Gut–Brain Axis Drives Psychiatric Disease

Full assessment

From here the review is written for a technical reader: caps and thresholds, competing models, the independence basis, and the citation ledger.

Sources: 62 primary reports; claims: 66; verified: 66; unverified: 0. Verification is bounded attribution. Access: 7 partial-body and 55 complete abstract/status examinations. Cutoff: 2026-10-05. No complete-paper audit or blinded calibration claimed. Adversarial review: Generated by Sol 6.1; adversarially reviewed by Claude Opus 5.5 (Anthropic) on 2026-10-06. 2 corrections applied (see ADVERSARIAL REVIEW AND APPROVAL). Status: Approved as revised.

AT A GLANCE

Microbes causally influence defined animal behaviours, and some human adjunctive trials report benefit. A general upstream, disease-modifying microbial cause of depression, anxiety and psychosis remains unestablished.

Level	Confidence	Causal role	Momentum	Stability	Divergence
Β — Associational	Moderate	reinforcing / stabilizing factor	Mechanistic only	Stable	Contested

Deciding experiment: Randomize prospectively defined MDD or first-episode psychosis participants to a functionally characterized consortium versus an otherwise matched consortium lacking the nominated microbial enzyme, with an add-back arm, stable medication and diet, and independent multisite replication. Register diagnosis-specific primary endpoints, engraftment, mediator exposure, masking, missing-data policy and follow-up before enrollment.

Belief versus evidence: The preregistered weakest-link aggregation returns Β because SC1 is descriptive association. This is a consequence of that conjunction and does not downgrade the demonstrated animal mechanisms to association. Primary-subclaim and scope-matched sensitivities return Γ for SC3. Positive human treatment comparisons warrant causal credit for the tested treatment; they do not alone establish that the specified microbial mechanisms are an upstream general cause. No necessity requirement is imposed on a merely contributory mechanism.

Best-supported rival: Bounded microbial amplification. The neural-gland manipulation [42] distinguishes a host-to-gut arrow from a purely upstream microbial account. The trial labelled ITT versus compliant mITT comparison [13] and the PROVIT clinical/metabolic dissociation [15,56] distinguish robust clinical mechanism from parallel biomarker change.

COMPETING MECHANISTIC MODELS

Rank 1: Bounded microbial amplification (Ε). Microbial functions influence specific behaviours under defined antecedent stress, diet or developmental conditions. Vagal and central CB1 contrasts plus species and receptor negatives narrow distinct mechanisms [7,10,40–44]. These compatible pathways do not establish a single universal microbial signature or human disease initiator.

Rank 2: Stress-to-brain and stress-to-gut feedback (Ε). An upstream host state changes microbial ecology, which can feed back on brain function. Neural circuit/gland perturbations establish host-to-gut causation; donor-transfer rescue establishes a feedback leg [4,10,42]. One direction does not exclude the other; human relative contributions remain unresolved.

Rank 3: Medication-related dysbiosis (Γ). Psychotropics directly alter microbial growth and community composition. Drug screens and cohort exposure contrasts support medication effects [36,54,57,61]; that support is in vitro plus observational, so universal caps 1 and 2 hold this model at Γ. Drug-free cohorts and adjusted associations persist; medication cannot explain every signal [28,53,55,62].

Rank 4: Diet, transit and illness-state confounding (Γ). Host behaviour and GI physiology create microbial associations with diagnosis. IBS overlap, diet analyses and treatment-conditioned follow-up provide material alternatives [35–36,59–60]. The ASD diet example is an external comparator, not direct proof for every registered psychiatric disease.

Rank 5: General upstream microbial psychiatric cause (Γ). Dysbiosis initiates psychiatric disease and its correction durably modifies disease. Human-donor transfer and several clinical treatment signals are its strongest supports [1–2,16–19,23,37]. Animal proxies, heterogeneity, null trials and unverified human mediation block this full-scope interpretation.

The tournament does not alter the nominal level. Compatible pathways and flexible alternatives are not automatically proved by explaining observations.

FULL REVIEW

Registered claim and scope

Gut microbial dysbiosis is an upstream cause of human psychiatric disease, acting through microbial metabolites, immune and vagal pathways, and correction of dysbiosis is disease-modifying.

The index prescribes weakest-link aggregation. Its descriptive SC1 can impose an associational floor; this limitation and primary-subclaim/scope sensitivities will be exposed rather than changing the rule after evidence appraisal.

Selection: public-interest rank 7 from the editorial ranking compiled 2026-09-23; categorical prior verdicts were visible. Private fresh review; earlier documents are preserved.

Evidence and causal interpretation

The registered proposition contains a real biological communication system and a stronger etiological assertion. Human depression and anxiety cohorts reproduce some microbial associations, including after psychotropic adjustment, but alpha diversity can increase, decrease or be null [1,28,32,62]. A generic “dysbiosis” score therefore cannot substitute for a specified microbial function, exposure and host response.

Two 2016 studies transferred human depression-donor material to depleted rats or germ-free mice [1–2]. These experiments establish a recipient phenotype beyond ordinary association. They do not establish that microbes initiated the donors’ illness, exclude every fecal component, or turn rodent immobility into a human syndrome. Crude transfer earns scoped model credit while its human and mediator inference remains bounded.

The strongest model evidence goes beyond transfer. Vagotomy before stress-donor inoculation blocked behavioural and neurogenic effects [7]. Endocannabinoid enhancement rescued behaviour and central CB1 blockade reversed that rescue; a peripherally restricted antagonist did not reproduce that reversal [10]. Defined Bacteroides species showed positive and negative effects [40]. These interventions narrow causal mechanisms and expose boundaries rather than merely adding correlated biomarkers.

Engineered bacterial production of 4EP and downstream 4EPS, myelin abnormalities and oligodendrocyte-directed prevention identify a defined mouse metabolite pathway [41]. L. reuteri experiments add vagal and oxytocin-receptor dependence across neurodevelopmental models [44]. These are informative SC2 mechanisms; they do not expand the registered human diseases to autism after observing positive results. Early-life HPA and fear-extinction windows similarly identify developmental constraints [9,58].

Directionality is bidirectional. Neural circuit activation and Brunner-gland ablation alter Lactobacillus and host defence [42]. Psychotropic compounds inhibit commensal strains in vitro [57], and antidepressant exposure can explain more community variation than diagnosis in some human cohorts [61]. Yet diagnosis associations survive medication, GI and diet adjustment in Lifelines [62]. A medication-only explanation is inadequate, as is treating persistence after adjustment as proof of microbial initiation.

Human treatment evidence is mixed. CCFM1025, an older MDD probiotic trial and a probiotic/prebiotic comparison report positive depression-score contrasts [16–17,23]. IBS-associated depression improved with NCC3001, while anxiety and community profiles did not [18]. These trials support effects of particular interventions in enrolled populations; peripheral metabolite changes do not automatically establish the nominated mediator or disease onset.

Schaub is a particularly important full-text check. Sixty participants were included, 47 completed and the compliant mITT set contained 43. The primary HAM-D time-by-group interaction was significant in mITT but not the paper-labelled ITT sample. Adherence above 65% defined inclusion in mITT, rather than exclusion of highly adherent participants. Lactobacillus change was associated with symptoms within the probiotic group; it was not randomized mediation [13]. Consensus correctly located the analysis dependence, but its wording around the adherence cutoff required correction against original Results.

The Nikolova report was a feasibility and treatment-effect pilot, with a separately published mechanistic follow-up using the same NCT03893162 participants [14,26]. PROVIT changed bacterial profiles and stool metabolites while clinical comparisons and serum metabolomics remained null [15,56]. LP299v cognition and subsequent metabolomics also share a trial family [25,27]. These paired reports enrich a mechanism discussion but do not multiply independent treatment replications.

Null outcomes are load-bearing. PS128 did not improve depression scores relative to placebo [24], low-mood probiotics were null in ITT [51], and the schizophrenia trial improved bowel difficulty without a PANSS total advantage [50]. The immune and Candida reports belong to that same schizophrenia program [48–49]. Somatic benefit, selected cognitive tests and subgroup trends must not replace the registered psychiatric endpoint.

