEG parameters Resting-state EEG data were analyzed to extract four primary features: power spectra, frontal alpha asymmetry (FAA), multiscale sample entropy (MSE), and inter-site phase clustering (ISPC). All feature computations were performed using MATLAB scripts adapted from the publicly available Sheffield Autism Biomarkers toolbox (Dede et al., 2025 ), which has been previously applied to large-scale clinical EEG datasets. The original functions are available via GitHub ( https://github.com/adede1988/SheffieldAutismBiomarkers ). Power Spectra Power spectra were computed separately for each epoch. To minimize edge artifacts, epochs were mirrored, and time-frequency decomposition was performed using wavelet convolution across 100 logarithmically spaced frequencies (2–80 Hz) (Cohen, 2014 ). The resulting complex time series was converted to power values, and mirrored segments were removed. Mean power values were then averaged across epochs, yielding a 32 (electrodes) × 100 (frequencies) matrix per participant. For power spectrum comparisons, spectra from each channel were averaged within six standard frequency bands: delta (2–4 Hz), theta (4–8 Hz), alpha (8–14 Hz), beta (14–30 Hz) and gamma (30–80 Hz). Frontal Alpha Asymmetry Frontal Alpha Asymmetry (FAA) was computed using log-transformed power values in the alpha band (8–14 Hz). Alpha power was averaged across bilateral frontal electrodes: F3 and F7 (left hemisphere), and F4 and F8 (right hemisphere). FAA was calculated as the difference in log power between homologous electrode pairs (i.e., F4-F3 and F8-F7) (Allen et al., 2004 ). Multiscale Sample Entropy Multiscale sample entropy (MSE) was computed to quantify signal complexity across multiple temporal scales. Data were resampled to 1000 Hz and coarse-grained into 20-time scales. Sample entropy was estimated using a similarity threshold of 30% of the signal's standard deviation and a pattern length of m = 2 (Costa et al., 2005 ). Prior to entropy estimation, signals were normalized to reduce amplitude-related bias. The final output for each participant was a 32-channel × 20-time-scale matrix, averaged across epochs. Inter-Site Phase Clustering Inter-site phase clustering (ISPC) was used to assess phase synchrony between electrode pairs across the same 100 frequencies used for spectral power. Signals were first transformed using a surface Laplacian to improve spatial resolution. ISPC was then calculated based on the consistency of phase differences between electrode pairs across time (Cohen, 2014 ). ISPC was calculated according to the following formula: ISPC = \(\:\frac{1}{n}\left|{\sum\:}_{t=1}^{n}{e}^{i\left({\phi\:}_{xt}-{\phi\:}_{yt}\right)}\right|\) where \(\:{\phi\:}_{xt}\) and \(\:{\phi\:}_{yt}\) represent the phase angles of signals from two channels, \(\:x\) and \(\:y\) , at time \(\:t\:\) . The result was a 32 × 32 × 100 matrix per participant, averaged across epochs. Analyses of EEG-derived features Resting-state EEG data were harmonized across datasets into 32 channels and processed to extract a range of spectral and complexity-based features for group-level analysis. EEG-derived features were analyzed with respect to neuroanatomical dimension (D1; D2), treatment outcome status (responder (R), defined as > = 50% improvement in depressive symptoms from baseline; non-responder (NR), defined as having persistent symptoms with < 50% improvement from baseline). EEG features were analyzed separately within the full sample and within participants stratified by treatment group (antidepressant medication or placebo), to account for potential confounding between treatment type and dataset source. For each analysis, group-level comparisons were conducted using two-way Analysis of Covariance (ANCOVA) models, with each EEG-derived feature entered as a dependent variable. Fixed factors were Dimension (D1; D2), Outcome (R; NR), and age and sex were covariates. EEG features were standardized across participants. Where appropriate, log-transformations were applied to reduce skewness, and outliers exceeding ± 5 standard deviations (z-score) were excluded. Effect sizes were quantified using eta-squared (η²), representing the proportion of variance explained by each main effect or interaction. Parameter estimates (β), 95% confidence intervals (CIs), and corresponding p -values were extracted for all predictors. To control multiple comparison problems across features, False Discovery Rate (FDR) correction was applied. All statistical analyses and visualizations were performed in R (R Core Team, 2023 ), with graphical outputs generated using the ggplot2 package (Wickham, 2016 ). Threshold-free cluster enhancement (TFCE) analysis Threshold-free cluster enhancement (TFCE) is a nonparametric, cluster-based analysis which accounts for spatial and spectral dependencies in EEG data (Smith and Nichols, 2009 ; Mensen and Khatami, 2013 ). TFCE has higher sensitivity to weak but spatially or temporally extended effects and does not require predefined cluster-forming thresholds. TFCE was applied using a two-tailed, permutation-based cluster testing framework. EEG-derived metrics, including spectral power, multiscale sample entropy (MSE), and inter-site phase clustering (ISPC), were evaluated across all channels and either frequency bins (for power and ISPC) or time scales (for MSE). TFCE parameters were configured based on Mensen and Khatami ( 2013 ), with the height exponent (H) set to 2 and the extent exponent (E) set to 0.66. Statistical inference was performed using Monte Carlo permutation testing with 5,000 iterations. All analyses were conducted using the ept_TFCE function from the TFCE MATLAB toolbox ( https://github.com/Mensen/ept_TFCE-matlab ). Comparisons were conducted across combinations of neuroanatomical dimensions (D1; D2) and Outcomes (R; NR), including both overall group comparisons (e.g., D1 vs. D2; R vs. NR) and subgroup contrasts within and across dimensions by response combinations. To further explore the robustness of significant findings, follow-up TFCE analyses were also conducted separately within the antidepressant medication (ADM) and placebo (PLA) subgroups. Results Power Spectra ANCOVA conducted on the full sample revealed a significant interaction between neuroanatomical dimension and clinical outcome for relative gamma power at electrode CP6 ( η² = 0.043, p = 0.035 FDR-corrected). Relative gamma power was lower in D1 responders (mean = -0.073, SE = 0.144) than in D1 non-responders (mean = 0.074, SE = 0.122), while values were comparable between D2 responders (mean = 0.071, SE = 0.154) and D2 non-responders (mean = 0.084, SE = 0.116) (Fig. 3 ). Post hoc pairwise comparisons did not reach statistical significance following FDR correction. No significant main effects of dimension or clinical outcome status were observed in absolute, relative, or log-transformed spectral power features across the full sample (Supplementary Table 1). Similarly, stratified analyses with the antidepressant-treated and placebo-treated subgroups did not reveal any significant main effects or interaction (Supplementary Tables 2 and 3). TFCE analyses revealed significant group-level differences in EEG power across neuroanatomical dimensions (Supplementary Table 4). Participants classified in D1 showed significantly greater absolute theta and alpha power in frontal regions, elevated theta, alpha, and beta power in central regions, and increased beta power in parieto-occipital regions relative to those in D2. In contrast, D1 participants showed significantly lower relative delta power across frontal, central, and parieto-occipital regions compared to D2. For log-transformed power, D1 participants showed significantly greater theta, alpha, and beta power across widespread scalp regions. No significant TFCE-based differences in spectral power were observed as a function of clinical outcome status alone. Among participants who showed a clinical response, baseline TFCE analysis showed increased absolute and log-transformed alpha power in frontal and central regions, elevated log-transformed theta power in frontal regions, decreased relative delta power in frontal regions, and decreased beta power in frontal regions in D1 compared to D2 (Supplementary Table 4). No significant dimension-related differences were observed within the non-responder subgroup. When stratified by treatment modality, SSRI-treated responders in D1 showed significantly greater absolute theta and alpha power across frontal, central, and parieto-occipital electrodes compared to D2 responders (Supplementary Table 5). In contrast, placebo-treated responders in D1 exhibited significantly lower relative delta power and increased relative alpha power in frontal, central, and posterior regions compared to D2 responders (Supplementary Table 6). Frontal Alpha Asymmetry There was a significant main effect of responder status on frontal alpha asymmetry (FAA) ( F (1, 227) = 5.10, p = 0.025 FDR-corrected, η² = 0.022), in which responders exhibited lower FAA values compared to non-responders at baseline, supported by post hoc Tukey’s HSD testing ( p = 0.025 FDR-corrected). No significant main effects were found for neuroanatomical dimension ( F (1, 227) = 0.25, p = 0.620 FDR-corrected, η² = 0.001) or treatment type ( F (1, 227) = 0.60, p = 0.440 FDR-corrected, η² = 0.003), nor were any two-way or three-way interactions statistically significant (all p > 0.20). As FAA is derived from an individual summary index across predefined electrode pairs and not assessed across spatially extended features or frequency bins, it was not suitable for TFCE analysis. Multiscale Sample Entropy At baseline, D1 responders (across all treatment types) showed significantly higher MSE, particularly at coarse time scales (time scales 14–20) over fronto-central regions compared to D2 responders (Supplementary Table 4). This effect was also observed when stratifying by treatment group. In participants receiving placebo treatment, D1 responders showed elevated MSE across a wider range of electrodes (frontal, central, parietal, and occipital) and broader scale range (time scales 6–20). All results were significant at p < .046 (TFCE-corrected) (Supplementary Table 6). In contrast, no significant baseline MSE differences were observed between D1 and D2 responders in the antidepressant medication group. No significant differences in MSE were found between non-responders in either treatment group. Inter-Site Phase Clustering ANCOVA analyses in the full sample revealed a significant main effect of treatment response status in theta-band ISPC at electrode F8 ( η² = 0.048, p = 0.027 FDR-corrected), with responders demonstrating increased ISPC at baseline relative to non-responders (Fig. 4 ; Supplementary Table 1). There were no significant main effects of dimension, nor any interaction effects. Subgroup analysis by treatment type (antidepressant medication, placebo) similarly showed no significant effects (Supplementary Tables 2 and 3). TFCE analyses did not reveal any significant differences in ISPC across groups (Supplementary Table 4). Discussion The present study sought to investigate whether MDD neuroanatomical dimensions previously derived from structural MRI in the COORDINATE-MDD consortium, correspond to distinct patterns of EEG metrics from resting-state eyes-closed brain activity, namely spectral power, frontal alpha asymmetry (FAA), multiscale entropy (MSE), and inter-site phase clustering (ISPC). We found that the structural dimensions, D1 characterized by relatively preserved brain volumes, and D2 associated with widespread grey and white matter reductions, also differ in electrophysiological dynamics, in particular among participants who showed a clinical response to antidepressant treatment. EEG power analyses using TFCE revealed that MDD individuals classified within D1 showed elevated absolute theta and alpha power in frontal regions, increased beta power in both central and parieto-occipital regions, as well as widespread elevations in log-transformed theta, alpha and beta power, as well as significantly lower relative delta power across frontal and posterior regions compared to D1. Increases in alpha are associated with top-down inhibitory control that facilitates selective attention by suppressing task-irrelevant brain areas (Jensen and Mazaheri, 2010 ). While alpha oscillations are typically associated with global corticocortical dynamics, delta oscillations may index more locally dominant neocortical processes (Nunez, 1995 ; Nunez Srinivasan, 2006). The pattern of reduced relative delta power and increased absolute alpha power observed in D1 individuals may indicate a shift from locally driven cortical inhibition toward more global, top-down regulatory control. This shift could support more flexible and selective engagement of task-relevant brain regions, facilitating improved cognitive control and emotion regulation. In participants who subsequently showed a treatment response, this spectral pattern was similarly observed at baseline, indicating that this EEG profile may be predictive of clinical response. Increased alpha power, particularly in posterior and prefrontal regions, has been observed in in treatment responders across a range of modalities, including SSRI antidepressant medication (Bruder et al., 2008 ) and neuromodulatory approaches such as repetitive transcranial magnetic stimulation (rTMS; Noda et al., 2013 ). Wu et al. ( 2020 ) found that an alpha-based EEG signature predicted response to antidepressant medication but not placebo in EMBARC. In the present study, alpha power increase was observed only in D1 responders at baseline in the antidepressant treatment arm, not in the placebo treatment group, suggesting a treatment-specific electrophysiological profile. The observation of reduced relative delta power in D1 responders at baseline has been observed for SSRI antidepressant medication (Iosifescu et al., 2009 ) and cognitive behavioral therapy (Schwartzmann et al., 2023 ). Delta oscillations, particularly in frontal and medial prefrontal regions, have been associated with internal attention and inhibitory control, especially during cognitively demanding tasks such as mental arithmetic and working memory (Harmony, 2013 ). In such contexts, elevated delta is thought to support internal attention by inhibiting irrelevant environmental input. Increased delta activity has also been observed during states of reduced cognitive and emotional engagement, such as meditation, where it might reflect cortical inhibition or downregulation (Tei et al., 2009 ), a more active and engaged neural state which could facilitate adaptive processing and responsiveness. Notably, no significant power differences were observed between D1 and D2 non-responders, supporting the specificity of these effects to positive treatment outcomes. In frontal alpha asymmetry (FAA), a main effect of clinical response status was observed at baseline in which MDD participants who subsequently showed a good clinical response to have a lower FAA, reflecting greater left-lateralised frontal alpha activity, compared to those who had persistent symptoms following treatment. The approach-withdrawal model proposes two motivational systems corresponding to positive and negative affects which are reflected in relative lateralisation of EEG activity (Henriques and Davidson, 1991 ). Lower FAA has been observed in MDD, which has been hypothesised to be associated with anhedonia (Nusslock et al, 2015 ), however van der Vinne et al. ( 2017 ) meta-analysis reports a small effect which was not significant citing high heterogeneity across studies. Multiscale entropy (MSE) analysis revealed greater signal complexity at coarse time scales in D1 responders at baseline