Takamiya et al. Study Hippocampal Dentate Gyrus Plasticity and Its Association With Dimensional Clinical Improvement Across Four Depression Treatment Interventions Akihiro Takamiya et al. American Journal of Psychiatry. Published online August 26, 2026. DOI: 10.1176/appi.ajp.20251330 Research Question Do cognitive-behavioral therapy (CBT), pharmacotherapy, repetitive transcranial magnetic stimulation (TMS), and electroconvulsive therapy (ECT) produce measurable changes in hippocampal dentate gyrus volume, and are those changes associated with improvement in particular depressive symptom dimensions or cognitive function? Brief Background The dentate gyrus is a hippocampal subregion involved in learning, memory, emotional regulation, and multiple forms of neuroplasticity. Preclinical studies have identified increased dentate gyrus neurogenesis, synaptic remodeling, and related plasticity following antidepressant medications and electroconvulsive stimulation. Human MRI studies have reported increases in hippocampal or dentate gyrus volume following pharmacotherapy, repetitive transcranial magnetic stimulation, and ECT, although findings have varied across treatments and studies. It remains unclear whether dentate gyrus plasticity represents a final common pathway shared by effective depression treatments or a biological effect that is more specific to particular interventions. Study Location This was a prospective, multisite observational study conducted at four sites in Japan. MRI data were acquired using six 3-Tesla scanners. Study Sponsor The study was supported by the Japan Agency for Medical Research and Development through grant JP19dm0307102h0001. Inclusion Criteria Patients were eligible if they: Were 20-85 years old Met DSM-IV criteria for major depressive disorder Had a baseline 17-item Hamilton Depression Rating Scale score of at least 8 Were scheduled in routine clinical care to receive cognitive-behavioral therapy, a pharmacologic treatment change, TMS, or ECT In addition to patients, healthy controls were also recruited. Exclusion Criteria Key exclusions were: Bipolar disorder Dementia Current significant suicidal ideation or recurrent suicide attempts considered incompatible with safe participation in the longitudinal study How Many Participants 450 participants: 181 healthy controls and 269 patients with major depressive disorder. Study Intervention This was a nonrandomized observational study. Treatment was selected by the attending physician before enrollment, and all interventions were delivered in routine clinical settings rather than assigned or modified by the study team. Cognitive-behavioral therapy: N = 106; Sixteen weekly sessions Pharmacotherapy: N = 83; A clinically indicated medication initiation, switch, or augmentation after baseline assessment TMS: N = 39; 10-Hz stimulation; 3,000 pulses per day; Five treatment days per week for 4-6 weeks ECT: N = 41; pulse width of 0.5 ms; Treatments administered two or three times per week; Electrode placement and number of sessions were selected by treating physicians; Mean number of ECT sessions was 10.5, with a range of 4-19 Follow-Up MRI and clinical assessments were conducted at three time points: TP1: Within two weeks before treatment TP2: Within two weeks after the treatment course TP3: Six months after treatment Endpoints Primary imaging outcome Bilateral CA4/dentate gyrus volume measured using the longitudinal FreeSurfer hippocampal subfield segmentation pipeline Primary clinical association Relationship between CA4/dentate gyrus volume and the HAM-D-17 core depressive symptom dimension, comprising psychomotor retardation, depressed mood, suicidality, work and interests, and guilt Results Baseline Findings Patients with major depressive disorder had lower CA4/dentate gyrus volume than healthy controls. Within the MDD group, baseline CA4/dentate gyrus volume was not associated with baseline core depressive symptom severity. The ECT group differed substantially from the other treatment groups at baseline. Participants receiving ECT were older, had more severe depressive symptoms, had a later age at illness onset, and were less likely to be experiencing a first depressive episode. Treatment-Associated CA4/Dentate Gyrus Changes Across all patients, CA4/dentate gyrus volume increased significantly over time. Post hoc analysis showed that only the ECT group experienced a statistically significant increase in CA4/dentate gyrus volume. No significant CA4/dentate gyrus volume increase was detected following CBT, pharmacotherapy, or TMS. At baseline, CA4/dentate gyrus volume in the ECT group was lower than in healthy controls. Following ECT, the difference was smaller and was no longer statistically significant: Relationship to Depressive Symptoms Changes in CA4/dentate gyrus volume were associated with changes in core depressive symptoms. In post hoc treatment-specific analyses, this relationship was found only in the ECT group. Volume changes were not significantly associated with the anxiety, somatic, or