The newest randomized FMT report materially updates the literature [19]. Its abstract reports greater HAMD-17 reduction at two and eight weeks, durable engraftment and bile-acid statistical mediation, but no significant remission difference at eight weeks. Neither Consensus nor original-body retrieval supplied complete Methods here. Remission is foregrounded but cannot be asserted to be the registered primary endpoint from the abstract alone. The registration identifier does not establish prospective timing. The review therefore records a new treatment signal and analytic mediator evidence, with an unresolved endpoint hierarchy and no confirmed mediator-blockade test.

Human drug-naive schizophrenia and multiomics results show that medication is not the sole source of every association [53–55], while shared causes and direction remain open. The linked erratum on the first-episode multiomics study is retained and no affected numerical claim is used. An ASD diet-confounding study is an explicitly external comparator [60]; it demonstrates a serious design possibility without directly refuting depression or psychosis causation.

Consensus improved this review by directing the counterevidence search, proposing temporal milestones and extracting trial analysis details that the abstract alone obscured. SciSpace additionally supplied the PROVIT metabolomics lead, which was independently retrieved and examined. Provider meters, citation counts and search-funnel numbers were not graded or counted as our examined papers. The final Β follows the locked conjunction rule; the Γ primary sensitivity and Ε model mechanisms are displayed so that this aggregation choice cannot be mistaken for a denial of gut–brain biology.

HYPOTHESIS STATE EVALUATION FRAMEWORK (HSEF v5.3)

Subclaim matrix and aggregation

Subclaim	Scope	Level	Decision
SC1 Gut microbial composition correlates reproducibly with specific human psychiatric states.	human depression, anxiety and psychotic disorders	Β	Human associations recur in independent cohorts, including medication-adjusted analyses, but diversity direction and taxonomic signatures vary. This existence subclaim contains association rather than causation [28,31–32,52–54,61–62].
SC2 Defined microbial manipulation changes behavioral phenotypes through identified mediators.	named experimental systems; human attribution separate	Ε	Defined strains, vagal subtraction, central CB1 blockade and metabolite exposure/rescue establish context-bounded mouse mechanisms. Negative species, developmental and receptor contrasts map boundaries [7,10,39–41,44,58–59].
SC3 These microbiota-to-brain mechanisms make a clinically meaningful causal contribution to human psychiatric disease.	human depressive, anxiety and psychotic disorders; diagnoses fixed at baseline	Γ	Human RCTs include positive, null and analysis-dependent results. Clinical efficacy has scoped causal support, but attribution to the nominated dysbiosis-to-brain pathways and durable disease modification across these diagnoses is unresolved [13–27,47–51].

Preregistered aggregation: weakest-link. The preregistered weakest-link aggregation returns Β because SC1 is descriptive association. This is a consequence of that conjunction and does not downgrade the demonstrated animal mechanisms to association. Primary-subclaim and scope-matched sensitivities return Γ for SC3. Positive human treatment comparisons warrant causal credit for the tested treatment; they do not alone establish that the specified microbial mechanisms are an upstream general cause. No necessity requirement is imposed on a merely contributory mechanism. Sensitivities: weakest-link Β; primary-subclaim Γ; scope-matched Γ.

Decisive-perturbation inventory and causal tally

[1] Depression-donor transfer induced anhedonia and anxiety-like behaviour in microbiota-depleted rats, with tryptophan changes. Crude transfer and recipient depletion do not identify a microbial mediator or show initiation in the human donors.

[2] MDD microbiota transferred behavioural and metabolic differences to germ-free recipients. Forced-swim immobility is a proxy, and metabolic covariation does not establish the required human mediator.

[3] Cohousing, transfer and defined commensal reconstitution altered social behaviour and synaptic plasticity after maternal high-fat diet. This is a neurodevelopmental model outside the registered human depression, anxiety and psychosis scope.

[5] Stress-donor microbiota transferred behaviour and complement-associated synaptic abnormalities to recipients. C3/CR3 changes accompany the phenotype; the abstract does not independently resolve every mediator necessity contrast.

[6] Transferred microbiota impaired fluoxetine response and neurogenesis; 5-hydroxytryptophan alleviated the deficits. A downstream precursor rescue is not proof that the same pathway initiates human MDD.

[7] Vagotomy before stress-donor inoculation abrogated behavioural, neurogenic and inflammatory effects; Methods describe blinded analysis. Subdiaphragmatic vagotomy is physiologically broad; effects and surgical boundaries are within the recipient model.

[9] Microbial depletion impaired fear extinction, with developmental-window dependence of restoration and altered neuronal activity. Fear learning is a defined behavioural function; identified metabolite differences alone do not establish human anxiety disease causation.

[10] Central eCB enhancement rescued transferred behaviours, central CB1 antagonism blocked rescue, and Lactobacillus complementation restored behavioural and lipid signals. Some individual contrasts were nonsignificant; downstream rescue narrows this model rather than establishing all human depression causes.

[11] Acute tryptophan depletion changed the behavioural phenotype and brain serotonin of germ-free mice more strongly than SPF mice. Depletion is a broad nutritional challenge and germ-free development differs from adult human dysbiosis.

[12] Roseburia hominis reduced microglial activation and depressive behaviours; SCFA treatments reproduced molecular anti-inflammatory effects. SCFA molecular rescue is not a demonstrated exclusive mediator of the behavioural effect in humans.

[16] Randomized CCFM1025 supplementation improved HDRS and MADRS change relative to placebo and altered serotonin-turnover measures. Small short-term trial; peripheral turnover and parallel microbial changes do not isolate the cause of benefit or prove durability.

[17] Probiotic supplementation improved BDI relative to placebo and prebiotic in MDD; prebiotic benefit and unadjusted mediator ratios were null. Attrition was material and adjusted metabolite associations do not establish a necessary microbial mediator.

[18] NCC3001 reduced depression scores and altered emotional brain responses in IBS, with no significant anxiety or IBS-symptom benefit. IBS with mood scores is distinct from primary MDD; fecal community profiles were similar and mediator attribution remains unresolved.

[23] An MDD randomized probiotic trial reported greater BDI improvement with changes in insulin, hs-CRP and glutathione. Multiple biomarkers and short follow-up do not identify a microbial mediator or sustained disease modification.

[29] Roseburia intestinalis rescued stress-related mouse behaviour and tryptophan metabolism alongside adolescent microbial associations. Human first diagnosis and post-sertraline profiles are observational; mouse rescue is not a randomized adolescent treatment result.

[34] CCFM1025 but not a comparator B. breve strain improved stress behaviour and produced distinct microbial/metabolic profiles. Strain comparisons narrow function but genomic predictions and parallel metabolites do not identify every mediator.

[37] A three-strain formula improved depression and GI scores relative to placebo; a separate stress model investigated serotonergic effects. The abstract does not supply complete human design or mediator-subtraction details; mouse mechanisms cannot be substituted for human mediation.

[38] B. fragilis ameliorated selected behaviours and barrier abnormalities; administration of an elevated metabolite induced behavioural abnormalities. Antecedent maternal immune activation and neurodevelopmental outcomes are outside the registered human SC3 scope.

[39] JB-1 changed stress behaviour and regional GABA receptor expression; effects were absent after vagotomy. Healthy-mouse stress responses and a broad nerve lesion do not prove human psychiatric onset or generalize to all strains.

[40] Specific Bacteroides species recapitulated behaviour and neurogenesis effects of MDD-donor transfer, while B. ovatus did not. Species boundaries are informative; depletion protocols and recipient proxies limit endogenous human inference.

[41] Defined 4EP biosynthesis increased 4EPS exposure, altered myelination and anxiety-like behaviour; oligodendrocyte-directed treatment prevented behaviour. Engineered exposure establishes a molecule pathway in mice, not its sufficient concentration or necessity in human anxiety disorders.