compared to D2 responders, consistent with proposals that greater neural complexity reflects more flexible and adaptive brain states (von Stein and Sarnthein, 2000 ; Schnitzler and Gross, 2005 ). Greater coarse-scale entropy has been associated with better clinical outcomes, which predicted clinical outcome in CAN-BIND (Zhdanov et al., 2020 ) as well as in EMBARC (Schwartzmann et al., 2023 ). Among participants who showed a response to placebo treatment, we found increased MSE in D1 at baseline compared to D2, suggesting that coarse-scale signal complexity may reflect a more general response which includes non-pharmacological treatments. Inter-site phase clustering (ISPC) analyses indicated higher theta-band synchrony at electrode F8 in responders at baseline compared to non-responders, implicating increased frontotemporal connectivity in treatment response. Pizzagalli et al. ( 2018 ) observed increased frontal theta activity, in particular in the rostral anterior cingulate cortex, predicted clinical response to antidepressant medication, which has been consistently implicated as a predictor of treatment response (Pizzagalli, 2011 ; Fu et al., 2013 ). Increased synchrony in these low-frequency bands may indicate stronger coordination within frontoparietal networks. These findings suggest that D1 may reflect a more neurophysiologically adaptive profile compared to D2 which is characterised by preserved oscillatory dynamics aligning with the neuroanatomical classification (Fu et al., 2024 ) and external validation in UK Biobank (Xiao et al., 2025 ). D1 could reflect a biologically distinct, more resilient MDD dimension with preserved brain structure and function, while D2 was associated with subtle but widespread cortical atrophy, increased systemic inflammation, metabolic disturbances, and a history of greater childhood adversity, suggesting a pathophysiological profile marked by greater vulnerability in neurodevelopmental and immuno-metabolic pathways (Fu et al., 2024 ; Xiao et al., 2025 ). Limitations in the present study include the sample size which was limited to available data in resting-state EEG which may not capture the full complexity of neural dynamics. While resting-state metrics are robust and clinically accessible, task-based or dynamic EEG measures, such as during emotional or cognitive tasks, might provide additional information about functional differences between dimensions and treatment responsiveness (Webb et al., 2016). Incorporation of longitudinal designs would clarify predictive and state-dependent effects. In summary, the present study observed that the neuroanatomical dimensions of MDD, previously identified via structural MRI, are reflected in distinct resting-state EEG profiles. Individuals classified within D1 showed increased absolute and log-transformed theta, alpha, and beta power alongside reduced relative delta power, a pattern suggesting enhanced global cortical regulation and reduced local inhibition. These spectral features were associated with antidepressant treatment response. These findings suggest that D1 may reflect a biologically and neurophysiologically more resilient MDD dimension, associated with preserved brain structure, lower inflammation, and more adaptive electrophysiological profiles that confer improved treatment responsiveness. In contrast, D2 appears linked to subtle widespread cortical atrophy, immuno-metabolic dysregulation, and a history of higher childhood adversity, consistent with greater vulnerability. References Allen JJ, Urry HL, Hitt SK, Coan JA (2004) The stability of resting frontal electroencephalographic asymmetry in depression. Psychophysiology 41(2):269–280. https://doi.org/10.1111/j.1469-8986.2003.00149.x Bruder GE, Sedoruk JP, Stewart JW, McGrath PJ, Quitkin FM, Tenke CE (2008) Electroencephalographic alpha measures predict therapeutic response to a selective serotonin reuptake inhibitor antidepressant: Pre- and post-treatment findings. Biol Psychiatry 63(12):1171–1177. https://doi.org/10.1016/j.biopsych.2007.10.009 Čukić M, Stokić M, Radenković S, Ljubisavljević M, Simić S, Savić D (2020) Nonlinear analysis of EEG complexity in episode and remission phase of recurrent depression. Int J Methods Psychiatr Res 29(2):e1816. https://doi.org/10.1002/mpr.1816 Cohen MX (2014) Analyzing neural time series data: Theory and practice. MIT Press Costa M, Goldberger AL, Peng C-K (2005) Multiscale entropy analysis of biological signals. Phys Rev E 71(2 Pt 1):021906. https://doi.org/10.1103/PhysRevE.71.021906 Dede AJ, Xiao W, Vaci N et al (2025) Exploring EEG resting state differences in autism: Sparse findings from a large cohort. Mol Autism 16:13. https://doi.org/10.1186/s13229-025-00647-3 Farzan F, Atluri S, Frehlich M et al (2017) Standardization of electroencephalography for multi-site, multi-platform and multi-investigator studies: Insights from the Canadian Biomarker Integration Network in Depression. Sci Rep 7:7473. https://doi.org/10.1038/s41598-017-07613-x Farzan F, Atluri S, Mei Y, Moreno S, Levinson AJ, Blumberger DM, Daskalakis ZJ (2017) Brain temporal complexity in explaining the therapeutic and cognitive effects of seizure therapy. Brain 140(4):1011–1025. https://doi.org/10.1093/brain/awx030 Fitzgerald PJ (2024) Frontal alpha asymmetry and its modulation by monoaminergic neurotransmitters in depression. Clin Psychopharmacol Neurosci 22(3):405–415. https://doi.org/10.9758/cpn.23.1138 Fried EI, Nesse RM (2015) Depression is not a consistent syndrome: An investigation of unique symptom patterns in the STARD study. J Affect Disord 172:96–102. https://doi.org/10.1016/j.jad.2014.10.010 Fu CHY, Steiner H, Costafreda SG (2013) Predictive neural biomarkers of clinical response in depression: A meta-analysis of functional and structural neuroimaging studies of pharmacological and psychological therapies. Neurobiol Dis 52:75–83. https://doi.org/10.1016/j.nbd.2012.05.008 Fu CHY, Fan Y, Davatzikos C (2019) Addressing heterogeneity (and homogeneity) in treatment mechanisms in depression and the potential to develop diagnostic and predictive biomarkers. NeuroImage: Clin 24:101997. https://doi.org/10.1016/j.nicl.2019.101997 Fu CHY, Erus G, Fan Y, Antoniades M, Arnone D, Arnott SR, Chen T, Choi KS, Fatt CC, Frey BN, Frokjaer VG, Ganz M, Garcia J, Godlewska BR, Hassel S, Ho K, McIntosh AM, Qin K, Rotzinger S, Davatzikos C (2023) AI-based dimensional neuroimaging system for characterizing heterogeneity in brain structure and function in major depressive disorder: COORDINATE-MDD consortium design and rationale. BMC Psychiatry 23:59. https://doi.org/10.1186/s12888-022-04509-7 Fu CHY, Fan Y, Davatzikos C, Erus G, Godlewska BR, Woodham RD, Xiao W, Choi KS, Frokjaer VG, Sacchet MD, Savitz J, Qin K, McIntosh AM (2024) Neuroanatomical dimensions in medication-free individuals with major depressive disorder and treatment response to SSRI antidepressant medications or placebo. Nat Mental Health. https://doi.org/10.1038/s44220-023-00187-w Harmony T (2013) The functional significance of delta oscillations in cognitive processing. Front Integr Nuerosci 7:83. https://doi.org/10.3389/fnint.2013.00083 Henriques JB, Davidson RJ (1991) Left frontal hypoactivation in depression. J Abnorm Psychol 100(4):535–545. https://doi.org/10.1037/0021-843X.100.4.535 Iosifescu DV, Greenwald S, Devlin P, Mischoulon D, Denninger JW, Alpert JE, Fava M (2009) Frontal EEG predictors of treatment outcome in major depressive disorder. Eur Neuropsychopharmacol 19(11):772–777. https://doi.org/10.1016/j.euroneuro.2009.06.001 Jaworska N, Blondeau C, Tessier P, Norris S, Fusee W, Blier P, Knott V (2013) Response prediction to antidepressants using scalp and source-localized loudness dependence of auditory evoked potential (LDAEP) slopes. Prog Neuropsychopharmacol Biol Psychiatry 44:100–107. https://doi.org/10.1016/j.pnpbp.2013.01.012 Jaworska N, de la Salle S, Ibrahim MH, Blier P, Knott V (2019) Leveraging machine learning approaches for predicting antidepressant treatment response using electroencephalography (EEG) and clinical data. Front Psychiatry 9:768. https://doi.org/10.3389/fpsyt.2018.00768 Jensen O, Mazaheri A (2010) Shaping functional architecture by oscillatory alpha activity: Gating by inhibition. Front Hum Neurosci 4:186. https://doi.org/10.3389/fnhum.2010.00186 Kennedy SH, Lam RW, Rotzinger S, Milev RV, Blier P, Downar J, Evans KR, Foster JA, MacQueen GM, Parikh SV, Uher R (2019) Symptomatic and functional outcomes and early prediction of response to escitalopram monotherapy and sequential adjunctive aripiprazole therapy in patients with major depressive disorder: A CAN-BIND-1 report. J Clin Psychiatry 80(2):18m12202. https://doi.org/10.4088/JCP.18m12202 Lam RW, Milev R, Rotzinger S, Andreazza AC, Blier P, Brenner C, Daskalakis Z, Dharsee M, Downar J, Evans KR, Farzan F, Foster JA, Frey BN, Giacobbe P, Hall GB, Ismail Z, Kennedy SH, MacQueen GM, McAndrews MP, Young LT (2016) Discovering biomarkers for antidepressant response: Protocol from the Canadian Biomarker Integration Network in Depression (CAN-BIND) and clinical characteristics of the first patient cohort. BMC Psychiatry 16:105. https://doi.org/10.1186/s12888-016-0785-x Lamers F, Milaneschi Y, Vinkers CH, Schoevers RA, Giltay EJ, Penninx BWJH (2020) Depression profilers and immuno-metabolic dysregulation: Longitudinal results from the NESDA study. Brain Behav Immun 88:174–183. https://doi.org/10.1016/j.bbi.2020.04.002 Leucht S, Fennema H, Engel RR, Kaspers-Janssen M, Szegedi A (2018) Translating the HAM-D into the MADRS and vice versa with equipercentile linking. J Affect Disord 226:326–331. https://doi.org/10.1016/j.jad.2017.09.042 MacQueen GM, Hassel S, McIntosh AR, Giacobbe P, Rizvi SJ, Kennedy SH, Lam RW, Milev RV, Parikh SV, Rotzinger S, Uher R, Young LT (2019) The Canadian Biomarker Integration Network in Depression (CAN-BIND): Magnetic resonance imaging protocols. J Psychiatry Neurosci 44(4):223–236. https://doi.org/10.1503/jpn.180202 Mensen A, Khatami R (2013) Advanced EEG analysis using threshold-free cluster-enhancement and non-parametric statistics. NeuroImage 67:111–118. https://doi.org/10.1016/j.neuroimage.2012.10.027 Noda Y, Nakamura M, Saeki T, Iwata Y, Ikeda Y, Aubert G, Honda S, Yoshino A, Kato TA (2013) Potentiation of quantitative electroencephalograms following prefrontal repetitive transcranial magnetic stimulation in patients with major depression. Neurosci Res 77(1–2):70–77. https://doi.org/10.1016/j.neures.2013.06.002 Nunez PL (1995) Neocortical dynamics and human EEG rhythms. Oxford University Press Nunez PL, Srinivasan R (2006) A theoretical basis for standing and traveling brain waves measured with human EEG with implications for an integrated consciousness. Clin Neurophysiol 117(11):2424–2435. https://doi.org/10.1016/j.clinph.2006.06.754 Nusslock R, Walden K, Harmon-Jones E (2015) Asymmetrical frontal cortical activity associated with differential risk for mood and anxiety disorder symptoms: An RDoC perspective. International Journal of Psychophysiology, 98 (2, Part 2), 249–261. https://doi.org/10.1016/j.ijpsycho.2015.06.004 Olbrich S, Tränkner A, Chittka T, Hegerl U, Schönknecht P (2014) Functional connectivity in major depression: Increased phase synchronization between frontal cortical EEG-source estimates. Psychiatry Research: Neuroimaging 222(1–2):91–99. https://doi.org/10.1016/j.pscychresns.2014.02.010 Pigott HE, Kim T, Xu C, Dubin MJ, Quitkin FM, Trivedi MH (2023) What are the treatment remission, response and extent of improvement rates after up to four trials of antidepressant therapies in real-world depressed patients? A reanalysis of the STAR*D study’s patient-level data with fidelity to the original research protocol. BMJ Open 13:e063095. https://doi.org/10.1136/bmjopen-2022-063095 Pizzagalli DA (2011) Frontocingulate dysfunction in depression: Toward biomarkers of treatment response. Neuropsychopharmacology 36:183–206. https://doi.org/10.1038/npp.2010.166 Pizzagalli DA, Webb CA, Dillon DG, Tenke CE, Kayser J, Goer F, Fava M, McGrath PJ, Weissman MM, Bruder GE (2018) Pretreatment rostral anterior cingulate cortex theta activity in relation to symptom improvement in depression: A randomized clinical trial. JAMA Psychiatry 75(6):547–554. https://doi.org/10.1001/jamapsychiatry.2018.0252 R Core Team (2023) R: A language and environment for statistical computing. R Foundation for Statistical Computing. https://www.R-project.org/ Schnitzler A, Gross J (2005) Normal and pathological oscillatory communication in the brain. Nat Rev Neurosci 6(4):285–296. https://doi.org/10.1038/nrn1650 Schwartzmann B, Quilty LC, Dhami P, Kolla N, Mulsant BH, Daskalakis ZJ, Blumberger DM, Kennedy SH, Farzan F (2023) Resting-state EEG delta and alpha power predict response to cognitive behavioral therapy in depression: A Canadian Biomarker Integration Network in Depression study. Sci Rep 13:8418. https://doi.org/10.1038/s41598-023-35179-4 Schwartzmann B, Dhami P, Uher R, Farzan F, Kennedy SH (2023) Developing an electroencephalography-based model for predicting response to antidepressant medication. JAMA Netw Open 6(9):e2336094. https://doi.org/10.1001/jamanetworkopen.2023.36094 Smith SM, Nichols TE (2009) Threshold-free cluster enhancement: Addressing problems of smoothing, threshold dependence and localisation in cluster inference. NeuroImage 44(1):83–98. https://doi.org/10.1016/j.neuroimage.2008.03.061 Sporns O, Tononi G, Edelman GM (2000) Connectivity and complexity: The relationship between neuroanatomy and brain dynamics. Neural Netw 13(8–9):909–922. https://doi.org/10.1016/S0893-6080(00)00053-8 Tei S, Faber PL, Lehmann D, Tsujiuchi T, Kumano H, Pascual-Marqui RD, Kinoshita T (2009) Meditators and non-meditators: EEG source imaging during resting. Brain Topogr 22(3):158–165. https://doi.org/10.1007/s10548-009-0107-4 Trivedi MH, McGrath PJ, Fava M, Parsey RV, Kurian BT, Phillips ML, Oquendo MA, Bruder GE, Pizzagalli DA, Toups M, Cooper C, Adams P, Weyandt S, McInnis MG, Kraemer HC (2016) Establishing moderators and biosignatures of antidepressant response in clinical care (EMBARC): Rationale and design. J Psychiatr Res 78:11–23. https://doi.org/10.1016/j.jpsychires.2016.03.001 Trivedi MH, Rush AJ, Wisniewski SR, Nierenberg AA, Warden D, Ritz L, Norquist G, Howland RH, Lebowitz B, McGrath PJ, Shores-Wilson K, Biggs MM, Balasubramani GK, Fava M (2006) Evaluation of outcomes with citalopram for depression using measurement-based care in STAR*D: Implications for clinical practice. Am J Psychiatry 163(1):28–40. https://doi.org/10.1176/appi.ajp.163.1.28 Ulke C, Kayser J, Tenke CE, Mergl R, Sander C, Panier LYX, Alvarenga JE, Fava M, McGrath PJ, Deldin PJ, McInnis MG, Trivedi MH, Weissman MM, Pizzagalli DA, Hegerl U, Bruder GE (2024) EEG measures of brain arousal in relation to symptom improvement in patients with major depressive disorder: Results from a randomized placebo-controlled clinical trial. Psychiatry Res 342:116165. https://doi.org/10.1016/j.psychres.2024.116165 van der Vinne N, Vollebregt MA, van Putten MJAM, Arns M (2017) Frontal alpha asymmetry as a diagnostic marker in depression: Fact or fiction? A meta-analysis. NeuroImage: Clin 16:79–87. https://doi.org/10.1016/j.nicl.2017.07.006 von Stein A, Sarnthein J (2000) Different frequencies for different scales of cortical integration: From local gamma to long range alpha/theta synchronization. Int J Psychophysiol 38(3):301–313. https://doi.org/10.1016/S0167-8760(00)00172-0 Watts D, Pulice RF, Reilly J, Deldin PJ, Olbrich S, Tenke CE, Kayser J, Bruder GE, McGrath PJ (2022) Predicting treatment response using EEG in major depressive disorder: A machine-learning meta-analysis. Translational Psychiatry 12:332. https://doi.org/10.1038/s41398-022-02064-z Wickham H (2016) ggplot2: Elegant graphics