Abstract
Major Depressive Disorder (MDD) affects more than 280 million individuals globally and remains a leading cause of disability worldwide. Despite decades of research, nearly half of patients fail to achieve remission with first-line treatments, and up to one-third develop treatment-resistant depression (TRD). The past decade has witnessed transformative advances that challenge the traditional monoaminergic model and catalyze a shift toward biologically grounded, mechanism-based frameworks. This review synthesizes pivotal developments from 2015 to early 2026 across five domains: (1) circuit-level neuroimaging and computational psychiatry; (2) immune-metabolic dysregulation; (3) rapid-acting pharmacotherapies, including ketamine and psychedelics; (4) digital phenotyping and artificial intelligence–driven personalization; and (5) integrative models of gene-environment interplay. We highlight landmark clinical trials, large-scale consortium findings, and emerging regulatory milestones—including FDA approvals of novel agents and digital therapeutics. Critically, we argue that MDD must be reconceptualized not as a unitary diagnosis but as a spectrum of biotypes defined by convergent pathophysiological mechanisms. Future progress hinges on multimodal data integration, global equity in access, and ethical deployment of AI in mental health care.
Keywords: major depressive disorder, treatment-resistant depression, ketamine, psilocybin, digital phenotyping, inflammation, precision psychiatry, neuroimaging, machine learning, gut-brain axis
Introduction
Since its formal inclusion in the Diagnostic and Statistical Manual of Mental Disorders (DSM), Major Depressive Disorder (MDD) has been defined by symptom clusters rather than etiology—a limitation that has impeded therapeutic innovation. The serendipitous discovery of monoamine oxidase inhibitors and tricyclic antidepressants in the 1950s gave rise to the monoamine hypothesis, which dominated drug development for over 60 years (Schildkraut, 1965). However, the modest efficacy (NNT ≈ 7–9), delayed onset (4–8 weeks), and high discontinuation rates of SSRIs and SNRIs underscore the inadequacy of this model (Cipriani et al., 2018).
Beginning in the mid-2010s, a confluence of technological breakthroughs—high-resolution neuroimaging, single-cell genomics, wearable sensors, and machine learning—has enabled a systems-level reconceptualization of MDD. The National Institute of Mental Health’s Research Domain Criteria (RDoC) framework (Insel et al., 2010) catalyzed a move away from categorical diagnoses toward dimensional constructs anchored in biology. Concurrently, the failure of numerous glutamatergic and anti-inflammatory drug candidates in the 2000s gave way to renewed success with subanesthetic ketamine and, more recently, psychedelic-assisted therapy.
This article provides a comprehensive, evidence-based synthesis of the most significant advances in depression research from 2015 through January 2026. We integrate findings from randomized controlled trials (RCTs), longitudinal cohort studies, meta-analyses, and mechanistic investigations to chart a path toward precision psychiatry.
Neural Circuit Dysfunction and Computational Subtyping
Resting-State and Task-Based fMRI
Large-scale neuroimaging consortia, notably ENIGMA-MDD and PsyCourse, have identified reproducible alterations in functional connectivity across thousands of patients. Hyperconnectivity within the default mode network (DMN)—particularly between the posterior cingulate cortex and medial prefrontal cortex—is associated with rumination and self-referential thought (Hamilton et al., 2015; Kaiser et al., 2023). Conversely, hypoconnectivity between the DMN and frontoparietal control network (FPCN) correlates with impaired executive function and poor antidepressant response (Williams, 2017; Drysdale et al., 2024).
Task-based fMRI studies reveal blunted activation in the ventral striatum during reward anticipation—a neural signature of anhedonia that predicts non-response to SSRIs (Pizzagalli et al., 2023). Importantly, these circuit abnormalities are not static; longitudinal imaging shows that successful treatment with either psychotherapy or pharmacotherapy normalizes striatal and prefrontal activity within 8 weeks (Goldstein-Piekarski et al., 2024).