[44] L. reuteri rescued social deficits across models through vagus and oxytocin-receptor dependence, without restoring overall community composition. Strong bounded mechanism outside human SC3; therapeutic action without composition restoration challenges a generic normalization narrative.

[45] SCFA supplementation alleviated selected anhedonia, stress response and permeability changes while weight and some receptor outcomes remained unaffected. Supplementation does not establish endogenous deficiency as the initiating cause; effects were behavioural-test and stress-context specific.

[55] A metagenomic classifier replicated, and schizophrenia-enriched S. vestibularis induced mouse social and neurotransmitter changes. Human direction remains observational and mouse social deficits are not the full schizophrenia syndrome.

[59] Proline-linked human signals were complemented by dietary, transfer, transporter and microbial perturbations in mice and flies. Cross-system convergence supports a bounded pathway; no randomized human proline/microbial prevention or treatment mechanism was tested.

Tally: 25 discriminating contrasts. Null and negative contrasts are discriminators, not positive threshold crossings. Reports and replications are not independent causal lines. Positive clinical treatment comparisons have treatment-level credit; they are not counted as selective human mediator tests. The decisive flags for broad physiological host-to-gut and HPA endpoints are excluded from psychiatric-mechanism crossings.

Cap audit, precedence and Β/Γ/Δ boundary

“A perturbation crosses a threshold within the system it was actually run in. A universal cap governs the inference from that system to the hypothesis’s claimed scope.” Microbiome rubric04 governs transfer and strain function; psychiatry rubric07 governs diagnosis, endpoints and clinical masking. Inert placebo probiotic trials do not inherit an automatic psychedelic unblinding penalty. Rodent immobility alone is a weak human disease proxy, but convergent subtraction and rescue retain their actual model credit. SC3 is below Δ for the registered mechanism and generalization, not because a contributory mechanism must be universally necessary.

Gamma hard-rule audit: Crude transfer supports a recipient phenotype but does not identify the mediator in the human donor. Clinical treatment efficacy, statistical mediation and mediator necessity are separately assessed; no automatic observational cap is re-applied to a convergent set.

SC1 cap: “Pure association evidence — however large, however replicated — cannot exceed Β.” Human associations recur in independent cohorts, including medication-adjusted analyses, but diversity direction and taxonomic signatures vary. This existence subclaim contains association rather than causation [28,31–32,52–54,61–62].

SC1 field-valid definition: “ - Gnotobiotic or transfer experiment showing phenotype movement, at minimum - Even without defined-consortium precision or robust rescue, a transfer paradigm in vivo is the minimum that crosses this threshold

Per Cap precedence above, a gnotobiotic or forced-transfer experiment crosses Δ for the recipient-animal subclaim it was actually run in, while universal cap 5 (seeding or forced-transfer paradigms cap at Γ) still blocks the inference from that result to endogenous human biology — score the two scopes separately rather than choosing between the rules.” Positive full-scope result: no. Human associations recur in independent cohorts, including medication-adjusted analyses, but diversity direction and taxonomic signatures vary. This existence subclaim contains association rather than causation [28,31–32,52–54,61–62].

SC2 field-valid definition: “ - Gnotobiotic or transfer experiment showing phenotype movement, at minimum - Even without defined-consortium precision or robust rescue, a transfer paradigm in vivo is the minimum that crosses this threshold

Per Cap precedence above, a gnotobiotic or forced-transfer experiment crosses Δ for the recipient-animal subclaim it was actually run in, while universal cap 5 (seeding or forced-transfer paradigms cap at Γ) still blocks the inference from that result to endogenous human biology — score the two scopes separately rather than choosing between the rules.” Positive full-scope result: yes. Defined strains, vagal subtraction, central CB1 blockade and metabolite exposure/rescue establish context-bounded mouse mechanisms. Negative species, developmental and receptor contrasts map boundaries [7,10,39–41,44,58–59].

SC3 field-valid definition: “ - Gnotobiotic or transfer experiment showing phenotype movement, at minimum - Even without defined-consortium precision or robust rescue, a transfer paradigm in vivo is the minimum that crosses this threshold

Per Cap precedence above, a gnotobiotic or forced-transfer experiment crosses Δ for the recipient-animal subclaim it was actually run in, while universal cap 5 (seeding or forced-transfer paradigms cap at Γ) still blocks the inference from that result to endogenous human biology — score the two scopes separately rather than choosing between the rules.” Positive full-scope result: no. Human RCTs include positive, null and analysis-dependent results. Clinical efficacy has scoped causal support, but attribution to the nominated dysbiosis-to-brain pathways and durable disease modification across these diagnoses is unresolved [13–27,47–51].

Upward threshold ladder

SC1 Δ: “Requires a transfer paradigm or causal perturbation in vivo.” Satisfied: no. A causal interpretation would require an amended causal subclaim rather than upgrading descriptive association by sample size.

SC2 Δ: “Requires a transfer paradigm or causal perturbation in vivo.” Satisfied: yes.

SC2 Ε: “Requires context-specific causal effect with defined limits. A transferable phenotype exists, but microbial identity or generalizability remains incomplete.” Satisfied: yes.

SC2 Ζ: “Requires: - Defined consortium or strain-level causal evidence - Rescue logic - Replication across at least two relevant models or cohorts” Satisfied: no. Orthogonal microbial gene subtraction and add-back at measured endogenous exposure, with matched behaviour and off-target controls across independent systems.

SC3 Δ: “Requires a transfer paradigm or causal perturbation in vivo.” Satisfied: no. Prospectively registered human psychiatric benefit linked to a specified microbial function by mediator-selective subtraction/add-back or a valid instrument, with durability, replication and diagnosis-specific negative controls.

Upper ladder: Valid human therapeutic perturbations and natural experiments remain available; unethical disease induction is unnecessary. Clinical symptom RCTs supply treatment-level causal evidence, and animal strain/metabolite/vagal tests support distinct mechanisms. Their convergence is assessed as a set, but no examined set establishes the general upstream human attribution, sustained function and cross-diagnosis prediction needed for the higher thresholds.

Independence: No count of laboratories is equated with independence. The experimental pathways are biologically distinct but share behavioural-proxy and model-exposure limitations. Human RCTs distinguish allocation from mediator causation; reports from one cohort remain one family. No higher-level human convergence is claimed.

Recipient model subtraction and rescue [7, 10, 44]: Excludes relying only on human association; retains surgical/pharmacological and shared-model limitations.

Randomized human clinical treatment [16, 17, 23, 19]: Reduces treatment-allocation confounding but does not randomize the microbial mediator.

Human adjusted population associations [28, 62, 61]: Tests medication-only explanations while leaving temporality and residual confounding.

Falsification and pitfalls

Target engagement without clinical advantage. PROVIT changed microbiota/stool metabolites without clinical superiority [15,56]. Additional downgrade: False.

Somatic and psychiatric endpoints dissociate. Schizophrenia bowel difficulty improved while PANSS did not [50]. Additional downgrade: False.

Community normalization is not required for every microbial treatment effect. L. reuteri rescue occurred without community restoration [44]. Additional downgrade: False.

Therapeutic effect misread as mechanistic proof. Parallel symptom, taxon or metabolite changes in clinical trials [13,16,19,26].

Transferability misread as endogenous human primacy. Recipient depletion and behavioural proxies in donor-transfer experiments [1–2].

Counting overlapping reports as independent trials. Schaub, Nikolova, Rudzki, PROVIT and Dickerson secondary analyses share trial families.

Scope fitted to the result. No microbial responder subgroup is adopted after seeing benefit; baseline diagnoses and diet are fixed before outcome.

Adversarial steelman: above, then below

The case for Γ overall is that experimental biology is extensive and SC3 contains randomized human signals, including FMT with engraftment and statistical mediation [19]. This is compelling for the primary-subclaim sensitivity. A literal weakest-link rule, however, cannot promote the associational SC1 by importing SC2 evidence.

The case for Α is that human signatures are heterogeneous and several treatments are null [15,24,50–51]. It fails because reproducible adjusted associations exist and negative treatment formulations do not erase them [28,62].