for data analysis. Springer-. https://doi.org/10.1007/978-3-319-24277-4_9 Wu W, Zhang Y, Jiang J, Lucas MV, Fonzo GA, Rolle CE, Cooper C, Chin-Fatt C, Krepel N, Cornelssen CA, Wright R, Toll RT, Harrewijn A, Etkin A (2020) An electroencephalographic signature predicts antidepressant response in major depression. Nat Biotechnol 38(4):439–447. https://doi.org/10.1038/s41587-019-0397-3 Xiao W, Woodham RD, Cui Y, Antoniades M, Srinivasan D, Fan Y, Erus G, Garcia JA, Arnott SR, Chen T, Choi KS, Chin-Fatt C, Frey BN, Frokjaer VG, Ganz M, Godlewska BR, Hassel S, Ho K, McIntosh AM, Davatzikos C (2025) Neuroanatomical dimensions in major depression: External validation and links with cognition, adverse life events, self-harm, metabolomics and genetics [Preprint]. Res Square. https://doi.org/10.21203/rs.3.rs-5975731/v1 Yan G, Zhang Y, Wang S, Chen Q, Zhu L, Fang Y, Tang W (2024) Global, regional, and national temporal trend in burden of major depressive disorder from 1990 to 2019: An analysis of the global burden of disease study. Psychiatry Res 337:115958. https://doi.org/10.1016/j.psychres.2024.115958 Zhdanov A, Atluri S, Wong W, Levinson AJ, Blumberger DM, Daskalakis ZJ, Farzan F (2020) Use of machine learning for predicting escitalopram treatment outcome from electroencephalography recordings in adult patients with depression. JAMA Netw Open 3(1):e1918377. https://doi.org/10.1001/jamanetworkopen.2019.18377 Additional Declarations The authors declare no competing interests. Supplementary Files supplementarytables.xlsx Cite Share Download PDF Status: Posted Version 1 posted 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-6658719","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":456175047,"identity":"cd895328-706b-4ff2-b340-15937f0750e4","order_by":0,"name":"Wenyi Xiao","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAuElEQVRIiWNgGAWjYFACxgYGhgoGBj4IL4FYLWcYGNhI0ALS1UaKFoMbyY0fPs47LMfGwPzwA2NbGhFazhxslpy57bAxGwObsQRjWw5hLWbHGxukebcdTgS6zQzowgoitBxmbP7NO+dwfRsD+zcitRxvbJPmbTicwMbAA7KFCIfZnznYZjnjWLphGzNPsUTCOSK8Lzkj/fGNDzXW8vzs7Rs/fChLJqwFAZgZiI/IUTAKRsEoGAUEAADFdjPll0sjuAAAAABJRU5ErkJggg==","orcid":"","institution":"University of East London","correspondingAuthor":true,"prefix":"","firstName":"Wenyi","middleName":"","lastName":"Xiao","suffix":""},{"id":456175376,"identity":"a0f54739-92fd-428e-b9f8-436e6d683114","order_by":1,"name":"Rachel D. Woodham","email":"","orcid":"","institution":"University of East London","correspondingAuthor":false,"prefix":"","firstName":"Rachel","middleName":"D.","lastName":"Woodham","suffix":""},{"id":456175480,"identity":"1c897554-ea43-4647-930d-d2ba06cc2925","order_by":2,"name":"Mathilde Antoniades","email":"","orcid":"","institution":"University of Pennsylvania","correspondingAuthor":false,"prefix":"","firstName":"Mathilde","middleName":"","lastName":"Antoniades","suffix":""},{"id":456175507,"identity":"0f57b597-0aa1-4826-b521-17127ec17b0f","order_by":3,"name":"Dhivya Srinivasan","email":"","orcid":"","institution":"University of Pennsylvania","correspondingAuthor":false,"prefix":"","firstName":"Dhivya","middleName":"","lastName":"Srinivasan","suffix":""},{"id":456175508,"identity":"f0d2e183-3ff2-46ef-9076-71e3e0f68464","order_by":4,"name":"Yong Fan","email":"","orcid":"","institution":"University of Pennsylvania","correspondingAuthor":false,"prefix":"","firstName":"Yong","middleName":"","lastName":"Fan","suffix":""},{"id":456175724,"identity":"51627c4f-2e68-4671-a9b9-ae4f3b3febc7","order_by":5,"name":"Guray Erus","email":"","orcid":"","institution":"University of Pennsylvania","correspondingAuthor":false,"prefix":"","firstName":"Guray","middleName":"","lastName":"Erus","suffix":""},{"id":456175725,"identity":"8f9d6b68-9166-4436-8acf-d5e9635007ec","order_by":6,"name":"Jose A. Garcia","email":"","orcid":"","institution":"University of Pennsylvania","correspondingAuthor":false,"prefix":"","firstName":"Jose","middleName":"A.","lastName":"Garcia","suffix":""},{"id":456175726,"identity":"a812fd9b-c52e-4f46-a950-cce7ea1ce9e9","order_by":7,"name":"Stephen R. Arnott","email":"","orcid":"","institution":"Rotman Research Institute, Baycrest Centre","correspondingAuthor":false,"prefix":"","firstName":"Stephen","middleName":"R.","lastName":"Arnott","suffix":""},{"id":456177153,"identity":"abb86865-61a9-49e7-888d-061a186bb8e7","order_by":8,"name":"Taolin Chen","email":"","orcid":"","institution":"West China Hospital of Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Taolin","middleName":"","lastName":"Chen","suffix":""},{"id":456177154,"identity":"690019c2-5310-4738-8534-460badfe79eb","order_by":9,"name":"Ki Sueng Choi","email":"","orcid":"","institution":"Nash Family Center for Advanced Circuit Therapeutics","correspondingAuthor":false,"prefix":"","firstName":"Ki","middleName":"Sueng","lastName":"Choi","suffix":""},{"id":456177155,"identity":"9502805f-5ae7-4331-bd31-df85751ae11c","order_by":10,"name":"Cherise Chin Fatt","email":"","orcid":"","institution":"University of Texas Southwestern Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Cherise","middleName":"Chin","lastName":"Fatt","suffix":""},{"id":456177156,"identity":"901f0a95-bdb6-4e09-b345-5a77051be113","order_by":11,"name":"Benicio N. Frey","email":"","orcid":"","institution":"McMaster University","correspondingAuthor":false,"prefix":"","firstName":"Benicio","middleName":"N.","lastName":"Frey","suffix":""},{"id":456177157,"identity":"8aa2e2a2-a531-4f0b-96b0-5b7296de4d02","order_by":12,"name":"Vibe G. Frokjaer","email":"","orcid":"","institution":"University Hospital Rigshospitalet","correspondingAuthor":false,"prefix":"","firstName":"Vibe","middleName":"G.","lastName":"Frokjaer","suffix":""},{"id":456177286,"identity":"e518b91c-8148-4199-8dbf-61c3e059b63c","order_by":13,"name":"Melanie Ganz","email":"","orcid":"","institution":"University Hospital Rigshospitalet","correspondingAuthor":false,"prefix":"","firstName":"Melanie","middleName":"","lastName":"Ganz","suffix":""},{"id":456177287,"identity":"d7a3316b-cbac-4913-88a3-25430e153460","order_by":14,"name":"Beata R. Godlewska","email":"","orcid":"","institution":"University of Oxford","correspondingAuthor":false,"prefix":"","firstName":"Beata","middleName":"R.","lastName":"Godlewska","suffix":""},{"id":456177288,"identity":"45cae8d1-cad3-44c9-8170-a6b9f2d5aee5","order_by":15,"name":"Stefanie Hassel","email":"","orcid":"","institution":"University of Calgary","correspondingAuthor":false,"prefix":"","firstName":"Stefanie","middleName":"","lastName":"Hassel","suffix":""},{"id":456177439,"identity":"f13c0d5a-9565-4dcc-bbca-0cc8a3688ab9","order_by":16,"name":"Keith Ho","email":"","orcid":"","institution":"University Health Network","correspondingAuthor":false,"prefix":"","firstName":"Keith","middleName":"","lastName":"Ho","suffix":""},{"id":456177440,"identity":"2eb5c710-d651-4487-b665-bb014094f05b","order_by":17,"name":"Andrew M. McIntosh","email":"","orcid":"","institution":"Royal Infirmary Edinburgh","correspondingAuthor":false,"prefix":"","firstName":"Andrew","middleName":"M.","lastName":"McIntosh","suffix":""},{"id":456177927,"identity":"b534ab42-4ac5-46ff-8b7a-74e350012c9b","order_by":18,"name":"Kun Qin","email":"","orcid":"","institution":"West China Hospital of Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Kun","middleName":"","lastName":"Qin","suffix":""},{"id":456177928,"identity":"7c6ffca6-e20b-4937-b4f9-ee57bb4266fc","order_by":19,"name":"Susan Rotzinger","email":"","orcid":"","institution":"University Health Network","correspondingAuthor":false,"prefix":"","firstName":"Susan","middleName":"","lastName":"Rotzinger","suffix":""},{"id":456177929,"identity":"7567175a-e5d2-453d-b772-2671568ab1f6","order_by":20,"name":"Matthew D. Sacchet","email":"","orcid":"","institution":"Harvard Medical School","correspondingAuthor":false,"prefix":"","firstName":"Matthew","middleName":"D.","lastName":"Sacchet","suffix":""},{"id":456177930,"identity":"9f23c1c3-1b02-4906-b748-d2665f1f9432","order_by":21,"name":"Jonathan Savitz","email":"","orcid":"","institution":"Laureate Institute for Brain Research","correspondingAuthor":false,"prefix":"","firstName":"Jonathan","middleName":"","lastName":"Savitz","suffix":""},{"id":456177931,"identity":"5b3c4378-b2ba-4a41-9edf-6a446d07f7f2","order_by":22,"name":"Haochang Shou","email":"","orcid":"","institution":"University of Pennsylvania","correspondingAuthor":false,"prefix":"","firstName":"Haochang","middleName":"","lastName":"Shou","suffix":""},{"id":456177932,"identity":"815c1f95-f226-4ec7-bb71-dce5a13ceaee","order_by":23,"name":"Ashish Singh","email":"","orcid":"","institution":"University of Pennsylvania","correspondingAuthor":false,"prefix":"","firstName":"Ashish","middleName":"","lastName":"Singh","suffix":""},{"id":456177933,"identity":"cdd3cd7b-73e0-4ed6-abc2-6beb322f969f","order_by":24,"name":"Aleks Stolicyn","email":"","orcid":"","institution":"Royal Infirmary Edinburgh","correspondingAuthor":false,"prefix":"","firstName":"Aleks","middleName":"","lastName":"Stolicyn","suffix":""},{"id":456177934,"identity":"8880a5a1-9f55-43af-9324-739d0a0f0021","order_by":25,"name":"Irina Strigo","email":"","orcid":"","institution":"University of California San Francisco","correspondingAuthor":false,"prefix":"","firstName":"Irina","middleName":"","lastName":"Strigo","suffix":""},{"id":456177935,"identity":"3c708dcb-a712-4a1f-a8d2-ee86759022f0","order_by":26,"name":"Stephen C. Strother","email":"","orcid":"","institution":"Baycrest Centre","correspondingAuthor":false,"prefix":"","firstName":"Stephen","middleName":"C.","lastName":"Strother","suffix":""},{"id":456177936,"identity":"0c194aec-81b2-4d7f-9ecf-de107b8f4460","order_by":27,"name":"Duygu Tosun","email":"","orcid":"","institution":"University of California San Francisco","correspondingAuthor":false,"prefix":"","firstName":"Duygu","middleName":"","lastName":"Tosun","suffix":""},{"id":456177937,"identity":"07807af1-3849-4bcd-a9d5-bc70d798f92e","order_by":28,"name":"Dongtao Wei","email":"","orcid":"","institution":"Southwest University","correspondingAuthor":false,"prefix":"","firstName":"Dongtao","middleName":"","lastName":"Wei","suffix":""},{"id":456177938,"identity":"0cd4c237-7f53-4f0a-a5ee-7adce91777b1","order_by":29,"name":"Roland Zahn","email":"","orcid":"","institution":"King’s College London","correspondingAuthor":false,"prefix":"","firstName":"Roland","middleName":"","lastName":"Zahn","suffix":""},{"id":456177939,"identity":"2e8a8c6d-71a6-445a-9897-41bc9075f90b","order_by":30,"name":"Ian M. Anderson","email":"","orcid":"","institution":"University of Manchester","correspondingAuthor":false,"prefix":"","firstName":"Ian","middleName":"M.","lastName":"Anderson","suffix":""},{"id":456177940,"identity":"6877e2e8-7462-4d50-b7ff-6da935af8c11","order_by":31,"name":"W. Edward Craighead","email":"","orcid":"","institution":"Emory University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"W.","middleName":"Edward","lastName":"Craighead","suffix":""},{"id":456177941,"identity":"f81d203e-6bd2-459e-a088-fa040abc0579","order_by":32,"name":"J. F. William Deakin","email":"","orcid":"","institution":"University of Manchester","correspondingAuthor":false,"prefix":"","firstName":"J.","middleName":"F. William","lastName":"Deakin","suffix":""},{"id":456180246,"identity":"a3d590c2-c6b6-4531-a729-8e601f484bdc","order_by":33,"name":"Boadie W. Dunlop","email":"","orcid":"","institution":"Emory University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Boadie","middleName":"W.","lastName":"Dunlop","suffix":""},{"id":456180247,"identity":"b38eb451-6286-4010-bf7d-af4952e222b7","order_by":34,"name":"Rebecca Elliott","email":"","orcid":"","institution":"University of Manchester","correspondingAuthor":false,"prefix":"","firstName":"Rebecca","middleName":"","lastName":"Elliott","suffix":""},{"id":456180248,"identity":"48b6c348-5007-4d06-a393-935d768b4f46","order_by":35,"name":"Qiyong Gong","email":"","orcid":"","institution":"West China Hospital of Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Qiyong","middleName":"","lastName":"Gong","suffix":""},{"id":456180249,"identity":"a59ab2d8-0a73-4557-846a-ae89eb566dc7","order_by":36,"name":"Ian H. Gotlib","email":"","orcid":"","institution":"Stanford University","correspondingAuthor":false,"prefix":"","firstName":"Ian","middleName":"H.","lastName":"Gotlib","suffix":""},{"id":456180250,"identity":"dba8bfef-dc75-4337-af43-7add67293666","order_by":37,"name":"Catherine J. Harmer","email":"","orcid":"","institution":"University of Oxford","correspondingAuthor":false,"prefix":"","firstName":"Catherine","middleName":"J.","lastName":"Harmer","suffix":""},{"id":456180251,"identity":"266c1c25-268e-407e-9c3e-aeb308100303","order_by":38,"name":"Sidney H. Kennedy","email":"","orcid":"","institution":"University Health Network","correspondingAuthor":false,"prefix":"","firstName":"Sidney","middleName":"H.","lastName":"Kennedy","suffix":""},{"id":456180252,"identity":"748ba287-0200-4ead-9f02-6c9aa186a906","order_by":39,"name":"Gitte M. Knudsen","email":"","orcid":"","institution":"University Hospital Rigshospitalet","correspondingAuthor":false,"prefix":"","firstName":"Gitte","middleName":"M.","lastName":"Knudsen","suffix":""},{"id":456180253,"identity":"6816e7eb-8ebf-4117-a1c4-c7bb720110ea","order_by":40,"name":"Helen S. Mayberg","email":"","orcid":"","institution":"Helen S. Mayberg","correspondingAuthor":false,"prefix":"","firstName":"Helen","middleName":"S.","lastName":"Mayberg","suffix":""},{"id":456180254,"identity":"7920203c-5471-4233-a9df-3a53ab2f6b3d","order_by":41,"name":"Martin P. Paulus","email":"","orcid":"","institution":"Laureate Institute for Brain Research","correspondingAuthor":false,"prefix":"","firstName":"Martin","middleName":"P.","lastName":"Paulus","suffix":""},{"id":456180255,"identity":"530dca21-b1b7-4305-97e7-d1cc71b977a5","order_by":42,"name":"Jiang Qiu","email":"","orcid":"","institution":"Southwest University","correspondingAuthor":false,"prefix":"","firstName":"Jiang","middleName":"","lastName":"Qiu","suffix":""},{"id":456180256,"identity":"55a8ac93-f2a0-4892-ae70-dbc56e544af9","order_by":43,"name":"Madhukar H. Trivedi","email":"","orcid":"","institution":"University of Texas Southwestern Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Madhukar","middleName":"H.","lastName":"Trivedi","suffix":""},{"id":456180257,"identity":"7306cfb5-c732-4cc6-b610-8c83ce62134f","order_by":44,"name":"Heather C. Whalley","email":"","orcid":"","institution":"Royal Infirmary Edinburgh","correspondingAuthor":false,"prefix":"","firstName":"Heather","middleName":"C.","lastName":"Whalley","suffix":""},{"id":456180258,"identity":"df2c4f48-e403-4867-8812-f6dd4e8676d5","order_by":45,"name":"Chao-Gan Yan","email":"","orcid":"","institution":"Tsinghua University","correspondingAuthor":false,"prefix":"","firstName":"Chao-Gan","middleName":"","lastName":"Yan","suffix":""},{"id":456180259,"identity":"ac9298ff-ee50-4fba-b764-625e10640ff4","order_by":46,"name":"Allan H. Young","email":"","orcid":"","institution":"Imperial College London","correspondingAuthor":false,"prefix":"","firstName":"Allan","middleName":"H.","lastName":"Young","suffix":""},{"id":456180260,"identity":"f4ec2ec6-88f1-4d59-884e-602eb0c6b02a","order_by":47,"name":"Christos Davatzikos","email":"","orcid":"","institution":"University of Pennsylvania","correspondingAuthor":false,"prefix":"","firstName":"Christos","middleName":"","lastName":"Davatzikos","suffix":""},{"id":456180261,"identity":"8aa76297-3094-4b1c-9abf-8a3b2a46c49c","order_by":48,"name":"Cynthia H.Y. Fu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAArElEQVRIiWNgGAWjYFACxgcMHxgYZCCcA0RpYTZgnMHAwEOaFmYekrTwNzCzSdu22fAwsB9+wMxzhggtEgeAWnLb0ngYeNKA1t0gQosBA/8xoJbDQIflMDDzfCBKC9AWy7b/PAz8b0jRwth2gIdBAmQLMQ6TOMzMbNlzLpmHTeKZwcE5xHifv72Z8caPMjs5fv7khw/eHCNCCwMzlGZjIDYiR8EoGAWjYBQQBgDKcyexHRLW4gAAAABJRU5ErkJggg==","orcid":"","institution":"University of East London","correspondingAuthor":true,"prefix":"","firstName":"Cynthia","middleName":"H.Y.","lastName":"Fu","suffix":""}],"badges":[],"createdAt":"2025-05-13 21:20:00","currentVersionCode":1,"declarations":{"humanSubjects":true,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":true,"humanSubjectConsent":true,"humanSubjectClinicalTrial":true,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-6658719/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6658719/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":82890680,"identity":"c74e7f22-8b71-4cc2-9dd0-480230010c09","added_by":"auto","created_at":"2025-05-16 12:09:46","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":4122671,"visible":true,"origin":"","legend":"\u003cp\u003eEEG power differences between D1 and D2 groups. Topographical differences in panels A(i) absolute power, B(i) relative power, and C(i) log-transformed power across frequency bands (δ, θ, α, β, γ) and electrode channels are shown. All effects reached statistical significance following threshold-free cluster enhancement (TFCE; 5,000 permutations). Warmer colours indicate greater values in D1 group; cooler colours indicate greater values in D2. D1 participants showed significantly higher theta, alpha, and beta power (absolute and log-transformed) across frontal, central, and parieto-occipital sites, and lower relative delta power across widespread regions compared to D2 participants. Panels A(ii)–C(ii) show corresponding scalp topographies of channel-wise mean differences for each feature, visualising the spatial distribution of effects across the scalp. Each map reflects the average of statistically significant differences across all frequency bins for each channel, visualising the spatial distribution of effects.