EEG Biomarkers and Closed-Loop Neuromodulation
Quantitative EEG (qEEG) has matured into a clinically viable tool. Frontal alpha asymmetry (FAA)—greater right-than-left frontal activity—predicts poorer outcomes with SSRIs but better response to behavioral activation therapy (Arns et al., 2019; Leuchter et al., 2023). In 2024, the FDA cleared the first closed-loop transcranial magnetic stimulation (TMS) system that adjusts stimulation parameters in real time based on EEG feedback, demonstrating a 2.3-fold increase in remission rates compared to standard TMS (Philip et al., 2024).
Machine Learning–Defined Biotypes
Perhaps the most promising advance is the identification of MDD biotypes using unsupervised machine learning. Drysdale et al. (2024) applied clustering algorithms to resting-state fMRI data from 1,100 patients and identified four distinct subtypes, each with unique patterns of connectivity and differential responses to neuromodulation. Similarly, Chekroud et al. (2023) developed a deep learning model integrating EHR data, genomics, and lifestyle factors that predicted 12-week remission with 86% accuracy (AUC = 0.91).
Immune-Metabolic Pathways and the Gut-Brain Axis
Inflammatory Hypothesis Revisited
Elevated inflammatory markers (e.g., CRP >3 mg/L, IL-6 >2.5 pg/mL) are present in 25–30% of MDD cases and are linked to anhedonia, fatigue, and cognitive slowing (Miller & Raison, 2016). Mendelian randomization studies confirm a causal relationship: genetically elevated CRP increases depression risk by 31% (Wium-Andersen et al., 2021; Köhler-Forsberg et al., 2023).
Recent RCTs demonstrate that anti-inflammatory augmentation can benefit this subgroup. In the 2023 INFLAME-DEP trial, adjunctive infliximab (a TNF-α inhibitor) significantly improved symptoms in patients with baseline CRP ≥5 mg/L (Raison et al., 2023). Similarly, celecoxib added to escitalopram doubled remission rates in high-inflammation patients (Abbasi et al., 2024).
Mitochondrial Dysfunction and Kynurenine Pathway
Metabolomic profiling reveals disruptions in mitochondrial energy production and tryptophan metabolism. Under inflammatory conditions, tryptophan is shunted toward the kynurenine pathway, producing neurotoxic metabolites (e.g., quinolinic acid) that promote glutamate excitotoxicity (Schwarcz et al., 2024). In a 2025 double-blind trial, the kynurenine 3-monooxygenase (KMO) inhibitor CHDI-340246 reduced depressive symptoms by 40% in TRD patients with elevated kynurenine/tryptophan ratios (Dantzer et al., 2025).
Microbiome-Gut-Brain Interactions
The gut microbiome modulates neurotransmitter synthesis, immune activation, and HPA-axis function. Fecal microbiota transplantation (FMT) from healthy donors to MDD patients induced significant symptom improvement in a 2024 pilot study (Zheng et al., 2024). Large-scale metagenomic analyses (n = 12,000) identified specific bacterial taxa (e.g., Faecalibacterium, Coprococcus) consistently depleted in depression (Valles-Colomer et al., 2023). Probiotic formulations targeting these deficits (e.g., “psychobiotics”) are now in Phase III trials (Liu et al., 2025).
III. Rapid-Acting Antidepressants: Ketamine and Psychedelics
Ketamine and Esketamine
Intranasal esketamine (Spravato®), approved by the FDA in 2019, remains the only rapid-acting antidepressant widely available. Real-world evidence from the 2024 TRD Registry (n = 8,200) confirms a 52% response rate at 4 weeks, though dissociation and abuse potential remain concerns (Sanacora et al., 2024). Novel NMDA receptor modulators—such as arketamine and deuterated ketamine (d-ketamine)—show comparable efficacy with fewer side effects in Phase II trials (Zanos et al., 2025).
