Many Targets, One Disorder: Convergence and Specificity of Network Normalization Across Neuromodulation Modalities in Depression
DOI: Pending DOI
Synopsis
Observational models map depression onto static network dysfunctions, but how interventions normalize these circuits is poorly established. This review examines network changes across eight modalities: ketamine, rTMS, ECT, ezogabine, MBCT, neurofeedback, tVNS, and antidepressants. One convergent track emerged, where ketamine and ezogabine both targeted frontostriatal pathways to treat anhedonia. Conversely, seven dimension-specific profiles were found: neurofeedback, ECT, and citalopram modulated default mode network nodes to halt rumination, mindfulness altered salience pathways to boost interoception; and tVNS modified amygdala-prefrontal circuits to reduce anxiety. Broad diagnoses hide these distinct neural patterns, which can instead be used to guide targeted interventions.
Background
Observational neuroimaging has done a remarkable job linking the heterogeneous clinical presentation of depression to large-scale network abnormalities (1). Abnormalities in the canonical networks, specifically the Default mode network (DMN), Salience network (SN), Frontoparietal network (FPN) and Cingulo-opercular network (CO) serve as baseline markers of distinct symptom dimensions, namely anhedonia, rumination, anxiety, cognitive fatigue and cognitive appraisal (2). However, whether active clinical interventions rescue these systems by funneling through a single convergent route or via highly tailored dimension-specific pathways remains undetermined.
This review evaluates both the network-level convergence and mechanistic specificity of eight treatment modalities, including ketamine, transcranial magnetic stimulation (TMS), electroconvulsive therapy (ECT), ezogabine, mindfulness-based cognitive therapy (MBCT), real-time fMRI neurofeedback, transcutaneous vagus nerve stimulation (tVNS) and traditional antidepressants. If different modalities normalize the same circuit to reduce the same symptom, it proves that the circuit is a robust target for intervention. But if these modalities exhibit distinct symptom-circuit specificity, we can use those unique profiles to directly inform personalized treatment selection.
Methods
To identify relevant literature, a systematic search of PubMed was conducted in March 2026 using keywords for depressive disorder, large-scale cortical networks, functional connectivity methodology, and distinct symptom dimensions. A total of 751 papers were then screened by 2 independent reviewers based on the following inclusion criteria: 1) is an adult sample, 2) has a confirmed depressive episode (major depressive disorder, treatment-resistant depression and bipolar depression in depressive phase), 3) network-level connectivity methodology (resting state fMRI, task-based fMRI, EEG/MEG source based connectivity, or fNIRS) was used, 4) evaluated large scale networks or their canonical nodes (DMN, FPN, SN and CO or PCC, vmPFC, dlPFC, anterior insula, dACC or frontal operculum) and 5) a symptom dimension (rumination, anhedonia, interoception, anxiety, cognitive fatigue, cognitive reappraisal) as a correlate of connectivity change was studied. This yielded 145 papers with 36 conflicts, which were resolved by an independent reviewer.
Full text review produced a total of 58 papers that included eight treatment modalities. We assessed cross-modality convergence by identifying where independent lines of research pointed to the same circuit-symptom link, even when using vastly different clinical intervention or neuroimaging lenses. Symptom specificity was determined by analyzing internal controls, with greater emphasis placed on studies reporting a significant correlation for one symptom alongside a non-significant result.
Results
From the final 58 papers, 1 cross-modality convergence cluster and 7 dimension-specific profiles were identified, which are mapped by modality, circuit and clinical symptom in Figure 1.
Cross-modality convergence cluster
The Anticipatory Anhedonia Track (Pharmacological Convergence): Placebo-controlled clinical trials of intravenous Ketamine (3) and the KCNQ potassium channel opener Ezogabine (4) converged on frontostriatal connectivity. While ketamine modulated the subgenual anterior cingulate cortex (sgACC)- perigenual ACC (pgACC), sgACC-ventral striatal, and sgACC-dorsal ACC (dACC) connectivity (3), ezogabine reduced the rsFC between the reward seeds and the posterior cingulate cortex/precuneus node of the DMN (4). Both treatments target different hubs of this shared reward pipeline to selectively improve anticipatory anhedonia (3,4).