\u003c/p\u003e","description":"","filename":"Figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-6658719/v1/1df24817fe4c4375bd3e3ebf.png"},{"id":82892168,"identity":"3c9f18da-0b7a-4c39-9d7f-494db7140e53","added_by":"auto","created_at":"2025-05-16 12:17:46","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":5289029,"visible":true,"origin":"","legend":"\u003cp\u003eEEG differences between D1 and D2 responders. Group-level comparisons between D1 and D2 responders are shown for A(i) multiscale sample entropy, B(i) absolute power, C(i) relative power, and D(i) log-transformed power across frequency bands (δ, θ, α, β, γ) or time scales (for entropy). The y-axis represents EEG electrode channels. All displayed effects were statistically significant following threshold-free cluster enhancement (TFCE; 5,000 permutations). D1 responders showed significantly greater entropy at coarse time scales (particularly over prefrontal electrodes), higher theta and alpha power (absolute and log-transformed), and lower relative delta power compared to D2 responders. No statistically significant differences were observed across any EEG modality within either D1 or D2 groups. Panels A(ii)–D(ii) present the corresponding scalp topographies of mean differences across EEG channels, highlighting the spatial distribution of group-level effects. Each topomap shows the channel-wise average of statistically significant differences across frequency bins (for power) or time scales (for MSE), reflecting the spatial distribution of group effects.\u003c/p\u003e","description":"","filename":"Figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-6658719/v1/cb518e88361be886ceffe480.png"},{"id":82890669,"identity":"a9e87fc4-8b4f-4e07-a560-10e70c37c9a0","added_by":"auto","created_at":"2025-05-16 12:09:46","extension":"jpeg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":125418,"visible":true,"origin":"","legend":"\u003cp\u003eEstimated marginal means of scaled relative gamma power at electrode CP6 as a function of treatment response and Dimension group (D1, D2). Responders were defined as participants with ≥50% reduction in depressive symptom scores from baseline to post-treatment. Error bars represent 95% confidence intervals.\u003c/p\u003e","description":"","filename":"Figure3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6658719/v1/10ac848db482e3487b5c7d50.jpeg"},{"id":82890674,"identity":"9cea0501-a423-434b-a23b-d5871bd403d8","added_by":"auto","created_at":"2025-05-16 12:09:46","extension":"jpeg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":101436,"visible":true,"origin":"","legend":"\u003cp\u003eEstimated marginal means of scaled theta-band inter-site phase clustering (ISPC) at electrode F8 by treatment response status. Responders exhibited significantly greater ISPC than non-responders (\u003cem\u003ep\u003c/em\u003e = .027), controlling for age and sex. Error bars represent 95% confidence intervals.\u003c/p\u003e","description":"","filename":"Figure4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6658719/v1/e0d8250f0fefd1bb18d5dc7c.jpeg"},{"id":82895376,"identity":"f4c7c5a3-6809-4897-826f-281818e22347","added_by":"auto","created_at":"2025-05-16 12:42:13","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":20348334,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6658719/v1/b8f1bf6d-d42b-46cc-b894-80766a439c8a.pdf"},{"id":82890670,"identity":"ec7317d1-cdac-4539-88f8-124a3d450bd0","added_by":"auto","created_at":"2025-05-16 12:09:46","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1188203,"visible":true,"origin":"","legend":"","description":"","filename":"supplementarytables.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-6658719/v1/dafb4030ccd89c9d831f0e99.xlsx"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003eNeurophysiological correlates of neuroanatomical dimensions in major depressive disorder: EEG markers of brain function and treatment outcome\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003eMajor depressive disorder (MDD) is common and is a major contributor to the global burden of disease (Yan et al., \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). MDD is a clinically heterogeneous mental health condition, which is characterized by wide variability in symptom profiles, illness trajectories, and responses to treatment (Fried and Nesse, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Fu et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Current diagnostic systems are based on symptom checklists rather than underlying biological mechanisms and are not able to capture this complexity. Moreover, treatment decisions are largely by trial-and-error, and fewer than one-third of patients achieving remission after first-line antidepressant therapy (Pigott et al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). There are currently no biomarkers capable of predicting treatment response at the individual level (Fu et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eWe have sought to investigate the neurobiological heterogeneity of major depressive disorder (MDD) through data-driven machine learning analyses of structural MRI scans in the COORDINATE-MDD consortium (Fu et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Based on neuroanatomical measures in MDD individuals, all medication-free, with first episode or recurrent MDD and in current depressive episode, the data-driven machine learning analysis revealed two dimensions (Fu et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Dimension 1 (D1) was characterised by preserved grey and white matter volumes in which MDD individuals showed a significantly greater clinical response to treatment with selective serotonin reuptake inhibitors (SSRIs) antidepressant medication as compared to placebo. In contrast, Dimension 2 (D2) was characterised by widespread subtle reductions in grey and white matter in which MDD individuals showed limited clinical improvement to either SSRIs or placebo treatment. External validation in the UK Biobank confirmed the stability of these dimensions across clinical and general populations (Xiao et al., \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). D2 was aligned with an \u0026ldquo;immuno-metabolic\u0026rdquo; MDD profile which is characterised by systemic inflammation and metabolic dysregulation, as observed in the Netherlands Study of Depression and Anxiety (NESDA) (Lamers et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). D2 was further associated with cognitive impairments, greater childhood adversity, and greater rates of self-harm and suicidality, suggesting a biologically distinct, metabolically and immunologically dysregulated MDD dimension (Xiao et al., \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eComplementary to structural neuroimaging markers, electrophysiological measures, such as electroencephalography (EEG) provide a temporally sensitive assessment of brain function. Quantitative EEG (QEEG) involves spectral analysis of resting-state brain activity, capturing dynamic neural oscillations, and has shown potential in identifying neurophysiological predictors of treatment response to antidepressant medication as well as to repetitive transcranial magnetic stimulation (rTMS) treatments (Watts et al. \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). ​Schwartzmann et al. (\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) developed and validated a multivariate model based on resting-state EEG features which predicted treatment response to SSRI antidepressant in MDD. The model achieved a balanced accuracy of 64.2% in internal validation in the national Canadian Biomarker Integration Network in Depression (CAN-BIND) cohort and 63.7% in external validation in the Establishing Moderators and Biosignatures of Antidepressant Response for Clinical Care (EMBARC) multi-site clinical trial cohort. Notably, when applied to the EMBARC placebo treatment group, the model's balanced accuracy dropped to 48.7%, indicating its specificity to pharmacological treatment response.\u003c/p\u003e \u003cp\u003eMore advanced computational techniques provide deeper analysis of neural dynamics. Measures such as frontal alpha asymmetry (FAA), multiscale entropy (MSE) and inter-site phase clustering (ISPC) capture additional features of brain function, including asymmetry, complexity and connectivity. FAA quantifies the difference in alpha-band power between the left and right frontal cortices and has long been proposed as a marker of affective style. Greater left-sided activity has often been associated with approach-related motivation and positive affect and has been examined as a potential predictor of antidepressant treatment outcomes, though findings have been mixed (Fitzgerald, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Pizzagalli et al. (\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) reported that increased rostral anterior cingulate cortex theta activity is prognostic marker of treatment outcome in EMBARC which has been consistently observed in MDD treatment studies. MSE quantifies the complexity of neural signals across multiple time scales. Higher MSE values at coarse scales and lower MSE at fine scales may reflect greater adaptability and efficiency of brain function (Sporns et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2000\u003c/span\u003e). Higher MSE at coarse scales has been associated with better treatment outcomes to antidepressant medications in MDD (Jaworska et al., 2018) and reductions in MSE at fine time scales following seizure therapy have been associated with symptom improvement (Farzan et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). Zhdanov et al. (\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) identified MSE as among the most informative features for predicting treatment response to escitalopram medication in MDD. ISPC captures the consistency of phase relationships between signals from different brain regions and is commonly used as an index of functional connectivity, with stronger phase synchronisation potentially reflecting greater coordination within neural networks (Cohen, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Increased connectivity at baseline has been associated with better clinical outcomes to antidepressant medication (Olbrich et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2014\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn the present study, we sought to examine the electrophysiological correlates of the two neuroanatomical dimensions (D1 and D2) identified in the COORDINATE-MDD consortium (Fu et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2023\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Xiao et al., \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). We sought to determine whether these structurally defined dimensions exhibit distinct resting-state EEG signatures, which could provide complementary, temporally sensitive markers of underlying neural function. EEG data were eyes-closed resting-state recordings obtained from the Canadian Biomarker Integration Network in Depression (CAN-BIND) (MacQueen et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) and the multi-site clinical trial Establishing Moderators and Biosignatures of Antidepressant Response in Clinical Care (EMBARC) (Trivedi et al., \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2016\u003c/span\u003e) datasets, both of which are part of the COORDINATE-MDD consortium. We further assessed whether the EEG differences were associated with antidepressant treatment response and whether effects vary by treatment type (SSRI and placebo). Specifically, we investigate spectral power, FAA, MSE, and ISPC to identify EEG correlates of structural brain differences and clinical outcomes.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eDatasets\u003c/h2\u003e \u003cp\u003eData were integrated from two independent clinical trials: CAN-BIND (Kennedy et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) and EMBARC (Trivedi et al., \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2016\u003c/span\u003e), within the COORDINATE-MDD consortium (Fu et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Ethical approval was granted by the respective Research Ethics Boards. All participants had provided written informed consent.\u003c/p\u003e \u003cp\u003eCAN-BIND is a multi-site prospective treatment study with four recruitment sites in Canada (Kennedy et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Inclusion criteria included major depressive disorder (MDD) diagnosis, according to DSM-IV-TR criteria, using the Mini-International Neuropsychiatric Interview (MINI), in current depressive episode with Montgomery-\u0026Aring;sberg Depression Rating Scale (MADRS) score of a minimum of 24 and being medication-free. From CAN-BIND, EEG data were available in 138 MDD participants (82 women; mean age 36.24 years, SD 12.80 years) and 54 healthy control participants (35 women; mean age 33.32 years, SD 11.01 years), acquired at baseline before treatment initiation. All MDD participants received treatment with SSRI antidepressant, escitalopram, for 8 weeks. Treatment response was defined as a\u0026thinsp;\u0026gt;\u0026thinsp;=\u0026thinsp;50% reduction in HAMD equivalent scores from baseline to week 8.\u003c/p\u003e \u003cp\u003eEMBARC is a mulit-site prospective, double-blind, randomised controlled trial with four recruitment sites in USA (Trivedi et al., \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). Inclusion criteria included MDD diagnosis based on Structured Clinical Interview for DSM-IV (SCID), being in a current depressive episode of at least moderate severity defined as a 17-item Hamilton Depression Rating Scale (HAMD-17) score of a minimum of 14, and being medication-free. From EMBARC, EEG data were available in 281 MDD participants (185 women; mean age 37.18 years, SD 14.78 years) and 39 healthy control participants (24 women; mean age 37.21 years, SD 13.29 years), acquired at baseline before treatment initiation. In the initial treatment phase, MDD participants were randomised to receive either SSRI, sertraline, or placebo medication for 8 weeks. Treatment response was defined as achieving a\u0026thinsp;\u0026gt;\u0026thinsp;=\u0026thinsp;50% reduction in HAMD-17 score from baseline to week 8.