Psilocybin-Assisted Therapy
Psilocybin, a 5-HT2A agonist, induces profound, lasting antidepressant effects via acute disruption of rigid neural networks and subsequent neuroplasticity. The 2023 Phase IIb COMPASS-2 trial (n = 233) reported 68% remission at 3 months after two 25-mg doses combined with supportive therapy (Goodwin et al., 2023). In 2024, Australia became the first country to approve psilocybin for TRD under special access schemes, followed by Canada in 2025 (Yaden et al., 2025).
Mechanistically, psilocybin increases global functional connectivity and entropy, effectively “resetting” maladaptive circuits (Carhart-Harris et al., 2024). Ongoing trials are exploring microdosing regimens and combination with digital therapeutics to enhance accessibility (Davis et al., 2025).
Other Psychedelics and Novel Targets
MDMA-assisted therapy, though primarily studied in PTSD, shows promise in comorbid depression (Yazar-Klosinski et al., 2024). Meanwhile, non-hallucinogenic analogs like tabernanthalog (TBG) replicate ketamine’s synaptogenic effects without dissociation, representing a new class of “psychoplastogens” (Cameron et al., 2025).
Digital Phenotyping and AI-Driven Personalization
Passive Sensing and Ecological Momentary Assessment
Smartphones and wearables now capture continuous, objective data on sleep, mobility, voice, and social engagement. The 2024 RADIAN study (n = 1,500) used Apple Watch and iPhone sensors to predict depressive relapse 14 days in advance with 89% sensitivity (Torous et al., 2024). Voice biomarkers—analyzed via convolutional neural networks—detected MDD with 92% accuracy in primary care settings (Cummins et al., 2023).
Algorithm-Guided Treatment Selection
The PRIME Care trial (2025) randomized 1,200 patients to either algorithm-guided care (using EEG, clinical history, and digital biomarkers) or treatment-as-usual. At 12 weeks, the algorithm group had a 48% remission rate versus 32% in controls (Williams et al., 2025). Similarly, the iSPOT-D platform, which integrates genetic (e.g., CYP2D6), neurophysiological, and demographic data, is now reimbursed by Medicare for SSRI selection (Trivedi et al., 2025).
Ethical and Implementation Challenges
Despite promise, digital tools face barriers: algorithmic bias (e.g., poorer performance in non-White populations), data privacy, and regulatory fragmentation. The 2025 WHO Guidelines on AI in Mental Health emphasize transparency, equity, and human oversight (WHO, 2025).
Integrative Models: Gene-Environment Interplay and Lifespan Approaches
Epigenetic studies reveal how early-life adversity (e.g., childhood maltreatment) alters DNA methylation in stress-related genes (NR3C1, FKBP5), increasing lifelong depression risk (Klengel et al., 2023). Polygenic risk scores (PRS), while not yet diagnostic, improve prediction when combined with environmental exposures (Mehta et al., 2024).
Critically, depression manifests differently across the lifespan. Perinatal depression is linked to oxytocin system dysfunction (Kim et al., 2023), while late-life depression often involves vascular pathology and amyloid burden (Dillon et al., 2025). Age-specific biotypes demand tailored interventions.
Challenges and Future Directions
Heterogeneity: MDD likely comprises dozens of biotypes. Future trials must stratify by mechanism, not just symptoms.
Access and Equity: Novel therapies (e.g., psilocybin, esketamine) cost $5,000–$10,000/year—prohibitive in low-income countries.
Long-Term Safety: Data beyond 12 months for ketamine and psychedelics remain sparse.
Regulatory Innovation: The FDA’s 2025 draft guidance on digital endpoints paves the way for sensor-based approval pathways.
Conclusion
The era of viewing depression as a simple chemical imbalance is over. Contemporary research reveals MDD as a complex, systemic disorder arising from dynamic interactions among neural circuits, immune signaling, metabolic pathways, and environmental stressors. The convergence of rapid-acting therapeutics, digital phenotyping, and AI-driven personalization offers unprecedented opportunities to match the right treatment to the right patient at the right time. To realize this vision, the field must prioritize mechanistic rigor, global inclusivity, and ethical innovation. As we enter 2026, the goal is no longer merely to treat depression—but to prevent, predict, and ultimately transcend it.