Retain Β under the registered rule; display Γ as the primary and scope-matched sensitivities. Reject Α because the phenomenon is observed, and reject an undisclosed switch to Γ aggregation. The conjunction design, rather than a generic claim that all microbiome evidence is weak, explains this result.

Final score, causal direction and M/S/D

The preregistered weakest-link aggregation returns Β because SC1 is descriptive association. This is a consequence of that conjunction and does not downgrade the demonstrated animal mechanisms to association. Primary-subclaim and scope-matched sensitivities return Γ for SC3. Positive human treatment comparisons warrant causal credit for the tested treatment; they do not alone establish that the specified microbial mechanisms are an upstream general cause. No necessity requirement is imposed on a merely contributory mechanism.

Within animal stress-transfer models, microbial manipulation sustains or amplifies phenotypes after an antecedent stress or diet exposure. Human upstream initiation and rate limitation are not established. Brain-to-gut neural perturbations and drug effects demonstrate that direction can also run from host state to microbial ecology [42–43,57].

Momentum: Mechanistic only. The 2026 randomized FMT adjunctive MDD study directly perturbs microbial communities and measures symptoms and engraftment [19]. Its null remission contrast and abstract-only access qualify, rather than erase, its causal-test momentum.

Stability: Stable. No specific verified pending result is used to claim Fragile status. A prospectively registered mediator-specific multisite trial could materially change SC3; this is a proposed test, not a known forthcoming readout.

Divergence: Contested; no single implied-level arithmetic. These dimensions do not change level or confidence.

PROPOSED DECISIVE EXPERIMENTS

Randomize prospectively defined MDD or first-episode psychosis participants to a functionally characterized consortium versus an otherwise matched consortium lacking the nominated microbial enzyme, with an add-back arm, stable medication and diet, and independent multisite replication. Register diagnosis-specific primary endpoints, engraftment, mediator exposure, masking, missing-data policy and follow-up before enrollment.

Readout: Sustained psychiatric symptoms and functioning plus time-resolved microbial function, serum mediator and CNS-relevant target engagement. Controls: Matched consortia, off-target and GI symptom controls, credible masking, stable psychotropics, measured diet/transit, randomized mediator function and baseline strata fixed before outcomes. Supporting: Benefit follows the nominated microbial function, disappears with selective subtraction and returns with add-back, without being explained by GI comfort or altered drug exposure. Falsifying: Adequate target engagement leaves the registered psychiatric endpoint unchanged, or equivalent benefit occurs with the function-deleted comparator.

RESEARCH TIMELINE

Year	Milestone	Primary source	Interpretation
2004	Developmental microbial programming of HPA responses	PMID15133062	Timing and strain boundaries in mice
2011	Defined Lactobacillus and vagal subtraction	PMID21876150	Model pathway causality
2013	Maternal immune activation and microbial rescue	PMID24315484	Neurodevelopmental model; outside human SC3
2014	Schizophrenia probiotic trial	PMID24940526	Bowel benefit with null PANSS
2016	Human depression-donor transfer	PMID27491067	Recipient behaviour, not human initiation
2017	ITT low-mood probiotic null	PMID28068788	Formulation and endpoint limits
2019	Population metagenomics	PMID30718848	Adjusted replicated association
2020	Endocannabinoid rescue and blockade	PMID33311466	Narrows a transferred mouse phenotype
2022	Probiotic MDD trial with analysis dependence	PMID35654766	mITT advantage; paper-labelled ITT interaction null
2023	Clinical pilot and FMT feasibility	PMID37314797	Effect-size estimates rather than definitive mechanism
2025	Medication-associated community effects	PMID39794490	Confounding remains material
2026	Adjunctive randomized FMT trial	PMID42309058	Symptom signal and statistical mediation; remission null and full Methods unavailable

KNOWLEDGE GAPS

Prospectively registered, mediator-specific human perturbation; statistical mediation should report sequential exchangeability and alternative mediator assumptions.

Separate symptom benefit from remission, functioning and persistence after treatment withdrawal.

Recruit diagnosis-specific medication-naive or stable-medication cohorts with measured diet, transit and GI symptoms, rather than adjusting poorly measured confounders.

Map strain, microbial gene, metabolite exposure and host receptor boundaries; no single alpha-diversity target is justified.

Obtain complete trial protocols, supplements and correction notices before treating the newest FMT or corrected multiomics reports as decisive publication evidence.

METHODS AND CONSENSUS CONTRIBUTION

Consensus was used through the signed-in browser for discovery, counterevidence, research-gap and timeline suggestions, and checks of specific indexed full-text passages. Extracted Methods/Results details were reconciled with original-paper passages where available. Provider screening counts, citation counts and meters were not used as primary counts or HSEF scores. This source log records access depth; it does not claim complete-paper or supplement review. The monthly API allowance was conserved: zero Consensus API calls in this batch.

REVIEWER NOTES

62 primary reports have individually authored result/limit cards. Seven original bodies were examined in selected passages; the remaining sources were examined at complete abstract/status depth. Complete papers, supplements and all correction details were not audited. SciSpace discovery yielded mostly reviews and a duplicate review; only a separately examined primary PROVIT report entered the primary ledger. Consensus checked particular Schaub Methods/Results passages, while the 2026 FMT report remained abstract-only. This is a bounded critical review, not an exhaustive systematic review.

Search leads, reviews, protocols and correction notices do not pad the primary count. Downloaded XML does not earn examination credit. Linked notices remain attached; unexamined correction details limit inference. This is a coordinator source/claim consistency assessment, not independent blinded calibration.

ADVERSARIAL REVIEW AND APPROVAL

Authoring model: Sol 6.1. Adversarial reviewer: Claude Opus 5.5 (Anthropic). Date: 2026-10-06.

Scope of checks. All 62 reference PMIDs re-verified against PubMed records (title, first author, journal, year, DOI, status flags): 62/62 resolve to the stated work; no retractions or expressions of concern; the only year differences are print-issue versus e-pub year. 18 load-bearing claims spot-checked against PubMed abstracts and the local full-text/ and body-presentation-receipts/ passages, including the Schaub trial's ITT/mITT analysis dependence and compliance cut-off, the Nikolova pilot and its mechanistic follow-up sharing NCT03893162, PROVIT's clinical-versus-microbial dissociation, the 2026 adjunctive FMT trial's HAMD-17 and remission results, the PS128, Romijn and Dickerson nulls, the Lifelines adjustment set, the Kelly and Zheng transfer experiments, the vagotomy and endocannabinoid rescue experiments, and the ASD diet-confounding study. Scoring logic re-derived from framework/hsef-hypothesis-review-v5/SKILL.md and references/hsef-framework.md: the index-prescribed weakest-link aggregation, cap order, the Β/Γ/Δ boundary, the reported rule sensitivities, and the competing-model levels.

#	Severity	Location	Original → corrected	Evidence
1	Minor	COMPETING MECHANISTIC MODELS, Rank 3; `competing_models` in the score card	Medication-related dysbiosis graded Δ → Γ, with the cap stated in the rationale	The cited support is an in vitro drug screen (PMID 29555994) and observational or uncontrolled before-after human cohorts (PMIDs 34290352, 32065949, 39794490). Framework universal caps: "In vitro evidence alone cannot exceed Γ" and "Pure association evidence — however large, however replicated — cannot exceed Β"; the microbiome rubric's Δ minimum is an in vivo transfer or gnotobiotic paradigm, which none of these supply. Per Step 13 the tournament never changes the nominal level, so the headline Β is unaffected.
2	Minor	AT A GLANCE table; "Momentum:" line; `extended_scoring_MSD.momentum.value`	Momentum "Forward" → "Mechanistic only"	"Forward" is not a value on the framework's Momentum scale (Decisive work underway / Mechanistic only / Associational only / Little new work / Abandoned). The recorded artifact names completed clinical and mechanistic work and the review states that no pending threshold-crossing result is claimed. M/S/D never modify the level or confidence.