\u003c/p\u003e \u003cp\u003eThe combined dataset initially consisted of 250 MDD participants (CAN-BIND: 54; EMBARC: 196) who had both structural MRI dimension classification (D1, D2) and EEG data. During EEG preprocessing, 13 participant data were excluded due to data quality issues, resulting in a final sample of 237 MDD participants in the present analysis: D1: 116 MDD (70 women; mean age 37.58 years, SD 13.25 years; mean years of education 15.21 years, SD 2.49 years), and D2: 121 MDD (85 women; mean age 37.36 years, SD 13.51 years; mean years of education 14.60 years, SD 2.62 years)). Within CAN-BIND, D1: 15 MDD (9 women); D2: 39 MDD (29 women), and within EMBARC, D1: 101 MDD (61 women); D2: 82 MDD (56 women) (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\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 characteristics of participants\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\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=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eCAN-BIND\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003eEMBARC\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDemographic feature\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDimension 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDimension 2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eDimension 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eDimension 2\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNumber (women)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e15 (9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e39 (29)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e101(61)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e82 (56)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge (years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e34.2 (14.11)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e38.72 (12.96)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e38.08 (13.12)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e36.72 (13.8)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYears of Education\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e13.33 (1.54)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e13.79 (2.61)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e15.5 (2.49)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e14.98 (2.55)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDepressive severity (baseline)\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 \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMADRS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e29.08(5.66)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e30.39(5.28)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHAMD-17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e19.42 (3.82)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e19.68 (3.63)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTreatment\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 \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAntidepressant medication\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e39\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e54\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e33\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePlacebo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e49\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"5\"\u003eValues represent means with standard deviations in parentheses. Age and years of education are in years. Depression severity was assessed using the Montgomery\u0026ndash;\u0026Aring;sberg Depression Rating Scale (MADRS) in CAN-BIND and the 17-item Hamilton Depression Rating Scale (HAMD-17) in EMBARC. HAMD-17 scores reported for CAN-BIND are estimated from MADRS using equipercentile linking (Leucht et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) to allow comparison across studies. In CAN-BIND, all participants received escitalopram (ESC). In EMBARC, participants were randomised to receive either sertraline (SER) or placebo (PLA). Abbreviations: ESC\u0026thinsp;=\u0026thinsp;escitalopram; SER\u0026thinsp;=\u0026thinsp;sertraline; PLA\u0026thinsp;=\u0026thinsp;placebo; HAMD\u0026thinsp;=\u0026thinsp;Hamilton Depression Rating Scale; MADRS\u0026thinsp;=\u0026thinsp;Montgomery\u0026ndash;\u0026Aring;sberg Depression Rating Scale.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eMean baseline depressive severity scores were CAN-BIND D1 (MADRS mean\u0026thinsp;=\u0026thinsp;28.10, SD\u0026thinsp;=\u0026thinsp;5.22), CAN-BIND D2 (MADRS mean\u0026thinsp;=\u0026thinsp;30.18, SD\u0026thinsp;=\u0026thinsp;5.45), EMBARC D1 (HAMD-17 mean\u0026thinsp;=\u0026thinsp;19.42, SD\u0026thinsp;=\u0026thinsp;3.82), and EMBARC D2 (HAMD-17 mean\u0026thinsp;=\u0026thinsp;19.68, SD\u0026thinsp;=\u0026thinsp;3.63). Following treatment: CAN-BIND D1 (MADRS mean\u0026thinsp;=\u0026thinsp;17.53, SD\u0026thinsp;=\u0026thinsp;5.36), CAN-BIND D2 (MADRS mean\u0026thinsp;=\u0026thinsp;17.67, SD\u0026thinsp;=\u0026thinsp;5.89), EMBARC D1 (HAMD-17 mean\u0026thinsp;=\u0026thinsp;11.63, SD\u0026thinsp;=\u0026thinsp;7.50) and EMBARC D2 (HAMD-17 mean\u0026thinsp;=\u0026thinsp;11.50, SD\u0026thinsp;=\u0026thinsp;6.29). In CAN-BIND, all MDD participants received SSRI medication (escitalopram). In EMBARC, participants were randomized to either SSRI medication (D1: 54 participants; D2: 33 participants) or placebo (D1: 47 participants; D2: 49 participants). Treatment response was attained following the course of treatments in, D1: 50 MDD (34 women) and D2: 48 MDD (34 women), and persistent depressive symptoms were observed in, D1: 66 MDD (36 women) and D2: 73 MDD (51 women) (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eDemographic characteristics by Dimension (D1, D2) and treatment response status\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\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=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eDimension 1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003eDimension 2\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTreatment response\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePersistent symptoms\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eTreatment response\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003ePersistent symptoms\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNumber (women)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e50 (34)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e66 (36)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e48 (34)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e73 (51)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge (years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e35.02 (12.04)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e39.52 (13.87)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e37.42 (14.42)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e37.33 (12.98)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYear of Education\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e15.32 (2.32)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e15.14 (2.63)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e14.43 (2.81)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e14.71 (2.51)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHAMD baseline\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e19.74 (4.00)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e19.40 (3.64)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e20.71 (3.79)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e20.47 (4.19)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHAMD post-treatment\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4.90 (2.93)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e17.05 (4.90)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5.79 (3.62)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e16.29 (4.60)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTreatment\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 \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEMBARC placebo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e31\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e30\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEMBARC medication\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e29\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e14\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e19\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCAN-BIND medication\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e24\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"5\"\u003eValues represent means with standard deviations in parentheses. Age and years of education are in years. Depression severity was assessed using the Montgomery\u0026ndash;\u0026Aring;sberg Depression Rating Scale (MADRS) in CAN-BIND and the 17-item Hamilton Depression Rating Scale (HAMD-17) in EMBARC. HAMD-17 scores for CAN-BIND were estimated from MADRS using equipercentile linking (Leucht et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) for harmonisation. Original MADRS baseline and post-treatment values for CAN-BIND are in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. In CAN-BIND, participants received escitalopram (ESC). In EMBARC, participants were randomised to sertraline (SER) or placebo (PLA). Abbreviations: ESC\u0026thinsp;=\u0026thinsp;escitalopram; SER\u0026thinsp;=\u0026thinsp;sertraline; PLA\u0026thinsp;=\u0026thinsp;placebo; HAMD\u0026thinsp;=\u0026thinsp;Hamilton Depression Rating Scale; MADRS\u0026thinsp;=\u0026thinsp;Montgomery\u0026ndash;\u0026Aring;sberg Depression Rating Scale.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eEEG data recording\u003c/h3\u003e\n\u003cp\u003eIn CAN-BIND, participants underwent an 8-minute resting-state EEG recording with their eyes closed at baseline (Farzan et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Lam et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). In EMBARC, resting-state EEG was acquired in four 2-minute blocks at baseline, consisting of two eye-open segments and two eye-closed segments (Trivedi et al., \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). Data from the eyes-closed recordings only were included to ensure consistency across datasets.\u003c/p\u003e\n\u003ch3\u003eEEG data preprocessing\u003c/h3\u003e\n\u003cp\u003eEEG recordings were preprocessed to minimize artifacts and standardize the dataset for analysis. Data were first segmented into 2-second epochs to facilitate artifact detection and spectral analysis. To mitigate edge effects from filtering, each epoch was forward and backward reflected before applying standard EEG preprocessing procedures. EEG data were then high-pass filtered at 0.5 Hz using the MATLAB \u003cem\u003ehighpass\u003c/em\u003e function to remove slow drifts, low-pass filtered at 200 Hz using the \u003cem\u003elowpass\u003c/em\u003e function to remove high-frequency noise, and notch filtered at either 60 Hz or 50 Hz using \u003cem\u003edesignNotchPeakIIR\u003c/em\u003e and filtfilt to eliminate line noise. Specifically, a 60 Hz notch filter was applied to EEG data collected from both the EMBARC (United States) and CAN-BIND (Canada) datasets, consistent with the local electrical frequency.\u003c/p\u003e \u003cp\u003eTo ensure data quality, a voltage deflection threshold approach was applied for identifying and rejecting noisy epochs. A moving 80 ms window was used to compute the difference between the maximum and minimum voltage within each epoch. Epochs were flagged as potentially noisy if any channel exhibited a voltage deflection exceeding 100 \u0026micro;V. After noisy epochs were identified, hannels exhibiting high-amplitude artifacts, defined as exceeding 100 \u0026micro;V threshold in more than 50% trials, were identified and removed. Artifact rejection was then applied to the remaining data to exclude trials containing residual noise, having more than 25% of remaining channels exceeding 100 \u0026micro;V noise threshold.\u003c/p\u003e \u003cp\u003eAt the participant level, additional data quality criteria were applied to ensure reliable EEG signals base on the following criteria: original EEG recording contained fewer than 20 usable EEG channels or if more than 50% of channels were excluded due to excessive noise contamination. Based on these criteria, from the initial sample of 250 MDD, 13 MDD participant data were excluded, and the final dataset consisted of 237 MDD participants (Tables\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eTo ensure consistency between datasets, EMBARC EEG data were re-referenced to an average reference, and 58 common electrodes from the eyes-closed condition were retained to match the CAN-BIND dataset. Subsequently, all EEG recordings were re-referenced to the average reference and spatially interpolated to a standardized 32-channel montage to enable harmonized analyses.\u003c/p\u003e\n\u003ch3\u003eAnalysis of primary EEG parameters\u003c/h3\u003e\n\u003cp\u003eResting-state EEG data were analyzed to extract four primary features: power spectra, frontal alpha asymmetry (FAA), multiscale sample entropy (MSE), and inter-site phase clustering (ISPC). All feature computations were performed using MATLAB scripts adapted from the publicly available Sheffield Autism Biomarkers toolbox (Dede et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2025\u003c/span\u003e), which has been previously applied to large-scale clinical EEG datasets. The original functions are available via GitHub (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/adede1988/SheffieldAutismBiomarkers\u003c/span\u003e\u003cspan address=\"https://github.com/adede1988/SheffieldAutismBiomarkers\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e).\u003c/p\u003e\n\u003ch3\u003ePower Spectra\u003c/h3\u003e\n\u003cp\u003ePower spectra were computed separately for each epoch. To minimize edge artifacts, epochs were mirrored, and time-frequency decomposition was performed using wavelet convolution across 100 logarithmically spaced frequencies (2\u0026ndash;80 Hz) (Cohen, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). The resulting complex time series was converted to power values, and mirrored segments were removed. Mean power values were then averaged across epochs, yielding a 32 (electrodes) \u0026times; 100 (frequencies) matrix per participant. For power spectrum comparisons, spectra from each channel were averaged within six standard frequency bands: delta (2\u0026ndash;4 Hz), theta (4\u0026ndash;8 Hz), alpha (8\u0026ndash;14 Hz), beta (14\u0026ndash;30 Hz) and gamma (30\u0026ndash;80 Hz).