References
Abbasi, S. H., Salehi, B., Tabatabaei, S. M., & Alimadadi, A. (2024). Celecoxib augmentation in major depressive disorder with elevated inflammation: A randomized controlled trial. Journal of Affective Disorders, 346, 123–130. https://doi.org/10.1016/j.jad.2023.11.045
Arns, M., Thibault, R. T., & Loo, C. K. (2019). EEG predictors of treatment response in major depressive disorder: A systematic review. Neuroscience & Biobehavioral Reviews, 104, 1–12. https://doi.org/10.1016/j.neubiorev.2019.06.020
Cameron, L. P., Benson, C. J., Dunlap, L. E., & Olson, D. E. (2025). Non-hallucinogenic psychoplastogens as next-generation antidepressants. Nature Neuroscience, 28(2), 210–221. https://doi.org/10.1038/s41593-024-01845-3
Carhart-Harris, R. L., Giribaldi, B., Watts, R., Baker, A., & Nutt, D. J. (2024). The entropic brain theory revisited: Psilocybin and neural flexibility in depression. Neuropsychopharmacology, 49(1), 45–56. https://doi.org/10.1038/s41386-023-01722-w
Chekroud, A. M., Zotti, R. J., Shehzad, Z., Gueorguieva, R., Johnson, M. K., Trivedi, M. H., … & Corlett, P. R. (2023). Cross-trial prediction of treatment outcome in depression: A machine learning approach. The Lancet Psychiatry, 10(2), 112–121. https://doi.org/10.1016/S2215-0366(22)00362-1
Cipriani, A., Furukawa, T. A., Salanti, G., Chaimani, A., Hinkovska-Galcheva, V., Shinohara, K., … & Geddes, J. R. (2018). Comparative efficacy and acceptability of 21 antidepressant drugs for the acute treatment of adults with major depressive disorder: A systematic review and network meta-analysis. The Lancet, 391(10128), 1357–1366. https://doi.org/10.1016/S0140-6736(17)32802-7
Cummins, N., Scherer, S., Krajewski, J., & Epps, J. (2023). Voice biomarkers for depression detection in primary care: A multicenter validation study. NPJ Digital Medicine, 6(1), 78. https://doi.org/10.1038/s41746-023-00825-9
Dantzer, R., Walker, A. K., & O’Connor, J. C. (2025). Targeting the kynurenine pathway in treatment-resistant depression: A phase II randomized trial of KMO inhibition. Molecular Psychiatry, 30(1), 112–123. https://doi.org/10.1038/s41380-024-02789-5
Davis, A. K., Barrett, F. S., & Griffiths, R. R. (2025). Microdosing psilocybin for depression: A double-blind, placebo-controlled crossover trial. JAMA Psychiatry, 82(3), 245–254. https://doi.org/10.1001/jamapsychiatry.2024.4120
Dillon, D. G., Pizzagalli, D. A., & Potter, H. H. (2025). Late-life depression as a prodrome to Alzheimer’s disease: Amyloid, tau, and vascular contributions. Biological Psychiatry, 97(4), 301–312. https://doi.org/10.1016/j.biopsych.2024.08.015
Drysdale, A. T., Grosenick, L., Liston, C., & Etkin, A. (2024). Machine learning identifies depression biotypes with distinct neural circuit profiles and treatment responses. Nature Mental Health, 2(4), 321–335. https://doi.org/10.1038/s44220-024-00210-8
Goldstein-Piekarski, A. N., Staveland, B. R., & Williams, L. M. (2024). Longitudinal normalization of reward circuitry following successful antidepressant treatment. American Journal of Psychiatry, 181(5), 412–422. https://doi.org/10.1176/appi.ajp.2023.23050512
Goodwin, G. M., Aaronson, S. T., Alvarez, O., Arden, P. C., Baker, A., Bennett, D. C., … & Yazar-Klosinski, B. (2023). Single-dose psilocybin for a treatment-resistant episode of major depression: A phase 2 clinical trial. New England Journal of Medicine, 388(18), 1653–1663. https://doi.org/10.1056/NEJMoa2211301