insomnia dimensions after correction for multiple comparisons. Six-Month Findings Across the three time points, CA4/dentate gyrus volume followed a nonlinear trajectory. This pattern was again driven by the ECT group: Significant increase from baseline to immediately after ECT Significant decrease from the post-treatment assessment to six months No significant longitudinal changes in the CBT, medication, or TMS groups Changes in CA4/dentate gyrus volume across the three assessments remained associated with changes in core depressive symptoms. Limitations CA4 and the dentate gyrus cannot be completely separated using the FreeSurfer atlas employed in this study. The reported region should therefore be understood as a combined CA4/dentate gyrus MRI segmentation rather than a purely dentate gyrus measurement. Treatment assignment was not randomized. Patients received interventions selected through routine clinical decision-making, creating a substantial risk of confounding by indication. The treatment groups differed at baseline. The ECT group was older, had more severe depression, and had lower baseline CA4/dentate gyrus volume than several other groups. Nearly all ECT participants received bilateral treatment. The findings may not generalize to a predominantly right-unilateral ECT course or to specific pulse widths and dosing strategies. The association between volume change and core depressive symptom improvement was strong in the ECT mixed-effects analysis but modest in the complementary change-score analysis. This difference underscores that the estimated relationship depends partly on the statistical model. Summary and Implications This large, multisite longitudinal study found a robust increase in MRI-measured CA4/dentate gyrus volume following ECT but not following CBT, pharmacotherapy, or TMS. The ECT-associated volume increase tracked improvement in core depressive symptoms and was not associated with changes in measured cognitive performance. At six months, the initial enlargement had partially receded, producing an increase-decrease trajectory that paralleled changes in core depressive symptoms. The results argue against MRI-detectable dentate gyrus enlargement as a universal structural response to depression treatment. Instead, this imaging phenotype appears particularly associated with ECT under the conditions studied. The clinical correlation is noteworthy because it was specific to the core depressive symptom dimension rather than anxiety, somatic symptoms, insomnia, or cognitive outcomes. Comments This study addresses an important mechanistic question by directly comparing longitudinal hippocampal subfield measurements across four commonly used depression treatments. Its principal strength is not simply the relatively large sample, but the combination of repeated imaging, multisite harmonization, dimensional symptom analyses, and six-month follow-up. The findings suggest that ECT produces a particularly large biological perturbation within the hippocampus. This may relate to the combination of electrical stimulation, generalized seizure activity, repeated treatment exposure, neurotrophic signaling, synaptic remodeling, glial responses, and changes in local tissue water or vasculature. Structural MRI alone cannot determine which of these processes is responsible. An important analytic choice was the emphasis on core depressive symptoms rather than reliance exclusively on the total HAM-D-17 score. Depression rating scales combine symptoms with potentially different biological substrates, and a treatment-related brain change may track one domain while having little relationship to others. The specificity of the observed association supports dimensional clinical analyses in future ECT biomarker studies. At the same time, the nonrandomized design substantially limits causal interpretation. The ECT participants were older, more severely depressed, and had lower baseline CA4/dentate gyrus volume. These differences are inherent to naturalistic treatment selection and make it difficult to determine whether the observed specificity reflects ECT itself, the characteristics of patients selected for ECT, the intensity of treatment exposure, or an interaction among these factors. Ultimately, the study provides compelling evidence that ECT is associated with a distinctive, clinically relevant, and partly reversible structural MRI signal in the CA4/dentate gyrus. It does not yet establish what microscopic process produces that signal or whether the structural change is causal, compensatory, or simply a marker of treatment exposure. AI Use Disclaimer Generative artificial intelligence was used as a supplemental editorial tool to assist with organizing and summarizing the source article in the established ISEN blog format. All study details, numerical results, and statistical estimates were restricted to information reported in the cited article. The final content has been reviewed and edited by the author, who retains responsibility for its accuracy and interpretation.