Dimension Specificity profiles
Rumination subtypes
DLPFC-PCC Axis (Neurofeedback): Real-time fMRI neurofeedback targeting the FPN and DMN networks highlights this resolution. Training patients to consciously restore the natural anticorrelation between the dorsolateral prefrontal cortex (dlPFC) and the precuneus selectively reduced brooding rumination (5). This normalization strongly predicts reductions in brooding, showing no association with trait anxiety, reflection, or global depression severity (5).
Electroconvulsive therapy (ECT): ECT targets the right orbitofrontal cortex (rOFC), where lower baseline connectivity predicts brooding (6). By enhancing rOFC integration and rOFC–DMN coupling, ECT enables the prefrontal cortex to downregulate hyperactive DMN nodes that sustain rumination (6).
Citalopram: Acute citalopram infusion selectively altered DMN connectivity based on baseline rumination traits, without causing a global effect across the broader MDD diagnosis. Higher baseline rumination scores predicted citalopram-induced network changes (e.g., excluding the amygdala and restoring precuneus integration), demonstrating that chemical manipulation targets the specific circuit driving negative thinking (7).
RSC–rTPJ Axis (Neurofeedback): A real-time fMRI neurofeedback trial targeted repetitive negative thinking without altering the broader diagnosis. Down-regulating connectivity between the retrosplenial cortex (RSC) and right temporoparietal junction (rTPJ) directly predicted reductions in baseline brooding (8).
Anhedonia
Repeated Ketamine infusions: An open-label trial mapping repeated ketamine infusions demonstrated that treatments relieved specific symptoms by targeting entirely separate brain pathways. An early decrease in amygdala-salience connectivity uniquely predicted reduced anxiety, whereas an early increase in hippocampus-executive communication predicted reduced anhedonia (9).
Interoception
Mindfulness-based Cognitive Therapy (MBCT): In a randomized controlled fMRI study, MBCT decoupled the SN from the lingual gyrus during active rumination, driving selective improvements in attention to bodily sensations. DMN connectivity remained unaffected despite symptom improvement, dissociating salience-mediated interoceptive regulation from the DMN role in self-referential thought (10).
Anxiety
Transcutaneous Vagus Nerve Stimulation (tVNS): tVNS administered as a monotherapy increased rsFC between the right amygdala and left dlPFC, resulting in a reduction of anxiety and psychic retardation HAMD subscales (11).
Methodological Exclusion
Transcranial Magnetic Stimulation: TMS literature failed to demonstrate dimension-specificity independent of global severity. TMS was identified as a baseline predictive marker for anhedonia rather than mapping a post-interventional dimension change (12). Both DLPFC and DMPFC rTMS were shown to improve total HAMD/HAMA scores without dimensional specificity (13).
Discussion
This review moves beyond global scales, mapping how each treatment modality targets distinct brain circuits and symptoms. While anticipatory anhedonia responds to frontostriatal-DMN decoupling across modalities (3,4), rumination subtypes, interoceptive regulation and anxiety maintain distinct neural signatures. Instead of driving global depression reduction, these treatments act on distinct behavioral dimensions.
Treatment modalities can be mapped to match the patient’s dominant symptom over total severity: neurofeedback targets the dlPFC-PCC tract to calm brooding (5), mindfulness decouples salience-lingual pathways for interoceptive focus (10), and tVNS couples the amygdala to L-dlPFC for anxiety (11). Citalopram and mindfulness’s failure to normalize the DMN challenges it as a baseline therapeutic target (7,10). While TMS met initial criteria, it functioned as a predictive biomarker in the studies tracking only global severity and lacked the specific symptom-connectivity correlations needed to map dimensional mechanisms. These distinct circuit signatures prove that global symptom scales blur highly specific, mechanistically separable subtypes and that dimension-matched interventions help bypass the standard trial-and-error process.
Limitations include varied patient severity, scanning parameters, and connectivity methodologies across the literature. Furthermore, the lack of placebo controls in several studies limits the ability to confirm direct causality. Ultimately, prospective trials designed to directly test these dimensional circuit targets are the critical next step.