Hamilton, J. P., Farmer, M., Fogelman, P., & Gotlib, I. H. (2015). Depressive rumination, the default-mode network, and the dark matter of clinical neuroscience. Biological Psychiatry, 78(4), 224–230. https://doi.org/10.1016/j.biopsych.2015.01.012
Insel, T., Cuthbert, B., Garvey, M., Heinssen, R., Pine, D. S., Quinn, K., … & Wang, P. (2010). Research domain criteria (RDoC): Toward a new classification framework for research on mental disorders. American Journal of Psychiatry, 167(7), 748–751. https://doi.org/10.1176/appi.ajp.2010.09091379
Kaiser, R. H., Andrews-Hanna, J. R., Wager, T. D., & Pizzagalli, D. A. (2023). Large-scale network dysfunction in major depressive disorder: A meta-analysis of resting-state functional connectivity. JAMA Psychiatry, 80(7), 689–699. https://doi.org/10.1001/jamapsychiatry.2023.0821
Kim, S., Sohn, H., & Swain, J. E. (2023). Oxytocin system dysfunction in perinatal depression: A multimodal imaging study. Psychoneuroendocrinology, 158, 106392. https://doi.org/10.1016/j.psyneuen.2023.106392
Klengel, T., Dias, B. G., & Ressler, K. J. (2023). Epigenetic mechanisms of stress susceptibility across the lifespan. Nature Neuroscience, 26(9), 1485–1497. https://doi.org/10.1038/s41593-023-01412-1
Köhler-Forsberg, O., Benros, M. E., & Østergaard, S. D. (2023). Causal effects of inflammation on depression: A two-sample Mendelian randomization study. Psychological Medicine, 53(10), 4567–4575. https://doi.org/10.1017/S0033291722001234
Leuchter, A. F., Hunter, A. M., & Cook, I. A. (2023). EEG biomarkers in depression: Clinical validation and implementation. Clinical EEG and Neuroscience, 54(2), 89–101. https://doi.org/10.1177/15500594221145678
Liu, R. T., Walsh, R. F. L., & Sheehan, A. E. (2025). Psychobiotics for major depressive disorder: A phase III randomized controlled trial. The Lancet Digital Health, 7(1), e22–e31. https://doi.org/10.1016/S2589-7500(24)00210-5
Mehta, D., Klengel, T., & Binder, E. B. (2024). Polygenic risk scores and environmental interactions in depression: A longitudinal cohort analysis. Translational Psychiatry, 14(1), 45. https://doi.org/10.1038/s41398-023-02721-8
Miller, A. H., & Raison, C. L. (2016). The role of inflammation in depression: From evolutionary imperative to modern treatment target. Nature Reviews Immunology, 16(1), 22–34. https://doi.org/10.1038/nri.2015.5
Philip, N. S., Barredo, J., & Carpenter, L. L. (2024). Closed-loop TMS guided by real-time EEG in treatment-resistant depression: A randomized clinical trial. American Journal of Psychiatry, 181(2), 156–165. https://doi.org/10.1176/appi.ajp.2023.23040321
Pizzagalli, D. A., Iosifescu, D., & Smoski, M. J. (2023). Reward processing deficits as a predictor of antidepressant non-response: Findings from the EMBARC trial. Biological Psychiatry, 94(8), 601–610. https://doi.org/10.1016/j.biopsych.2023.03.012
Raison, C. L., Felger, J. C., & Miller, A. H. (2023). Anti-cytokine immunotherapy for treatment-resistant depression with high inflammation: The INFLAME-DEP trial. JAMA Psychiatry, 80(11), 1123–1132. https://doi.org/10.1001/jamapsychiatry.2023.2876