Rejected findings (checked, no change). (a) Headline Β under weakest-link while SC2 is Ε — the index prescribes weakest-link for this entry and the review reports the primary-subclaim and scope-matched sensitivities (both Γ) as the framework requires; the rule was pre-registered, not chosen after scoring. (b) SC1 at Β — universal cap 2 binds a purely associational existence subclaim; sample size cannot lift it. (c) The Schaub description ("60 included, 47 completed, mITT 43, time-by-group interaction significant only in the mITT sample, compliance cut-off >65% defining inclusion") — verified verbatim against the article's Results. (d) "Diagnosis associations survive medication, GI and diet adjustment in Lifelines" — verified against the article's Methods and Results (adjustment for psychotropics, functional GI disorders and IBS, antibiotics and diet-quality score). (e) Divergence "Contested" without arithmetic — Contested is a defined value for a genuinely split field. (f) Citation years differing by one from the PubMed e-pub year — print/issue-year convention.

Residual limitations. Access remains bounded: 7 partial-body and 55 abstract/status examinations, no complete-paper or supplement audit, and the 2026 FMT trial was examined at abstract depth only, so its endpoint hierarchy remains unresolved. This pass checked citation integrity, attribution and framework conformance; it is not independent blinded peer review.

Approval: Following adversarial review and the corrections listed above, the revised document is approved. — Claude Opus 5.5 (Anthropic), 2026-10-06.