Disclosures
Nothing to disclose by any author(s).
References
1. Drysdale AT, Grosenick L, Downar J, et al. Resting-state connectivity biomarkers define neurophysiological subtypes of depression. Nat Med. 2017;23(1):28-38. doi:10.1038/nm.4246
2. Williams LM. Precision psychiatry: a neural circuit taxonomy for depression and anxiety. Lancet Psychiatry. 2016;3(5):472-480. doi:10.1016/S2215-0366(15)00579-9
3. Alexander L, Hawkins PCT, Evans JW, Mehta MA, Zarate CA Jr. Preliminary evidence that ketamine alters anterior cingulate resting-state functional connectivity in depressed individuals. Transl Psychiatry. 2023;13(1):371. Published 2023 Dec 1. doi:10.1038/s41398-023-02674-1
4. Chowdhury A, Boukezzi S, Costi S, et al. Effects of the KCNQ (Kv7) Channel Opener Ezogabine on Resting-State Functional Connectivity of Striatal Brain Reward Regions, Depression, and Anhedonia in Major Depressive Disorder: Results From a Randomized Controlled Trial. Biol Psychiatry. 2025;98(7):568-577. doi:10.1016/j.biopsych.2025.02.897
5. Taylor JE, Yamada T, Kawashima T, et al. Depressive symptoms reduce when dorsolateral prefrontal cortex-precuneus connectivity normalizes after functional connectivity neurofeedback. Sci Rep. 2022;12(1):2581. Published 2022 Feb 16. doi:10.1038/s41598-022-05860-1
6. Wang H, Zhang T, Zu M, et al. Electroconvulsive therapy enhances degree centrality in the orbitofrontal cortex in depressive rumination. Psychiatry Res Neuroimaging. 2024;337:111765. doi:10.1016/j.pscychresns.2023.111765
7. Dutta, A., McKie, S., Downey, D. et al. Regional default mode network connectivity in major depressive disorder: modulation by acute intravenous citalopram. Transl Psychiatry 9, 116 (2019). https://doi.org/10.1038/s41398-019-0447-0
Tsuchiyagaito A, Misaki M, Kirlic N, et al. Real-Time fMRI Functional Connectivity Neurofeedback Reducing Repetitive Negative 8. Thinking in Depression: A Double-Blind, Randomized, Sham-Controlled Proof-of-Concept Trial. Psychother Psychosom. 2023;92(2):87-100. doi:10.1159/000528377
9. Vasavada MM, Loureiro J, Kubicki A, et al. Effects of Serial Ketamine Infusions on Corticolimbic Functional Connectivity in Major Depression. Biol Psychiatry Cogn Neurosci Neuroimaging. 2021;6(7):735-744. doi:10.1016/j.bpsc.2020.06.015
10. van der Velden AM, Scholl J, Elmholdt EM, et al. Mindfulness Training Changes Brain Dynamics During Depressive Rumination: A Randomized Controlled Trial. Biol Psychiatry. 2023;93(3):233-242. doi:10.1016/j.biopsych.2022.06.038
11. Liu J, Fang J, Wang Z, et al. Transcutaneous vagus nerve stimulation modulates amygdala functional connectivity in patients with depression. J Affect Disord. 2016;205:319-326. doi:10.1016/j.jad.2016.08.003
12. Wade B, Barbour T, Ellard K, Camprodon J. Predicting Dimensional Antidepressant Response to Repetitive Transcranial Magnetic Stimulation using Pretreatment Resting-state Functional Connectivity. Preprint. Res Sq. 2023;rs.3.rs-3204245. Published 2023 Aug 7. doi:10.21203/rs.3.rs-3204245/v1
13. Jin J, Wang Y, Liang S, et al. Effects of individualized rTMS on functional connectivity related to the default mode network and frontal-parietal network in major depressive disorder: exploratory analysis of a randomized controlled trial. Neuroimage Clin. 2025;48:103902. doi:10.1016/j.nicl.2025.103902