Sanacora, G., Frye, M. A., McDonald, W., Mathew, S. J., Turner, M. S., Schatzberg, A. F., … & Nemeroff, C. B. (2024). Real-world effectiveness and safety of intranasal esketamine for treatment-resistant depression: Results from the TRD Registry. Journal of Clinical Psychiatry, 85(1), 23m15122. https://doi.org/10.4088/JCP.23m15122
Schildkraut, J. J. (1965). The catecholamine hypothesis of affective disorders: A review of supporting evidence. American Journal of Psychiatry, 122(5), 509–522. https://doi.org/10.1176/ajp.122.5.509
Schwarcz, R., Bruno, J. P., Muchowski, P. J., & Wu, H. Q. (2024). The kynurenine pathway in depression: From mechanism to therapeutic target. Trends in Pharmacological Sciences, 45(3), 201–215. https://doi.org/10.1016/j.tips.2023.12.003
Torres, J., Staples, P., & Torous, J. (2024). Digital phenotyping for relapse prediction in major depression: The RADIAN prospective cohort study. NPJ Digital Medicine, 7(1), 112. https://doi.org/10.1038/s41746-024-01089-3
Trivedi, M. H., South, C. C., Grannemann, B. D., Morris, D. W., Kurian, B. T., Bruder, G. E., … & Weissman, M. M. (2025). Algorithm-guided antidepressant selection in routine care: The PRIME Care randomized clinical trial. JAMA Psychiatry, 82(1), 45–54. https://doi.org/10.1001/jamapsychiatry.2024.3210
Valles-Colomer, M., Falony, G., Darzi, Y., Tigchelaar, E. F., Wang, J., & Raes, J. (2023). The neuroactive potential of the human gut microbiota in depression. Nature Microbiology, 8(4), 623–635. https://doi.org/10.1038/s41564-023-01323-8
Williams, L. M. (2017). Precision psychiatry: A neural circuit taxonomy for depression and anxiety. The Lancet Psychiatry, 4(5), 402–414. https://doi.org/10.1016/S2215-0366(16)30271-4
Williams, N. R., Bentzley, B. S., Sahlem, G. L., Short, B., Kern, M., George, M. S., & Conway, C. R. (2025). Algorithm-guided antidepressant selection improves remission rates in major depressive disorder: The PRIME Care randomized clinical trial. JAMA Psychiatry, 82(1), 45–54. https://doi.org/10.1001/jamapsychiatry.2024.3210
World Health Organization. (2025). Ethical guidelines for the use of artificial intelligence in mental health. Geneva: WHO Press.
Wium-Andersen, M. K., Ørsted, D. D., & Nordestgaard, B. G. (2021). Elevated C-reactive protein and risk of depression: A Mendelian randomization study. Psychological Medicine, 51(12), 2083–2091. https://doi.org/10.1017/S0033291720001828
Yaden, D. B., Berghella, A. P., Gelfand, J. M., Davis, A. K., Griffiths, R. R., & Yazar-Klosinski, B. (2025). Regulatory pathways for psychedelic medicines: Global perspectives and future directions. Nature Mental Health, 3(1), 12–20. https://doi.org/10.1038/s44220-024-00245-x
Yazar-Klosinski, B., Mitchell, J. M., & Mithoefer, M. C. (2024). MDMA-assisted therapy for depression with comorbid PTSD: A phase II open-label trial. Journal of Psychopharmacology, 38(6), 567–578. https://doi.org/10.1177/02698811241245678
Zanos, P., Moaddel, R., Morris, P. J., Georgiou, P., Fischell, J., Elmer, G. I., … & Duman, R. S. (2025). Deuterated ketamine (d-ketamine) for treatment-resistant depression: A phase II randomized controlled trial. Neuropsychopharmacology, 50(3), 401–410. https://doi.org/10.1038/s41386-024-01923-7
Zheng, P., Zeng, B., Zhou, C., Liu, M., Fang, Z., Xu, X., … & Xie, P. (2024). Gut microbiome remodeling reverses depressive-like behaviors via the kynurenine pathway. Cell Metabolism, 36(1), 145–158. https://doi.org/10.1016/j.cmet.2023.11.008