REFERENCES

Kelly JR, Borre Y, O' Brien C et al. Transferring the blues: Depression-associated gut microbiota induces neurobehavioural changes in the rat. Journal of psychiatric research. 2016. PMID 27491067. DOI: 10.1016/j.jpsychires.2016.07.019. Access: abstract_metadata. Study: PMID27491067.
Zheng P, Zeng B, Zhou C et al. Gut microbiome remodeling induces depressive-like behaviors through a pathway mediated by the host's metabolism. Molecular psychiatry. 2016. PMID 27067014. DOI: 10.1038/mp.2016.44. Access: abstract_metadata. Study: PMID27067014.
Buffington SA, Di Prisco GV, Auchtung TA et al. Microbial Reconstitution Reverses Maternal Diet-Induced Social and Synaptic Deficits in Offspring. Cell. 2016. PMID 27315483. DOI: 10.1016/j.cell.2016.06.001. Access: abstract_metadata. Study: PMID27315483.
Deng Y, Zhou M, Wang J et al. Involvement of the microbiota-gut-brain axis in chronic restraint stress: disturbances of the kynurenine metabolic pathway in both the gut and brain. Gut microbes. 2021. PMID 33535879. DOI: 10.1080/19490976.2020.1869501. Access: abstract_metadata. Study: PMID33535879.
Hao W, Ma Q, Wang L et al. Gut dysbiosis induces the development of depression-like behavior through abnormal synapse pruning in microglia-mediated by complement C3. Microbiome. 2024. PMID 38378622. DOI: 10.1186/s40168-024-01756-6. Access: abstract_metadata. Study: PMID38378622.
Siopi E, Chevalier G, Katsimpardi L et al. Changes in Gut Microbiota by Chronic Stress Impair the Efficacy of Fluoxetine. Cell reports. 2020. PMID 32187541. DOI: 10.1016/j.celrep.2020.02.099. Access: abstract_metadata. Study: PMID32187541.
Siopi E, Galerne M, Rivagorda M et al. Gut microbiota changes require vagus nerve integrity to promote depressive-like behaviors in mice. Molecular psychiatry. 2023. PMID 37131071. DOI: 10.1038/s41380-023-02071-6. Access: full_text_partial. Study: PMID37131071.
Hoban AE, Stilling RM, Ryan FJ et al. Regulation of prefrontal cortex myelination by the microbiota. Translational psychiatry. 2016. PMID 27045844. DOI: 10.1038/tp.2016.42. Access: abstract_metadata. Study: PMID27045844.
Chu C, Murdock MH, Jing D et al. The microbiota regulate neuronal function and fear extinction learning. Nature. 2019. PMID 31645720. DOI: 10.1038/s41586-019-1644-y. Access: abstract_metadata. Study: PMID31645720.
Chevalier G, Siopi E, Guenin-Macé L et al. Effect of gut microbiota on depressive-like behaviors in mice is mediated by the endocannabinoid system. Nature communications. 2020. PMID 33311466. DOI: 10.1038/s41467-020-19931-2. Access: full_text_partial. Study: PMID33311466.
Lukić I, Getselter D, Koren O et al. Role of Tryptophan in Microbiota-Induced Depressive-Like Behavior: Evidence From Tryptophan Depletion Study. Frontiers in behavioral neuroscience. 2019. PMID 31231198. DOI: 10.3389/fnbeh.2019.00123. Access: abstract_metadata. Study: PMID31231198.
Song L, Sun Q, Zheng H et al. Roseburia hominis Alleviates Neuroinflammation via Short-Chain Fatty Acids through Histone Deacetylase Inhibition. Molecular nutrition & food research. 2022. PMID 35819092. DOI: 10.1002/mnfr.202200164. Access: abstract_metadata. Study: PMID35819092.
Schaub AC, Schneider E, Vazquez-Castellanos JF et al. Clinical, gut microbial and neural effects of a probiotic add-on therapy in depressed patients: a randomized controlled trial. Translational psychiatry. 2022. PMID 35654766. DOI: 10.1038/s41398-022-01977-z. Access: full_text_partial. Study: NCT02957591.
Nikolova VL, Cleare AJ, Young AH et al. Acceptability, Tolerability, and Estimates of Putative Treatment Effects of Probiotics as Adjunctive Treatment in Patients With Depression: A Randomized Clinical Trial. JAMA psychiatry. 2023. PMID 37314797. DOI: 10.1001/jamapsychiatry.2023.1817. Access: full_text_partial. Study: NCT03893162.
Reininghaus EZ, Platzer M, Kohlhammer-Dohr A et al. PROVIT: Supplementary Probiotic Treatment and Vitamin B7 in Depression-A Randomized Controlled Trial. Nutrients. 2020. PMID 33171595. DOI: 10.3390/nu12113422. Access: full_text_partial. Study: NCT03300440.
Tian P, Chen Y, Zhu H et al. Bifidobacterium breve CCFM1025 attenuates major depression disorder via regulating gut microbiome and tryptophan metabolism: A randomized clinical trial. Brain, behavior, and immunity. 2022. PMID 34875345. DOI: 10.1016/j.bbi.2021.11.023. Access: abstract_metadata. Study: PMID34875345.
Kazemi A, Noorbala AA, Azam K et al. Effect of probiotic and prebiotic vs placebo on psychological outcomes in patients with major depressive disorder: A randomized clinical trial. Clinical nutrition (Edinburgh, Scotland). 2019. PMID 29731182. DOI: 10.1016/j.clnu.2018.04.010. Access: abstract_metadata. Study: PMID29731182.
Pinto-Sanchez MI, Hall GB, Ghajar K et al. Probiotic Bifidobacterium longum NCC3001 Reduces Depression Scores and Alters Brain Activity: A Pilot Study in Patients With Irritable Bowel Syndrome. Gastroenterology. 2017. PMID 28483500. DOI: 10.1053/j.gastro.2017.05.003. Access: abstract_metadata. Study: PMID28483500.
Li J, Wang Y, Li R et al. Adjunctive fecal microbiota transplantation for major depressive disorder: A randomized, double-blind, placebo-controlled trial. Cell host & microbe. 2026. PMID 42309058. DOI: 10.1016/j.chom.2026.05.017. Access: abstract_metadata. Study: PMID42309058.
Green JE, Berk M, Mohebbi M et al. Feasibility, Acceptability, and Safety of Faecal Microbiota Transplantation in the Treatment of Major Depressive Disorder: A Pilot Randomized Controlled Trial. Canadian journal of psychiatry. Revue canadienne de psychiatrie. 2023. PMID 36637229. DOI: 10.1177/07067437221150508. Access: abstract_metadata. Study: PMID36637229.
Schneider E, Doll JPK, Schweinfurth N et al. Effect of short-term, high-dose probiotic supplementation on cognition, related brain functions and BDNF in patients with depression: a secondary analysis of a randomized controlled trial. Journal of psychiatry & neuroscience : JPN. 2023. PMID 36653035. DOI: 10.1503/jpn.220117. Access: abstract_metadata. Study: NCT02957591.
Slykerman RF, Hood F, Wickens K et al. Effect of Lactobacillus rhamnosus HN001 in Pregnancy on Postpartum Symptoms of Depression and Anxiety: A Randomised Double-blind Placebo-controlled Trial. EBioMedicine. 2017. PMID 28943228. DOI: 10.1016/j.ebiom.2017.09.013. Access: abstract_metadata. Study: PMID28943228.
Akkasheh G, Kashani-Poor Z, Tajabadi-Ebrahimi M et al. Clinical and metabolic response to probiotic administration in patients with major depressive disorder: A randomized, double-blind, placebo-controlled trial. Nutrition (Burbank, Los Angeles County, Calif.). 2016. PMID 26706022. DOI: 10.1016/j.nut.2015.09.003. Access: abstract_metadata. Study: PMID26706022.
Lin SK, Kuo PH, Hsu CY et al. The effects of Lactobacillus plantarum PS128 in patients with major depressive disorder: an eight-week double-blind, placebo-controlled study. Asian journal of psychiatry. 2024. PMID 39232392. DOI: 10.1016/j.ajp.2024.104210. Access: abstract_metadata. Study: PMID39232392.
Rudzki L, Ostrowska L, Pawlak D et al. Probiotic Lactobacillus Plantarum 299v decreases kynurenine concentration and improves cognitive functions in patients with major depression: A double-blind, randomized, placebo controlled study. Psychoneuroendocrinology. 2019. PMID 30388595. DOI: 10.1016/j.psyneuen.2018.10.010. Access: abstract_metadata. Study: LP299V-MDD-2019.
Nikolova VL, Cleare AJ, Young AH et al. Exploring the mechanisms of action of probiotics in depression: Results from a randomized controlled pilot trial. Journal of affective disorders. 2025. PMID 39924003. DOI: 10.1016/j.jad.2025.01.153. Access: abstract_metadata. Study: NCT03893162.
Godzien J, Kalaska B, Rudzki L et al. Probiotic Lactobacillus plantarum 299v supplementation in patients with major depression in a double-blind, randomized, placebo-controlled trial: A metabolomics study. Journal of affective disorders. 2025. PMID 39271063. DOI: 10.1016/j.jad.2024.09.058. Access: abstract_metadata. Study: LP299V-MDD-2019.
Valles-Colomer M, Falony G, Darzi Y et al. The neuroactive potential of the human gut microbiota in quality of life and depression. Nature microbiology. 2019. PMID 30718848. DOI: 10.1038/s41564-018-0337-x. Access: abstract_metadata. Study: PMID30718848.
Zhou M, Fan Y, Xu L et al. Microbiome and tryptophan metabolomics analysis in adolescent depression: roles of the gut microbiota in the regulation of tryptophan-derived neurotransmitters and behaviors in human and mice. Microbiome. 2023. PMID 37386523. DOI: 10.1186/s40168-023-01589-9. Access: abstract_metadata. Study: PMID37386523.
Strandwitz P, Kim KH, Terekhova D et al. GABA-modulating bacteria of the human gut microbiota. Nature microbiology. 2019. PMID 30531975. DOI: 10.1038/s41564-018-0307-3. Access: abstract_metadata. Study: PMID30531975.
Yang J, Zheng P, Li Y et al. Landscapes of bacterial and metabolic signatures and their interaction in major depressive disorders. Science advances. 2020. PMID 33268363. DOI: 10.1126/sciadv.aba8555. Access: abstract_metadata. Study: PMID33268363.
Jiang H, Ling Z, Zhang Y et al. Altered fecal microbiota composition in patients with major depressive disorder. Brain, behavior, and immunity. 2015. PMID 25882912. DOI: 10.1016/j.bbi.2015.03.016. Access: abstract_metadata. Study: PMID25882912.
Amin N, Liu J, Bonnechere B et al. Interplay of Metabolome and Gut Microbiome in Individuals With Major Depressive Disorder vs Control Individuals. JAMA psychiatry. 2023. PMID 37074710. DOI: 10.1001/jamapsychiatry.2023.0685. Access: abstract_metadata. Study: PMID37074710.
Tian P, Bastiaanssen TFS, Song L et al. Unraveling the Microbial Mechanisms Underlying the Psychobiotic Potential of a Bifidobacterium breve Strain. Molecular nutrition & food research. 2021. PMID 33594816. DOI: 10.1002/mnfr.202000704. Access: abstract_metadata. Study: PMID33594816.
Han L, Zhao L, Zhou Y et al. Altered metabolome and microbiome features provide clues in understanding irritable bowel syndrome and depression comorbidity. The ISME journal. 2022. PMID 34750528. DOI: 10.1038/s41396-021-01123-5. Access: abstract_metadata. Study: PMID34750528.
Shen Y, Yang X, Li G et al. The change of gut microbiota in MDD patients under SSRIs treatment. Scientific reports. 2021. PMID 34290352. DOI: 10.1038/s41598-021-94481-1. Access: abstract_metadata. Study: PMID34290352.
Tian P, Zou R, Wang L et al. Multi-Probiotics ameliorate Major depressive disorder and accompanying gastrointestinal syndromes via serotonergic system regulation. Journal of advanced research. 2023. PMID 35618633. DOI: 10.1016/j.jare.2022.05.003. Access: abstract_metadata. Study: PMID35618633.
Hsiao EY, McBride SW, Hsien S et al. Microbiota modulate behavioral and physiological abnormalities associated with neurodevelopmental disorders. Cell. 2013. PMID 24315484. DOI: 10.1016/j.cell.2013.11.024. Access: abstract_metadata. Study: PMID24315484.
Bravo JA, Forsythe P, Chew MV et al. Ingestion of Lactobacillus strain regulates emotional behavior and central GABA receptor expression in a mouse via the vagus nerve. Proceedings of the National Academy of Sciences of the United States of America. 2011. PMID 21876150. DOI: 10.1073/pnas.1102999108. Access: abstract_metadata. Study: PMID21876150.
Zhang Y, Fan Q, Hou Y et al. Bacteroides species differentially modulate depression-like behavior via gut-brain metabolic signaling. Brain, behavior, and immunity. 2022. PMID 35143877. DOI: 10.1016/j.bbi.2022.02.007. Access: abstract_metadata. Study: PMID35143877.
Needham BD, Funabashi M, Adame MD et al. A gut-derived metabolite alters brain activity and anxiety behaviour in mice. Nature. 2022. PMID 35165440. DOI: 10.1038/s41586-022-04396-8. Access: abstract_metadata. Study: PMID35165440.
Chang H, Perkins MH, Novaes LS et al. Stress-sensitive neural circuits change the gut microbiome via duodenal glands. Cell. 2024. PMID 39121857. DOI: 10.1016/j.cell.2024.07.019. Access: abstract_metadata. Study: PMID39121857.
McVey Neufeld KA, Bienenstock J, Bharwani A et al. Oral selective serotonin reuptake inhibitors activate vagus nerve dependent gut-brain signalling. Scientific reports. 2019. PMID 31582799. DOI: 10.1038/s41598-019-50807-8. Access: abstract_metadata. Study: PMID31582799.
Sgritta M, Dooling SW, Buffington SA et al. Mechanisms Underlying Microbial-Mediated Changes in Social Behavior in Mouse Models of Autism Spectrum Disorder. Neuron. 2019. PMID 30522820. DOI: 10.1016/j.neuron.2018.11.018. Access: abstract_metadata. Study: PMID30522820.
van de Wouw M, Boehme M, Lyte JM et al. Short-chain fatty acids: microbial metabolites that alleviate stress-induced brain-gut axis alterations. The Journal of physiology. 2018. PMID 30066368. DOI: 10.1113/JP276431. Access: abstract_metadata. Study: PMID30066368.
Williams BB, Van Benschoten AH, Cimermancic P et al. Discovery and characterization of gut microbiota decarboxylases that can produce the neurotransmitter tryptamine. Cell host & microbe. 2014. PMID 25263219. DOI: 10.1016/j.chom.2014.09.001. Access: abstract_metadata. Study: PMID25263219.
Chahwan B, Kwan S, Isik A et al. Gut feelings: A randomised, triple-blind, placebo-controlled trial of probiotics for depressive symptoms. Journal of affective disorders. 2019. PMID 31078831. DOI: 10.1016/j.jad.2019.04.097. Access: abstract_metadata. Study: PMID31078831.
Tomasik J, Yolken RH, Bahn S et al. Immunomodulatory Effects of Probiotic Supplementation in Schizophrenia Patients: A Randomized, Placebo-Controlled Trial. Biomarker insights. 2015. PMID 26052224. DOI: 10.4137/BMI.S22007. Access: abstract_metadata. Study: NCT01242371.
Severance EG, Gressitt KL, Stallings CR et al. Probiotic normalization of Candida albicans in schizophrenia: A randomized, placebo-controlled, longitudinal pilot study. Brain, behavior, and immunity. 2017. PMID 27871802. DOI: 10.1016/j.bbi.2016.11.019. Access: abstract_metadata. Study: NCT01242371.
Dickerson FB, Stallings C, Origoni A et al. Effect of probiotic supplementation on schizophrenia symptoms and association with gastrointestinal functioning: a randomized, placebo-controlled trial. The primary care companion for CNS disorders. 2014. PMID 24940526. DOI: 10.4088/PCC.13m01579. Access: abstract_metadata. Study: NCT01242371.
Romijn AR, Rucklidge JJ, Kuijer RG et al. A double-blind, randomized, placebo-controlled trial of Lactobacillus helveticus and Bifidobacterium longum for the symptoms of depression. The Australian and New Zealand journal of psychiatry. 2017. PMID 28068788. DOI: 10.1177/0004867416686694. Access: abstract_metadata. Study: PMID28068788.
Fan Y, Gao Y, Ma Q et al. Multi-Omics Analysis Reveals Aberrant Gut-Metabolome-Immune Network in Schizophrenia. Frontiers in immunology. 2022. PMID 35309369. DOI: 10.3389/fimmu.2022.812293. Access: abstract_metadata. Study: PMID35309369.
Wang Z, Yuan X, Zhu Z et al. Multiomics Analyses Reveal Microbiome-Gut-Brain Crosstalk Centered on Aberrant Gamma-Aminobutyric Acid and Tryptophan Metabolism in Drug-Naïve Patients with First-Episode Schizophrenia. Schizophrenia bulletin. 2024. PMID 37119525. DOI: 10.1093/schbul/sbad026. Access: abstract_metadata. Study: PMID37119525.
Ma X, Asif H, Dai L et al. Alteration of the gut microbiome in first-episode drug-naïve and chronic medicated schizophrenia correlate with regional brain volumes. Journal of psychiatric research. 2020. PMID 32065949. DOI: 10.1016/j.jpsychires.2020.02.005. Access: abstract_metadata. Study: PMID32065949.
Zhu F, Ju Y, Wang W et al. Metagenome-wide association of gut microbiome features for schizophrenia. Nature communications. 2020. PMID 32235826. DOI: 10.1038/s41467-020-15457-9. Access: abstract_metadata. Study: PMID32235826.
Kreuzer K, Reiter A, Birkl-Töglhofer AM et al. The PROVIT Study-Effects of Multispecies Probiotic Add-on Treatment on Metabolomics in Major Depressive Disorder-A Randomized, Placebo-Controlled Trial. Metabolites. 2022. PMID 36005642. DOI: 10.3390/metabo12080770. Access: abstract_metadata. Study: NCT03300440.
Maier L, Pruteanu M, Kuhn M et al. Extensive impact of non-antibiotic drugs on human gut bacteria. Nature. 2018. PMID 29555994. DOI: 10.1038/nature25979. Access: abstract_metadata. Study: PMID29555994.
Sudo N, Chida Y, Aiba Y et al. Postnatal microbial colonization programs the hypothalamic-pituitary-adrenal system for stress response in mice. The Journal of physiology. 2004. PMID 15133062. DOI: 10.1113/jphysiol.2004.063388. Access: abstract_metadata. Study: PMID15133062.
Mayneris-Perxachs J, Castells-Nobau A, Arnoriaga-Rodríguez M et al. Microbiota alterations in proline metabolism impact depression. Cell metabolism. 2022. PMID 35508109. DOI: 10.1016/j.cmet.2022.04.001. Access: abstract_metadata. Study: PMID35508109.
Yap CX, Henders AK, Alvares GA et al. Autism-related dietary preferences mediate autism-gut microbiome associations. Cell. 2021. PMID 34767757. DOI: 10.1016/j.cell.2021.10.015. Access: abstract_metadata. Study: PMID34767757.
Dilmore AH, Kuplicki R, McDonald D et al. Medication use is associated with distinct microbial features in anxiety and depression. Molecular psychiatry. 2025. PMID 39794490. DOI: 10.1038/s41380-024-02857-2. Access: full_text_partial. Study: PMID39794490.
Brushett S, Gacesa R, Vich Vila A et al. Gut feelings: the relations between depression, anxiety, psychotropic drugs and the gut microbiome. Gut microbes. 2023. PMID 38017662. DOI: 10.1080/19490976.2023.2281360. Access: full_text_partial. Study: PMID38017662.