Matheson et al. Study Convergent increases in serotonin 1B receptor binding following ketamine and electroconvulsive therapy: a multi-centrecenter bayesian re-analysis of PET data Granville J. Matheson et al. Molecular Psychiatry. 2026. https://doi.org/10.1038/s41380-026-03704-2 Research Question Do ketamine and electroconvulsive therapy (ECT) produce convergent changes in serotonin 1B receptor (5-HT1BR) binding, and can these changes be distinguished from placebo? Brief Background The serotonin 1B receptor (5-HT1BR) is an inhibitory receptor located both presynaptically on serotonin terminals as an autoreceptor and on non-serotonergic neurons as a heteroreceptor. It helps regulate serotonin release across cortical and limbic regions Prior PET studies have suggested reduced 5-HT1BR binding in major depressive disorder, particularly within the hippocampus, anterior cingulate cortex, and ventral striatal–ventral pallidal reward circuitry; however, these studies generally involved small samples. Ketamine and ECT have different primary mechanisms but may converge on serotonergic signaling. Preclinical findings indicate that both treatments can increase extracellular serotonin, potentially producing delayed, homeostatic changes in 5-HT1BR expression or activity. [¹¹C]AZ10419369 PET estimates 5-HT1BR binding potential, not serotonin concentration. Increased binding may reflect greater receptor density, altered receptor affinity or conformational state, reduced competition from endogenous serotonin, or a combination of these processes. Study Location Multi-center re-analysis of PET data collected at three centers: • Karolinska Institutet (KI), Stockholm, Sweden • Nippon Medical School (NMS), Tokyo, Japan • Neurobiology Research Unit at Rigshospitalet (NRU), Copenhagen, Denmark Study Sponsor This work was supported by Hjärnfonden (PS2020-0016), Vetenskapsrådet (2020-06356, 2024-03534), Region Stockholm (clinical research appointment), NIH grants P50MH090964 and R01EB024526, and OpenNeuroPET, funded through BRAIN Initiative grant 1R24MH120004-01A1 and Novo Nordisk Foundation grant NN20OC0063277. Inclusion Criteria The combined dataset included healthy volunteers and patients with major depressive disorder (MDD). All patients met DSM-IV or DSM-5 criteria for a major depressive episode. Within-subject treatment samples: Ketamine/placebo cohort: outpatients with SSRI-treatment-resistant depression, defined as failure to respond to at least 4 weeks of SSRI treatment at adequate doses during the current depressive episode. Patients were psychotropic medication-free at PET; those receiving antidepressants at inclusion completed a supervised washout of at least five drug half-lives before baseline PET. ECT cohort: hospitalised patients with severe major depressive episodes warranting ECT. Patients were receiving concurrent medication in steady state at the time of PET, with no medication changes between PET measurements; none of the medications acted on 5-HT1BR. No formal treatment-resistance criterion was applied. Exclusion Criteria Key exclusion criteria common to both treatment studies were: Psychotic disorder Substance abuse Organic brain disorder Pregnancy How Many Participants Treatment completers with usable pre- and post-treatment PET data: Ketamine: 19 Saline placebo: 10 ECT: 13 Study Intervention This study re-analysed previously collected data; it did not administer new treatments. Original ketamine/placebo study: Randomised, double-blind allocation to a single intravenous infusion of racemic ketamine 0.5 mg/kg or saline PET performed 1-3 days before and 24-72 hours after infusion Original ECT study: Open-label ECT series of 7-16 treatments, administered 2-3 times per week PET performed before treatment and up to one week after the ECT series No sham-ECT control was included The re-analysis applied SiMBA with the simplified reference tissue model to quantify and analyse the PET data simultaneously, borrow information across individuals and regions, and harmonise measurements across centers. Follow-Up The analysis used the original studies' pre- and post-treatment PET measurements. Post-treatment PET occurred 24-72 hours after the ketamine or saline infusion and up to one week after completion of the ECT series. No longer-term clinical or PET follow-up was reported in this re-analysis. Endpoints Primary analytical focus: Global within-individual change in 5-HT1BR binding after ketamine, saline placebo, and ECT Whether ketamine- and ECT-associated changes could be statistically distinguished from placebo and from one another Secondary aims: Regional specificity of treatment effects Differences in 5-HT1BR binding between patients with MDD before treatment and healthy controls Association between change in 5-HT1BR binding and change in depressive symptoms Effect of adding data from other studies and center on precision of treatment-effect estimates Results Global treatment-associated changes in 5-HT1BR binding: Ketamine: +6.4% (95% credible interval [CI], 3.1% to 9.6%; posterior probability of direction [pd] > 0.999) ECT: +9.3% (95% CI, 4.3% to 14.2%; pd > 0.999) Placebo: -0.7% (95% CI, -4.6% to 3.1%; pd = 0.632) ECT-associated changes were slightly larger than ketamine-associated changes (pd = 0.973) Difference-in-difference comparisons with placebo: Ketamine versus placebo: +7.1% (95% CI, 2.2% to 12.3%; pd = 0.996) ECT versus placebo: +10.0% (95% CI, 4.0% to 16.3%; pd > 0.999) The ECT-placebo comparison used the saline-placebo arm of the ketamine study because the ECT study did not include sham ECT. It therefore assumed that placebo-associated changes would be similar for sham ketamine and sham ECT; center and treatment effects were also partially confounded. Regional and clinical findings: No region was clearly statistically differentiable from the mean