\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eFrontal Alpha Asymmetry\u003c/h2\u003e \u003cp\u003eFrontal Alpha Asymmetry (FAA) was computed using log-transformed power values in the alpha band (8\u0026ndash;14 Hz). Alpha power was averaged across bilateral frontal electrodes: F3 and F7 (left hemisphere), and F4 and F8 (right hemisphere). FAA was calculated as the difference in log power between homologous electrode pairs (i.e., F4-F3 and F8-F7) (Allen et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2004\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eMultiscale Sample Entropy\u003c/h3\u003e\n\u003cp\u003eMultiscale sample entropy (MSE) was computed to quantify signal complexity across multiple temporal scales. Data were resampled to 1000 Hz and coarse-grained into 20-time scales. Sample entropy was estimated using a similarity threshold of 30% of the signal's standard deviation and a pattern length of m\u0026thinsp;=\u0026thinsp;2 (Costa et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2005\u003c/span\u003e). Prior to entropy estimation, signals were normalized to reduce amplitude-related bias. The final output for each participant was a 32-channel \u0026times; 20-time-scale matrix, averaged across epochs.\u003c/p\u003e\n\u003ch3\u003eInter-Site Phase Clustering\u003c/h3\u003e\n\u003cp\u003eInter-site phase clustering (ISPC) was used to assess phase synchrony between electrode pairs across the same 100 frequencies used for spectral power. Signals were first transformed using a surface Laplacian to improve spatial resolution. ISPC was then calculated based on the consistency of phase differences between electrode pairs across time (Cohen, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). ISPC was calculated according to the following formula:\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eISPC = \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\frac{1}{n}\\left|{\\sum\\:}_{t=1}^{n}{e}^{i\\left({\\phi\\:}_{xt}-{\\phi\\:}_{yt}\\right)}\\right|\\)\u003c/span\u003e\u003c/span\u003e\u003c/h2\u003e \u003cp\u003ewhere \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\phi\\:}_{xt}\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\phi\\:}_{yt}\\)\u003c/span\u003e\u003c/span\u003e represent the phase angles of signals from two channels, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:x\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:y\\)\u003c/span\u003e\u003c/span\u003e, at time \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:t\\:\\)\u003c/span\u003e\u003c/span\u003e. The result was a 32 \u0026times; 32 \u0026times; 100 matrix per participant, averaged across epochs.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eAnalyses of EEG-derived features\u003c/h2\u003e \u003cp\u003eResting-state EEG data were harmonized across datasets into 32 channels and processed to extract a range of spectral and complexity-based features for group-level analysis.\u003c/p\u003e \u003cp\u003eEEG-derived features were analyzed with respect to neuroanatomical dimension (D1; D2), treatment outcome status (responder (R), defined as \u0026gt;\u0026thinsp;=\u0026thinsp;50% improvement in depressive symptoms from baseline; non-responder (NR), defined as having persistent symptoms with \u0026lt;\u0026thinsp;50% improvement from baseline). EEG features were analyzed separately within the full sample and within participants stratified by treatment group (antidepressant medication or placebo), to account for potential confounding between treatment type and dataset source.\u003c/p\u003e \u003cp\u003eFor each analysis, group-level comparisons were conducted using two-way Analysis of Covariance (ANCOVA) models, with each EEG-derived feature entered as a dependent variable. Fixed factors were Dimension (D1; D2), Outcome (R; NR), and age and sex were covariates.\u003c/p\u003e \u003cp\u003eEEG features were standardized across participants. Where appropriate, log-transformations were applied to reduce skewness, and outliers exceeding\u0026thinsp;\u0026plusmn;\u0026thinsp;5 standard deviations (z-score) were excluded. Effect sizes were quantified using eta-squared (η\u0026sup2;), representing the proportion of variance explained by each main effect or interaction. Parameter estimates (β), 95% confidence intervals (CIs), and corresponding \u003cem\u003ep\u003c/em\u003e-values were extracted for all predictors. To control multiple comparison problems across features, False Discovery Rate (FDR) correction was applied. All statistical analyses and visualizations were performed in R (R Core Team, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), with graphical outputs generated using the ggplot2 package (Wickham, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2016\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eThreshold-free cluster enhancement (TFCE) analysis\u003c/h2\u003e \u003cp\u003eThreshold-free cluster enhancement (TFCE) is a nonparametric, cluster-based analysis which accounts for spatial and spectral dependencies in EEG data (Smith and Nichols, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2009\u003c/span\u003e; Mensen and Khatami, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). TFCE has higher sensitivity to weak but spatially or temporally extended effects and does not require predefined cluster-forming thresholds. TFCE was applied using a two-tailed, permutation-based cluster testing framework. EEG-derived metrics, including spectral power, multiscale sample entropy (MSE), and inter-site phase clustering (ISPC), were evaluated across all channels and either frequency bins (for power and ISPC) or time scales (for MSE). TFCE parameters were configured based on Mensen and Khatami (\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2013\u003c/span\u003e), with the height exponent (H) set to 2 and the extent exponent (E) set to 0.66. Statistical inference was performed using Monte Carlo permutation testing with 5,000 iterations. All analyses were conducted using the ept_TFCE function from the TFCE MATLAB toolbox (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/Mensen/ept_TFCE-matlab\u003c/span\u003e\u003cspan address=\"https://github.com/Mensen/ept_TFCE-matlab\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). Comparisons were conducted across combinations of neuroanatomical dimensions (D1; D2) and Outcomes (R; NR), including both overall group comparisons (e.g., D1 vs. D2; R vs. NR) and subgroup contrasts within and across dimensions by response combinations. To further explore the robustness of significant findings, follow-up TFCE analyses were also conducted separately within the antidepressant medication (ADM) and placebo (PLA) subgroups.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003ePower Spectra\u003c/h2\u003e \u003cp\u003eANCOVA conducted on the full sample revealed a significant interaction between neuroanatomical dimension and clinical outcome for relative gamma power at electrode CP6 (\u003cem\u003eη\u0026sup2;\u003c/em\u003e = 0.043, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.035 FDR-corrected). Relative gamma power was lower in D1 responders (mean = -0.073, SE\u0026thinsp;=\u0026thinsp;0.144) than in D1 non-responders (mean\u0026thinsp;=\u0026thinsp;0.074, SE\u0026thinsp;=\u0026thinsp;0.122), while values were comparable between D2 responders (mean\u0026thinsp;=\u0026thinsp;0.071, SE\u0026thinsp;=\u0026thinsp;0.154) and D2 non-responders (mean\u0026thinsp;=\u0026thinsp;0.084, SE\u0026thinsp;=\u0026thinsp;0.116) (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Post hoc pairwise comparisons did not reach statistical significance following FDR correction.\u003c/p\u003e \u003cp\u003eNo significant main effects of dimension or clinical outcome status were observed in absolute, relative, or log-transformed spectral power features across the full sample (Supplementary Table\u0026nbsp;1). Similarly, stratified analyses with the antidepressant-treated and placebo-treated subgroups did not reveal any significant main effects or interaction (Supplementary Tables\u0026nbsp;2 and 3).\u003c/p\u003e \u003cp\u003eTFCE analyses revealed significant group-level differences in EEG power across neuroanatomical dimensions (Supplementary Table\u0026nbsp;4). Participants classified in D1 showed significantly greater absolute theta and alpha power in frontal regions, elevated theta, alpha, and beta power in central regions, and increased beta power in parieto-occipital regions relative to those in D2. In contrast, D1 participants showed significantly lower relative delta power across frontal, central, and parieto-occipital regions compared to D2. For log-transformed power, D1 participants showed significantly greater theta, alpha, and beta power across widespread scalp regions. No significant TFCE-based differences in spectral power were observed as a function of clinical outcome status alone.\u003c/p\u003e \u003cp\u003eAmong participants who showed a clinical response, baseline TFCE analysis showed increased absolute and log-transformed alpha power in frontal and central regions, elevated log-transformed theta power in frontal regions, decreased relative delta power in frontal regions, and decreased beta power in frontal regions in D1 compared to D2 (Supplementary Table\u0026nbsp;4). No significant dimension-related differences were observed within the non-responder subgroup.\u003c/p\u003e \u003cp\u003eWhen stratified by treatment modality, SSRI-treated responders in D1 showed significantly greater absolute theta and alpha power across frontal, central, and parieto-occipital electrodes compared to D2 responders (Supplementary Table\u0026nbsp;5). In contrast, placebo-treated responders in D1 exhibited significantly lower relative delta power and increased relative alpha power in frontal, central, and posterior regions compared to D2 responders (Supplementary Table\u0026nbsp;6).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eFrontal Alpha Asymmetry\u003c/h2\u003e \u003cp\u003eThere was a significant main effect of responder status on frontal alpha asymmetry (FAA) (\u003cem\u003eF\u003c/em\u003e(1, 227)\u0026thinsp;=\u0026thinsp;5.10, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.025 FDR-corrected, \u003cem\u003eη\u0026sup2;\u003c/em\u003e = 0.022), in which responders exhibited lower FAA values compared to non-responders at baseline, supported by post hoc Tukey\u0026rsquo;s HSD testing (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.025 FDR-corrected).\u003c/p\u003e \u003cp\u003eNo significant main effects were found for neuroanatomical dimension (\u003cem\u003eF\u003c/em\u003e(1, 227)\u0026thinsp;=\u0026thinsp;0.25, p\u0026thinsp;=\u0026thinsp;0.620 FDR-corrected, η\u0026sup2; = 0.001) or treatment type (\u003cem\u003eF\u003c/em\u003e(1, 227)\u0026thinsp;=\u0026thinsp;0.60, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.440 FDR-corrected, \u003cem\u003eη\u0026sup2;\u003c/em\u003e = 0.003), nor were any two-way or three-way interactions statistically significant (all \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.20).\u003c/p\u003e \u003cp\u003eAs FAA is derived from an individual summary index across predefined electrode pairs and not assessed across spatially extended features or frequency bins, it was not suitable for TFCE analysis.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eMultiscale Sample Entropy\u003c/h2\u003e \u003cp\u003eAt baseline, D1 responders (across all treatment types) showed significantly higher MSE, particularly at coarse time scales (time scales 14\u0026ndash;20) over fronto-central regions compared to D2 responders (Supplementary Table\u0026nbsp;4). This effect was also observed when stratifying by treatment group. In participants receiving placebo treatment, D1 responders showed elevated MSE across a wider range of electrodes (frontal, central, parietal, and occipital) and broader scale range (time scales 6\u0026ndash;20). All results were significant at p\u0026thinsp;\u0026lt;\u0026thinsp;.046 (TFCE-corrected) (Supplementary Table\u0026nbsp;6). In contrast, no significant baseline MSE differences were observed between D1 and D2 responders in the antidepressant medication group. No significant differences in MSE were found between non-responders in either treatment group.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eInter-Site Phase Clustering\u003c/h2\u003e \u003cp\u003eANCOVA analyses in the full sample revealed a significant main effect of treatment response status in theta-band ISPC at electrode F8 (\u003cem\u003eη\u0026sup2;\u003c/em\u003e = 0.048, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.027 FDR-corrected), with responders demonstrating increased ISPC at baseline relative to non-responders (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e; Supplementary Table\u0026nbsp;1).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThere were no significant main effects of dimension, nor any interaction effects. Subgroup analysis by treatment type (antidepressant medication, placebo) similarly showed no significant effects (Supplementary Tables\u0026nbsp;2 and 3). TFCE analyses did not reveal any significant differences in ISPC across groups (Supplementary Table\u0026nbsp;4).\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThe present study sought to investigate whether MDD neuroanatomical dimensions previously derived from structural MRI in the COORDINATE-MDD consortium, correspond to distinct patterns of EEG metrics from resting-state eyes-closed brain activity, namely spectral power, frontal alpha asymmetry (FAA), multiscale entropy (MSE), and inter-site phase clustering (ISPC). We found that the structural dimensions, D1 characterized by relatively preserved brain volumes, and D2 associated with widespread grey and white matter reductions, also differ in electrophysiological dynamics, in particular among participants who showed a clinical response to antidepressant treatment.\u003c/p\u003e \u003cp\u003eEEG power analyses using TFCE revealed that MDD individuals classified within D1 showed elevated absolute theta and alpha power in frontal regions, increased beta power in both central and parieto-occipital regions, as well as widespread elevations in log-transformed theta, alpha and beta power, as well as significantly lower relative delta power across frontal and posterior regions compared to D1. Increases in alpha are associated with top-down inhibitory control that facilitates selective attention by suppressing task-irrelevant brain areas (Jensen and Mazaheri, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2010\u003c/span\u003e). While alpha oscillations are typically associated with global corticocortical dynamics, delta oscillations may index more locally dominant neocortical processes (Nunez, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e1995\u003c/span\u003e; Nunez Srinivasan, 2006). The pattern of reduced relative delta power and increased absolute alpha power observed in D1 individuals may indicate a shift from locally driven cortical inhibition toward more global, top-down regulatory control. This shift could support more flexible and selective engagement of task-relevant brain regions, facilitating improved cognitive control and emotion regulation.\u003c/p\u003e \u003cp\u003eIn participants who subsequently showed a treatment response, this spectral pattern was similarly observed at baseline, indicating that this EEG profile may be predictive of clinical response. Increased alpha power, particularly in posterior and prefrontal regions, has been observed in in treatment responders across a range of modalities, including SSRI antidepressant medication (Bruder et al., \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2008\u003c/span\u003e) and neuromodulatory approaches such as repetitive transcranial magnetic stimulation (rTMS; Noda et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). Wu et al. (\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) found that an alpha-based EEG signature predicted response to antidepressant medication but not placebo in EMBARC. In the present study, alpha power increase was observed only in D1 responders at baseline in the antidepressant treatment arm, not in the placebo treatment group, suggesting a treatment-specific electrophysiological profile.