Source-specific evidence cards

[1] Transferring the blues: Depression-associated gut microbiota induces neurobehavioural changes in the rat.

Shows: Depression-donor transfer induced anhedonia and anxiety-like behaviour in microbiota-depleted rats, with tryptophan changes.

Does not show: Crude transfer and recipient depletion do not identify a microbial mediator or show initiation in the human donors.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC2.

[2] Gut microbiome remodeling induces depressive-like behaviors through a pathway mediated by the host's metabolism.

Shows: MDD microbiota transferred behavioural and metabolic differences to germ-free recipients.

Does not show: Forced-swim immobility is a proxy, and metabolic covariation does not establish the required human mediator.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC2.

Shows: Cohousing, transfer and defined commensal reconstitution altered social behaviour and synaptic plasticity after maternal high-fat diet.

Does not show: This is a neurodevelopmental model outside the registered human depression, anxiety and psychosis scope.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2.

[4] Involvement of the microbiota-gut-brain axis in chronic restraint stress: disturbances of the kynurenine metabolic pathway in both the gut and brain.

Shows: IDO inhibition and microbiota interventions partially rescued stress-associated behaviour and kynurenine abnormalities.

Does not show: Stress is the initiating exposure and the interventions do not isolate one exclusively microbial pathway.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2.

[5] Gut dysbiosis induces the development of depression-like behavior through abnormal synapse pruning in microglia-mediated by complement C3.

Shows: Stress-donor microbiota transferred behaviour and complement-associated synaptic abnormalities to recipients.

Does not show: C3/CR3 changes accompany the phenotype; the abstract does not independently resolve every mediator necessity contrast.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2.

[6] Changes in Gut Microbiota by Chronic Stress Impair the Efficacy of Fluoxetine.

Shows: Transferred microbiota impaired fluoxetine response and neurogenesis; 5-hydroxytryptophan alleviated the deficits.

Does not show: A downstream precursor rescue is not proof that the same pathway initiates human MDD.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2.

[7] Gut microbiota changes require vagus nerve integrity to promote depressive-like behaviors in mice.

Shows: Vagotomy before stress-donor inoculation abrogated behavioural, neurogenic and inflammatory effects; Methods describe blinded analysis.

Does not show: Subdiaphragmatic vagotomy is physiologically broad; effects and surgical boundaries are within the recipient model.