treatment effect for ECT or ketamine. For ECT, the hippocampus showed the largest proportional change relative to the mean ECT effect (+6.2%; pd = 0.937), while the frontal cortex showed the smallest (-3.9%; pd = 0.907). Estimated regional heterogeneity was greater for ECT than ketamine: SD of regional deviations in change in log BPND was 0.044 for ECT and 0.013 for ketamine; pd = 0.90 that ECT produced more regionally heterogeneous effects. Changes in binding were not associated with individual symptom improvement beyond the average change in symptom scores. MDD versus healthy volunteers: No evidence of a global difference: -1.9% (95% CI, -7.3% to 4.1%; pd = 0.754). Age-by-MDD interaction: 3.0% greater reduction per decade (95% CI, -5.8% to 0.0%; pd = 0.972), suggesting lower binding among older patients with MDD relative to older controls. Exploratory regional analysis suggested lower amygdala binding in MDD: -6.3% (95% CI, -11.7% to -0.3%; pd = 0.985). The authors advised cautious interpretation. Other Relevant Information 5-HT1BR functions as an inhibitory auto- and heteroreceptor; presynaptic autoreceptors are located on axon terminals in serotonin projection regions and regulate extracellular serotonin. [11C]AZ10419369 PET cannot distinguish binding to 5-HT1BR autoreceptors from heteroreceptors, or determine whether binding changes reflect receptor density, receptor affinity state, or competition from endogenous serotonin. Because acute serotonin release would be expected to decrease rather than increase binding potential, and PET was performed days after treatment, the authors considered receptor-level changes more plausible than sustained increases in extracellular serotonin. The authors proposed that increased binding could reflect receptor upregulation or a greater proportion of receptors in an active conformation, potentially as a homeostatic response to treatment-induced serotonin release. Limitations This was a re-analysis of previously collected data, and the biological conclusions depend on the assumptions of the SiMBA model; independent replication with new data is needed. Sample sizes were determined by available data from the original studies; no prospective power calculation was performed for this re-analysis. The only experimental placebo group was the saline arm of the ketamine trial. The ECT-placebo comparison was cross-center and assumed that sham-ECT and sham-ketamine placebo effects on 5-HT1BR binding would be similar. ECT, ketamine and placebo were administered at separate locationsat KI, so center and treatment effects were partially confounded despite within-individual estimates and model-based harmonisation. The ketamine study was randomised and double-blind, whereas the ECT study was open-label and had no sham-ECT control. The treatment cohorts differed clinically: ketamine/placebo participants were medication-free outpatients with SSRI-resistant depression, while ECT participants were medicated inpatients with severe depressive episodes. The ketamine study included results following a single intravenous infusion of racemic ketamine compared to an acute, multi-treatment course of ECT. Current standard of care for ketamine infusion therapy similarly includes and acute, multi-treatment typically consisting of 6 to 8, twice weekly consecutive treatments. The results presented here may therefore under-represent biological changes to ketamine in a more clinically relevant treatment context. PET cannot determine the cellular source or precise biological meaning of increased [11C]AZ10419369 binding. Binding changes were not associated with individual symptom improvement, leaving their relationship to therapeutic efficacy uncertain. Summary and Implications This study showed global increases in 5-HT1BR binding after exposure to both ketamine and ECT. The estimated increases were distinguishable from saline placebo, although the ECT comparison was exploratory because it relied on the placebo arm of a different study conducted at another center. The magnitude of binding change was not associated with individual symptom improvement, however, the sample sizes in the treatment conditions were small. The convergent findings support a possible shared downstream serotonergic response to two rapid-acting treatments with different primary mechanisms. The authors interpret the increased binding as more consistent with receptor-level adaptation than persistent elevation of extracellular serotonin, but PET cannot establish whether the signal reflects autoreceptors or heteroreceptors, receptor number or density, affinity state, or endogenous neurotransmitter competition. The results also do not preclude changes to subcellular neuronal morphology, such synaptic architecture, which these treatments are known to influence or changes to the glial contribution of the tripartite synapse. The clinical relevance of the finding therefore remains unresolved and requires further mechanistic study and independent replication. Comments Ketamine and ECT have markedly different proximal mechanisms of action. Ketamine is generally understood to initiate its antidepressant effects through NMDA-receptor antagonism, altered glutamate release, and increased AMPA-receptor signaling, followed by activation of neuroplasticity-related pathways such as BDNF–TrkB and mTOR. ECT begins with an electrically induced generalized tonic-clonic seizure that produces widespread changes in neuronal activity, network connectivity, neurotransmitter systems, and neurotrophic signaling. No single mechanism fully explains the clinical effects of either treatment.

Confirmatory efficacy and safety trial of magnetic seizure therapy versus right unilateral ultra-brief electroconvulsive therapy in depression (CREST–MST): a randomised, double-blind, non-inferiority trial in Canada and the USA