\u003c/p\u003e \u003cp\u003eThe observation of reduced relative delta power in D1 responders at baseline has been observed for SSRI antidepressant medication (Iosifescu et al., \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2009\u003c/span\u003e) and cognitive behavioral therapy (Schwartzmann et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Delta oscillations, particularly in frontal and medial prefrontal regions, have been associated with internal attention and inhibitory control, especially during cognitively demanding tasks such as mental arithmetic and working memory (Harmony, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). In such contexts, elevated delta is thought to support internal attention by inhibiting irrelevant environmental input. Increased delta activity has also been observed during states of reduced cognitive and emotional engagement, such as meditation, where it might reflect cortical inhibition or downregulation (Tei et al., \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2009\u003c/span\u003e), a more active and engaged neural state which could facilitate adaptive processing and responsiveness. Notably, no significant power differences were observed between D1 and D2 non-responders, supporting the specificity of these effects to positive treatment outcomes.\u003c/p\u003e \u003cp\u003eIn frontal alpha asymmetry (FAA), a main effect of clinical response status was observed at baseline in which MDD participants who subsequently showed a good clinical response to have a lower FAA, reflecting greater left-lateralised frontal alpha activity, compared to those who had persistent symptoms following treatment. The approach-withdrawal model proposes two motivational systems corresponding to positive and negative affects which are reflected in relative lateralisation of EEG activity (Henriques and Davidson, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e1991\u003c/span\u003e). Lower FAA has been observed in MDD, which has been hypothesised to be associated with anhedonia (Nusslock et al, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2015\u003c/span\u003e), however van der Vinne et al. (\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2017\u003c/span\u003e) meta-analysis reports a small effect which was not significant citing high heterogeneity across studies.\u003c/p\u003e \u003cp\u003eMultiscale entropy (MSE) analysis revealed greater signal complexity at coarse time scales in D1 responders at baseline compared to D2 responders, consistent with proposals that greater neural complexity reflects more flexible and adaptive brain states (von Stein and Sarnthein, \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2000\u003c/span\u003e; Schnitzler and Gross, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2005\u003c/span\u003e). Greater coarse-scale entropy has been associated with better clinical outcomes, which predicted clinical outcome in CAN-BIND (Zhdanov et al., \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) as well as in EMBARC (Schwartzmann et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Among participants who showed a response to placebo treatment, we found increased MSE in D1 at baseline compared to D2, suggesting that coarse-scale signal complexity may reflect a more general response which includes non-pharmacological treatments.\u003c/p\u003e \u003cp\u003eInter-site phase clustering (ISPC) analyses indicated higher theta-band synchrony at electrode F8 in responders at baseline compared to non-responders, implicating increased frontotemporal connectivity in treatment response. Pizzagalli et al. (\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) observed increased frontal theta activity, in particular in the rostral anterior cingulate cortex, predicted clinical response to antidepressant medication, which has been consistently implicated as a predictor of treatment response (Pizzagalli, \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2011\u003c/span\u003e; Fu et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). Increased synchrony in these low-frequency bands may indicate stronger coordination within frontoparietal networks.\u003c/p\u003e \u003cp\u003eThese findings suggest that D1 may reflect a more neurophysiologically adaptive profile compared to D2 which is characterised by preserved oscillatory dynamics aligning with the neuroanatomical classification (Fu et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) and external validation in UK Biobank (Xiao et al., \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). D1 could reflect a biologically distinct, more resilient MDD dimension with preserved brain structure and function, while D2 was associated with subtle but widespread cortical atrophy, increased systemic inflammation, metabolic disturbances, and a history of greater childhood adversity, suggesting a pathophysiological profile marked by greater vulnerability in neurodevelopmental and immuno-metabolic pathways (Fu et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Xiao et al., \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eLimitations in the present study include the sample size which was limited to available data in resting-state EEG which may not capture the full complexity of neural dynamics. While resting-state metrics are robust and clinically accessible, task-based or dynamic EEG measures, such as during emotional or cognitive tasks, might provide additional information about functional differences between dimensions and treatment responsiveness (Webb et al., 2016). Incorporation of longitudinal designs would clarify predictive and state-dependent effects.\u003c/p\u003e \u003cp\u003eIn summary, the present study observed that the neuroanatomical dimensions of MDD, previously identified via structural MRI, are reflected in distinct resting-state EEG profiles. Individuals classified within D1 showed increased absolute and log-transformed theta, alpha, and beta power alongside reduced relative delta power, a pattern suggesting enhanced global cortical regulation and reduced local inhibition. These spectral features were associated with antidepressant treatment response. These findings suggest that D1 may reflect a biologically and neurophysiologically more resilient MDD dimension, associated with preserved brain structure, lower inflammation, and more adaptive electrophysiological profiles that confer improved treatment responsiveness. In contrast, D2 appears linked to subtle widespread cortical atrophy, immuno-metabolic dysregulation, and a history of higher childhood adversity, consistent with greater vulnerability.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAllen JJ, Urry HL, Hitt SK, Coan JA (2004) The stability of resting frontal electroencephalographic asymmetry in depression. Psychophysiology 41(2):269\u0026ndash;280. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/j.1469-8986.2003.00149.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1469-8986.2003.00149.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBruder GE, Sedoruk JP, Stewart JW, McGrath PJ, Quitkin FM, Tenke CE (2008) Electroencephalographic alpha measures predict therapeutic response to a selective serotonin reuptake inhibitor antidepressant: Pre- and post-treatment findings. Biol Psychiatry 63(12):1171\u0026ndash;1177. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.biopsych.2007.10.009\u003c/span\u003e\u003cspan address=\"10.1016/j.biopsych.2007.10.009\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eČukić M, Stokić M, Radenković S, Ljubisavljević M, Simić S, Savić D (2020) Nonlinear analysis of EEG complexity in episode and remission phase of recurrent depression. Int J Methods Psychiatr Res 29(2):e1816. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/mpr.1816\u003c/span\u003e\u003cspan address=\"10.1002/mpr.1816\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCohen MX (2014) Analyzing neural time series data: Theory and practice. MIT Press\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCosta M, Goldberger AL, Peng C-K (2005) Multiscale entropy analysis of biological signals. Phys Rev E 71(2 Pt 1):021906. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1103/PhysRevE.71.021906\u003c/span\u003e\u003cspan address=\"10.1103/PhysRevE.71.021906\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDede AJ, Xiao W, Vaci N et al (2025) Exploring EEG resting state differences in autism: Sparse findings from a large cohort. Mol Autism 16:13. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s13229-025-00647-3\u003c/span\u003e\u003cspan address=\"10.1186/s13229-025-00647-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFarzan F, Atluri S, Frehlich M et al (2017) Standardization of electroencephalography for multi-site, multi-platform and multi-investigator studies: Insights from the Canadian Biomarker Integration Network in Depression. Sci Rep 7:7473. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41598-017-07613-x\u003c/span\u003e\u003cspan address=\"10.1038/s41598-017-07613-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFarzan F, Atluri S, Mei Y, Moreno S, Levinson AJ, Blumberger DM, Daskalakis ZJ (2017) Brain temporal complexity in explaining the therapeutic and cognitive effects of seizure therapy. Brain 140(4):1011\u0026ndash;1025. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/brain/awx030\u003c/span\u003e\u003cspan address=\"10.1093/brain/awx030\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFitzgerald PJ (2024) Frontal alpha asymmetry and its modulation by monoaminergic neurotransmitters in depression. Clin Psychopharmacol Neurosci 22(3):405\u0026ndash;415. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.9758/cpn.23.1138\u003c/span\u003e\u003cspan address=\"10.9758/cpn.23.1138\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFried EI, Nesse RM (2015) Depression is not a consistent syndrome: An investigation of unique symptom patterns in the STARD study. J Affect Disord 172:96\u0026ndash;102. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jad.2014.10.010\u003c/span\u003e\u003cspan address=\"10.1016/j.jad.2014.10.010\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFu CHY, Steiner H, Costafreda SG (2013) Predictive neural biomarkers of clinical response in depression: A meta-analysis of functional and structural neuroimaging studies of pharmacological and psychological therapies. Neurobiol Dis 52:75\u0026ndash;83. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.nbd.2012.05.008\u003c/span\u003e\u003cspan address=\"10.1016/j.nbd.2012.05.008\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFu CHY, Fan Y, Davatzikos C (2019) Addressing heterogeneity (and homogeneity) in treatment mechanisms in depression and the potential to develop diagnostic and predictive biomarkers. NeuroImage: Clin 24:101997. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.nicl.2019.101997\u003c/span\u003e\u003cspan address=\"10.1016/j.nicl.2019.101997\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFu CHY, Erus G, Fan Y, Antoniades M, Arnone D, Arnott SR, Chen T, Choi KS, Fatt CC, Frey BN, Frokjaer VG, Ganz M, Garcia J, Godlewska BR, Hassel S, Ho K, McIntosh AM, Qin K, Rotzinger S, Davatzikos C (2023) AI-based dimensional neuroimaging system for characterizing heterogeneity in brain structure and function in major depressive disorder: COORDINATE-MDD consortium design and rationale. BMC Psychiatry 23:59. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s12888-022-04509-7\u003c/span\u003e\u003cspan address=\"10.1186/s12888-022-04509-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFu CHY, Fan Y, Davatzikos C, Erus G, Godlewska BR, Woodham RD, Xiao W, Choi KS, Frokjaer VG, Sacchet MD, Savitz J, Qin K, McIntosh AM (2024) Neuroanatomical dimensions in medication-free individuals with major depressive disorder and treatment response to SSRI antidepressant medications or placebo. Nat Mental Health. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s44220-023-00187-w\u003c/span\u003e\u003cspan address=\"10.1038/s44220-023-00187-w\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHarmony T (2013) The functional significance of delta oscillations in cognitive processing. Front Integr Nuerosci 7:83. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fnint.2013.00083\u003c/span\u003e\u003cspan address=\"10.3389/fnint.2013.00083\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHenriques JB, Davidson RJ (1991) Left frontal hypoactivation in depression. J Abnorm Psychol 100(4):535\u0026ndash;545. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1037/0021-843X.100.4.535\u003c/span\u003e\u003cspan address=\"10.1037/0021-843X.100.4.535\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIosifescu DV, Greenwald S, Devlin P, Mischoulon D, Denninger JW, Alpert JE, Fava M (2009) Frontal EEG predictors of treatment outcome in major depressive disorder. Eur Neuropsychopharmacol 19(11):772\u0026ndash;777. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.euroneuro.2009.06.001\u003c/span\u003e\u003cspan address=\"10.1016/j.euroneuro.2009.06.001\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJaworska N, Blondeau C, Tessier P, Norris S, Fusee W, Blier P, Knott V (2013) Response prediction to antidepressants using scalp and source-localized loudness dependence of auditory evoked potential (LDAEP) slopes. Prog Neuropsychopharmacol Biol Psychiatry 44:100\u0026ndash;107. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.pnpbp.2013.01.012\u003c/span\u003e\u003cspan address=\"10.1016/j.pnpbp.2013.01.012\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJaworska N, de la Salle S, Ibrahim MH, Blier P, Knott V (2019) Leveraging machine learning approaches for predicting antidepressant treatment response using electroencephalography (EEG) and clinical data. Front Psychiatry 9:768. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fpsyt.2018.00768\u003c/span\u003e\u003cspan address=\"10.3389/fpsyt.2018.00768\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJensen O, Mazaheri A (2010) Shaping functional architecture by oscillatory alpha activity: Gating by inhibition. Front Hum Neurosci 4:186. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fnhum.2010.00186\u003c/span\u003e\u003cspan address=\"10.3389/fnhum.2010.00186\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKennedy SH, Lam RW, Rotzinger S, Milev RV, Blier P, Downar J, Evans KR, Foster JA, MacQueen GM, Parikh SV, Uher R (2019) Symptomatic and functional outcomes and early prediction of response to escitalopram monotherapy and sequential adjunctive aripiprazole therapy in patients with major depressive disorder: A CAN-BIND-1 report. J Clin Psychiatry 80(2):18m12202. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.4088/JCP.18m12202\u003c/span\u003e\u003cspan address=\"10.4088/JCP.18m12202\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLam RW, Milev R, Rotzinger S, Andreazza AC, Blier P, Brenner C, Daskalakis Z, Dharsee M, Downar J, Evans KR, Farzan F, Foster JA, Frey BN, Giacobbe P, Hall GB, Ismail Z, Kennedy SH, MacQueen GM, McAndrews MP, Young LT (2016) Discovering biomarkers for antidepressant response: Protocol from the Canadian Biomarker Integration Network in Depression (CAN-BIND) and clinical characteristics of the first patient cohort. BMC Psychiatry 16:105. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s12888-016-0785-x\u003c/span\u003e\u003cspan address=\"10.1186/s12888-016-0785-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLamers F, Milaneschi Y, Vinkers CH, Schoevers RA, Giltay EJ, Penninx BWJH (2020) Depression profilers and immuno-metabolic dysregulation: Longitudinal results from the NESDA study. Brain Behav Immun 88:174\u0026ndash;183. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.bbi.2020.04.002\u003c/span\u003e\u003cspan address=\"10.1016/j.bbi.2020.04.002\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLeucht S, Fennema H, Engel RR, Kaspers-Janssen M, Szegedi A (2018) Translating the HAM-D into the MADRS and vice versa with equipercentile linking. J Affect Disord 226:326\u0026ndash;331. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jad.2017.09.042\u003c/span\u003e\u003cspan address=\"10.1016/j.jad.2017.09.042\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMacQueen GM, Hassel S, McIntosh AR, Giacobbe P, Rizvi SJ, Kennedy SH, Lam RW, Milev RV, Parikh SV, Rotzinger S, Uher R, Young LT (2019) The Canadian Biomarker Integration Network in Depression (CAN-BIND): Magnetic resonance imaging protocols. J Psychiatry Neurosci 44(4):223\u0026ndash;236. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1503/jpn.180202\u003c/span\u003e\u003cspan address=\"10.1503/jpn.180202\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMensen A, Khatami R (2013) Advanced EEG analysis using threshold-free cluster-enhancement and non-parametric statistics. NeuroImage 67:111\u0026ndash;118. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.neuroimage.2012.10.027\u003c/span\u003e\u003cspan address=\"10.1016/j.neuroimage.2012.10.027\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNoda Y, Nakamura M, Saeki T, Iwata Y, Ikeda Y, Aubert G, Honda S, Yoshino A, Kato TA (2013) Potentiation of quantitative electroencephalograms following prefrontal repetitive transcranial magnetic stimulation in patients with major depression. Neurosci Res 77(1\u0026ndash;2):70\u0026ndash;77. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.neures.2013.06.002\u003c/span\u003e\u003cspan address=\"10.1016/j.neures.2013.06.002\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNunez PL (1995) Neocortical dynamics and human EEG rhythms. Oxford University Press\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNunez PL, Srinivasan R (2006) A theoretical basis for standing and traveling brain waves measured with human EEG with implications for an integrated consciousness. Clin Neurophysiol 117(11):2424\u0026ndash;2435. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.clinph.2006.06.754\u003c/span\u003e\u003cspan address=\"10.1016/j.clinph.2006.06.754\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNusslock R, Walden K, Harmon-Jones E (2015) Asymmetrical frontal cortical activity associated with differential risk for mood and anxiety disorder symptoms: An RDoC perspective. \u003cem\u003eInternational Journal of Psychophysiology, 98\u003c/em\u003e(2, Part 2), 249\u0026ndash;261. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ijpsycho.2015.06.004\u003c/span\u003e\u003cspan address=\"10.1016/j.ijpsycho.2015.06.004\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOlbrich S, Tr\u0026auml;nkner A, Chittka T, Hegerl U, Sch\u0026ouml;nknecht P (2014) Functional connectivity in major depression: Increased phase synchronization between frontal cortical EEG-source estimates. Psychiatry Research: Neuroimaging 222(1\u0026ndash;2):91\u0026ndash;99. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.pscychresns.2014.02.010\u003c/span\u003e\u003cspan address=\"10.1016/j.pscychresns.2014.02.010\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePigott HE, Kim T, Xu C, Dubin MJ, Quitkin FM, Trivedi MH (2023) What are the treatment remission, response and extent of improvement rates after up to four trials of antidepressant therapies in real-world depressed patients? A reanalysis of the STAR*D study\u0026rsquo;s patient-level data with fidelity to the original research protocol. BMJ Open 13:e063095. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/bmjopen-2022-063095\u003c/span\u003e\u003cspan address=\"10.1136/bmjopen-2022-063095\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePizzagalli DA (2011) Frontocingulate dysfunction in depression: Toward biomarkers of treatment response. Neuropsychopharmacology 36:183\u0026ndash;206. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/npp.2010.166\u003c/span\u003e\u003cspan address=\"10.1038/npp.2010.166\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePizzagalli DA, Webb CA, Dillon DG, Tenke CE, Kayser J, Goer F, Fava M, McGrath PJ, Weissman MM, Bruder GE (2018) Pretreatment rostral anterior cingulate cortex theta activity in relation to symptom improvement in depression: A randomized clinical trial. JAMA Psychiatry 75(6):547\u0026ndash;554. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1001/jamapsychiatry.2018.0252\u003c/span\u003e\u003cspan address=\"10.1001/jamapsychiatry.2018.0252\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eR Core Team (2023) R: A language and environment for statistical computing. R Foundation for Statistical Computing. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.R-project.org/\u003c/span\u003e\u003cspan address=\"https://www.R-project.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchnitzler A, Gross J (2005) Normal and pathological oscillatory communication in the brain. Nat Rev Neurosci 6(4):285\u0026ndash;296. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/nrn1650\u003c/span\u003e\u003cspan address=\"10.1038/nrn1650\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchwartzmann B, Quilty LC, Dhami P, Kolla N, Mulsant BH, Daskalakis ZJ, Blumberger DM, Kennedy SH, Farzan F (2023) Resting-state EEG delta and alpha power predict response to cognitive behavioral therapy in depression: A Canadian Biomarker Integration Network in Depression study. Sci Rep 13:8418. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41598-023-35179-4\u003c/span\u003e\u003cspan address=\"10.1038/s41598-023-35179-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchwartzmann B, Dhami P, Uher R, Farzan F, Kennedy SH (2023) Developing an electroencephalography-based model for predicting response to antidepressant medication. JAMA Netw Open 6(9):e2336094. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1001/jamanetworkopen.2023.36094\u003c/span\u003e\u003cspan address=\"10.1001/jamanetworkopen.2023.36094\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSmith SM, Nichols TE (2009) Threshold-free cluster enhancement: Addressing problems of smoothing, threshold dependence and localisation in cluster inference. NeuroImage 44(1):83\u0026ndash;98. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.neuroimage.2008.03.061\u003c/span\u003e\u003cspan address=\"10.1016/j.neuroimage.2008.03.061\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSporns O, Tononi G, Edelman GM (2000) Connectivity and complexity: The relationship between neuroanatomy and brain dynamics. Neural Netw 13(8\u0026ndash;9):909\u0026ndash;922. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/S0893-6080(00)00053-8\u003c/span\u003e\u003cspan address=\"10.1016/S0893-6080(00)00053-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTei S, Faber PL, Lehmann D, Tsujiuchi T, Kumano H, Pascual-Marqui RD, Kinoshita T (2009) Meditators and non-meditators: EEG source imaging during resting. Brain Topogr 22(3):158\u0026ndash;165. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10548-009-0107-4\u003c/span\u003e\u003cspan address=\"10.1007/s10548-009-0107-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTrivedi MH, McGrath PJ, Fava M, Parsey RV, Kurian BT, Phillips ML, Oquendo MA, Bruder GE, Pizzagalli DA, Toups M, Cooper C, Adams P, Weyandt S, McInnis MG, Kraemer HC (2016) Establishing moderators and biosignatures of antidepressant response in clinical care (EMBARC): Rationale and design. J Psychiatr Res 78:11\u0026ndash;23. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jpsychires.2016.03.001\u003c/span\u003e\u003cspan address=\"10.1016/j.jpsychires.2016.03.001\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTrivedi MH, Rush AJ, Wisniewski SR, Nierenberg AA, Warden D, Ritz L, Norquist G, Howland RH, Lebowitz B, McGrath PJ, Shores-Wilson K, Biggs MM, Balasubramani GK, Fava M (2006) Evaluation of outcomes with citalopram for depression using measurement-based care in STAR*D: Implications for clinical practice. Am J Psychiatry 163(1):28\u0026ndash;40. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1176/appi.ajp.163.1.28\u003c/span\u003e\u003cspan address=\"10.1176/appi.ajp.163.1.28\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUlke C, Kayser J, Tenke CE, Mergl R, Sander C, Panier LYX, Alvarenga JE, Fava M, McGrath PJ, Deldin PJ, McInnis MG, Trivedi MH, Weissman MM, Pizzagalli DA, Hegerl U, Bruder GE (2024) EEG measures of brain arousal in relation to symptom improvement in patients with major depressive disorder: Results from a randomized placebo-controlled clinical trial. Psychiatry Res 342:116165. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.psychres.2024.116165\u003c/span\u003e\u003cspan address=\"10.1016/j.psychres.2024.116165\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evan der Vinne N, Vollebregt MA, van Putten MJAM, Arns M (2017) Frontal alpha asymmetry as a diagnostic marker in depression: Fact or fiction? A meta-analysis. NeuroImage: Clin 16:79\u0026ndash;87. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.nicl.2017.07.006\u003c/span\u003e\u003cspan address=\"10.1016/j.nicl.2017.07.006\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evon Stein A, Sarnthein J (2000) Different frequencies for different scales of cortical integration: From local gamma to long range alpha/theta synchronization. Int J Psychophysiol 38(3):301\u0026ndash;313. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/S0167-8760(00)00172-0\u003c/span\u003e\u003cspan address=\"10.1016/S0167-8760(00)00172-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWatts D, Pulice RF, Reilly J, Deldin PJ, Olbrich S, Tenke CE, Kayser J, Bruder GE, McGrath PJ (2022) Predicting treatment response using EEG in major depressive disorder: A machine-learning meta-analysis. Translational Psychiatry 12:332. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41398-022-02064-z\u003c/span\u003e\u003cspan address=\"10.1038/s41398-022-02064-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWickham H (2016) ggplot2: Elegant graphics for data analysis. Springer-. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/978-3-319-24277-4_9\u003c/span\u003e\u003cspan address=\"10.1007/978-3-319-24277-4_9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu W, Zhang Y, Jiang J, Lucas MV, Fonzo GA, Rolle CE, Cooper C, Chin-Fatt C, Krepel N, Cornelssen CA, Wright R, Toll RT, Harrewijn A, Etkin A (2020) An electroencephalographic signature predicts antidepressant response in major depression. Nat Biotechnol 38(4):439\u0026ndash;447. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41587-019-0397-3\u003c/span\u003e\u003cspan address=\"10.1038/s41587-019-0397-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXiao W, Woodham RD, Cui Y, Antoniades M, Srinivasan D, Fan Y, Erus G, Garcia JA, Arnott SR, Chen T, Choi KS, Chin-Fatt C, Frey BN, Frokjaer VG, Ganz M, Godlewska BR, Hassel S, Ho K, McIntosh AM, Davatzikos C (2025) Neuroanatomical dimensions in major depression: External validation and links with cognition, adverse life events, self-harm, metabolomics and genetics [Preprint]. Res Square. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.21203/rs.3.rs-5975731/v1\u003c/span\u003e\u003cspan address=\"10.21203/rs.3.rs-5975731/v1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYan G, Zhang Y, Wang S, Chen Q, Zhu L, Fang Y, Tang W (2024) Global, regional, and national temporal trend in burden of major depressive disorder from 1990 to 2019: An analysis of the global burden of disease study. Psychiatry Res 337:115958. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.psychres.2024.115958\u003c/span\u003e\u003cspan address=\"10.1016/j.psychres.2024.115958\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhdanov A, Atluri S, Wong W, Levinson AJ, Blumberger DM, Daskalakis ZJ, Farzan F (2020) Use of machine learning for predicting escitalopram treatment outcome from electroencephalography recordings in adult patients with depression. JAMA Netw Open 3(1):e1918377. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1001/jamanetworkopen.2019.18377\u003c/span\u003e\u003cspan address=\"10.1001/jamanetworkopen.2019.18377\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[{"identity":"153107c4-dc2d-456e-9fdc-9102345814a3","identifier":"10.13039/100000025","name":"National Institute of Mental Health","awardNumber":"r01mh134236","order_by":0}],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"University of East London","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Major depressive disorder, EEG, neurophysiology, neuroanatomical dimensions, treatment response, resting-state EEG, spectral power, frontal alpha asymmetry, multiscale entropy","lastPublishedDoi":"10.21203/rs.3.rs-6658719/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6658719/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eMajor depressive disorder (MDD) is a heterogeneous disorder with variable treatment responses and no established biomarkers for identification or predictors of treatment response. In the COORDINATE-MDD consortium, a data-driven classification identified two neuroanatomic-based dimensions: Dimension 1 (D1), with preserved grey and white matter volumes, and Dimension 2 (D2), with widespread reductions. Here, we investigated whether resting-state electroencephalography (EEG) features differ between these dimensions and whether such features predict treatment response.\u003c/p\u003e \u003cp\u003eParticipants were 237 MDD (155 women; mean age 37.47\u0026thinsp;\u0026plusmn;\u0026thinsp;13.36 year) from two clinical trials: CAN-BIND (escitalopram) and EMBARC (randomized to sertraline or placebo). All were medication-free at baseline, in a current depressive episode of at least moderate severity. Resting-state, eyes-closed EEG was recorded at baseline. EEG features included spectral power, frontal alpha asymmetry (FAA), multiscale sample entropy (MSE), and inter-site phase clustering (ISPC). Analyses examined effects of neuroanatomical dimension (D1, D2) and clinical outcome (responder, non-responder), using ANCOVA and threshold-free cluster enhancement (TFCE).\u003c/p\u003e \u003cp\u003eD1 participants showed greater absolute and log-transformed theta, alpha, and beta power, and lower relative delta power compared to D2, particularly in frontal and central regions. Among responders, D1 showed higher alpha and theta power and greater MSE at coarse time scales.\u003c/p\u003e \u003cp\u003eIndividuals with preserved neuroanatomy in D1 exhibit electrophysiological markers of more flexible and integrated brain function, possibly reflecting efficient top-down regulation and adaptive neural dynamics. In contrast, the D2 dimension, marked by lower complexity and elevated delta activity, may reflect disrupted network integrity, reduced cortical arousal, and impaired information processing.\u003c/p\u003e","manuscriptTitle":"Neurophysiological correlates of neuroanatomical dimensions in major depressive disorder: EEG markers of brain function and treatment outcome","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-05-16 12:09:41","doi":"10.21203/rs.3.rs-6658719/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"7e4c5448-4825-4b96-88e0-caef756c2d61","owner":[],"postedDate":"May 16th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":48490358,"name":"Psychiatry"}],"tags":[],"updatedAt":"2025-05-16T12:09:41+00:00","versionOfRecord":[],"versionCreatedAt":"2025-05-16 12:09:41","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6658719","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6658719","identity":"rs-6658719","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}