Access/status: Complete primary abstract and linked-status metadata examined. Selected body sections: 1:Materials and methods:paragraph3, 1:Materials and methods:paragraph4, 1:Materials and methods:paragraph13, 1:Materials and methods:paragraph18, 2:Results:paragraph9. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2.

[8] Regulation of prefrontal cortex myelination by the microbiota.

Shows: Post-weaning colonization reversed myelin-related gene changes in germ-free prefrontal cortex.

Does not show: Molecular and ultrastructural regulation is not a psychiatric clinical endpoint.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2.

[9] The microbiota regulate neuronal function and fear extinction learning.

Shows: Microbial depletion impaired fear extinction, with developmental-window dependence of restoration and altered neuronal activity.

Does not show: Fear learning is a defined behavioural function; identified metabolite differences alone do not establish human anxiety disease causation.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2.

[10] Effect of gut microbiota on depressive-like behaviors in mice is mediated by the endocannabinoid system.

Shows: Central eCB enhancement rescued transferred behaviours, central CB1 antagonism blocked rescue, and Lactobacillus complementation restored behavioural and lipid signals.

Does not show: Some individual contrasts were nonsignificant; downstream rescue narrows this model rather than establishing all human depression causes.

Access/status: Complete primary abstract and linked-status metadata examined. Selected body sections: 1:Results:paragraph9, 1:Results:paragraph16, 2:Discussion:paragraph4. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2.

[11] Role of Tryptophan in Microbiota-Induced Depressive-Like Behavior: Evidence From Tryptophan Depletion Study.

Shows: Acute tryptophan depletion changed the behavioural phenotype and brain serotonin of germ-free mice more strongly than SPF mice.

Does not show: Depletion is a broad nutritional challenge and germ-free development differs from adult human dysbiosis.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2.

[12] Roseburia hominis Alleviates Neuroinflammation via Short-Chain Fatty Acids through Histone Deacetylase Inhibition.

Shows: Roseburia hominis reduced microglial activation and depressive behaviours; SCFA treatments reproduced molecular anti-inflammatory effects.

Does not show: SCFA molecular rescue is not a demonstrated exclusive mediator of the behavioural effect in humans.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2.

[13] Clinical, gut microbial and neural effects of a probiotic add-on therapy in depressed patients: a randomized controlled trial.

Shows: HAM-D time-by-group benefit occurred in the compliant mITT sample, while the paper-labelled ITT interaction was nonsignificant; Lactobacillus enrichment correlated with symptoms.

Does not show: The ITT table has 47 completers out of 60 included; adherence exclusions, baseline imbalance and attrition limit inference. Correlation is not randomized mediation.

Access/status: Complete primary abstract and linked-status metadata examined. Selected body sections: 1:Patients and methods:paragraph3, 2:Results:paragraph12, 2:Results:paragraph14, 3:Discussion:paragraph1, 1:Patients and methods:paragraph0, 1:Patients and methods:paragraph6, 2:Results:paragraph0, 2:Results:paragraph2, 2:Results:paragraph3, 2:Results:paragraph4, 2:Results:paragraph5, 2:Results:paragraph6. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC3.

[14] Acceptability, Tolerability, and Estimates of Putative Treatment Effects of Probiotics as Adjunctive Treatment in Patients With Depression: A Randomized Clinical Trial.

Shows: A masked pilot add-on probiotic trial estimated favourable effects on selected depression and anxiety measures with low attrition.

Does not show: Feasibility and effect-size estimation preceded a definitive efficacy trial; this report does not establish microbial mediator necessity or disease modification.

Access/status: Complete primary abstract and linked-status metadata examined. Selected body sections: 1:Methods:paragraph0, 1:Methods:paragraph1, 1:Methods:paragraph2, 1:Methods:paragraph3, 1:Methods:paragraph4. No complete-paper or supplement audit claimed. none_found. Subclaims: SC3.

[15] PROVIT: Supplementary Probiotic Treatment and Vitamin B7 in Depression-A Randomized Controlled Trial.

Shows: PROVIT changed microbial profiles but clinical outcome measures did not differ from placebo plus biotin.

Does not show: Inpatient treatment improved both arms; microbiome change cannot be substituted for a positive psychiatric endpoint.

Access/status: Complete primary abstract and linked-status metadata examined. Selected body sections: 2:3. Results:paragraph5, 2:3. Results:paragraph6, 2:3. Results:paragraph7, 2:3. Results:paragraph10, 2:3. Results:paragraph11, 2:3. Results:paragraph12. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC3.

[16] Bifidobacterium breve CCFM1025 attenuates major depression disorder via regulating gut microbiome and tryptophan metabolism: A randomized clinical trial.

Shows: Randomized CCFM1025 supplementation improved HDRS and MADRS change relative to placebo and altered serotonin-turnover measures.

Does not show: Small short-term trial; peripheral turnover and parallel microbial changes do not isolate the cause of benefit or prove durability.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC3.

[17] Effect of probiotic and prebiotic vs placebo on psychological outcomes in patients with major depressive disorder: A randomized clinical trial.

Shows: Probiotic supplementation improved BDI relative to placebo and prebiotic in MDD; prebiotic benefit and unadjusted mediator ratios were null.

Does not show: Attrition was material and adjusted metabolite associations do not establish a necessary microbial mediator.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC3.

[18] Probiotic Bifidobacterium longum NCC3001 Reduces Depression Scores and Alters Brain Activity: A Pilot Study in Patients With Irritable Bowel Syndrome.

Shows: NCC3001 reduced depression scores and altered emotional brain responses in IBS, with no significant anxiety or IBS-symptom benefit.

Does not show: IBS with mood scores is distinct from primary MDD; fecal community profiles were similar and mediator attribution remains unresolved.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC3.

Shows: Adjunctive randomized FMT improved HAMD-17 reductions at weeks two and eight, without a significant week-eight remission difference; the abstract reports bile-acid mediation analysis.

Does not show: Only abstract accessible; primary endpoint hierarchy, registration timing, missing-data treatment and mediation assumptions cannot be confirmed. Statistical mediation is not mediator manipulation.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC3.

[20] Feasibility, Acceptability, and Safety of Faecal Microbiota Transplantation in the Treatment of Major Depressive Disorder: A Pilot Randomized Controlled Trial.

Shows: A pilot MDD FMT trial met feasibility and acceptability targets and reported successful masking checks.

Does not show: The trial was not designed to establish clinical efficacy; exploratory quality-of-life and gastrointestinal outcomes do not establish MDD causation.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC3.

Shows: A secondary analysis found immediate verbal recall and hippocampal activation differences; other cognitive measures were null.

Does not show: Same NCT02957591 cohort as the HAM-D report; low-adherence exclusions and multiple outcomes limit independent causal weight.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC3.

[22] Effect of Lactobacillus rhamnosus HN001 in Pregnancy on Postpartum Symptoms of Depression and Anxiety: A Randomised Double-blind Placebo-controlled Trial.

Shows: HN001 allocation lowered postpartum depression and anxiety screening scores.

Does not show: Mood was a secondary outcome of an offspring-eczema trial; screening scores and postpartum prevention do not establish the full psychiatric mechanism.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC3.

Shows: An MDD randomized probiotic trial reported greater BDI improvement with changes in insulin, hs-CRP and glutathione.

Does not show: Multiple biomarkers and short follow-up do not identify a microbial mediator or sustained disease modification.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC3.

Shows: An eight-week PS128 MDD trial found no between-group depression advantage or significant inflammatory, permeability or community changes.

Does not show: Small trial and one formulation do not refute every microbiota pathway; within-arm improvement is not treatment efficacy.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC3.

Shows: LP299v augmentation improved selected cognitive tests and reduced kynurenine among analysed completers.

Does not show: The abstract does not demonstrate depression-score superiority; cognition, kynurenine and mood must remain separate endpoints.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC3.

[26] Exploring the mechanisms of action of probiotics in depression: Results from a randomized controlled pilot trial.

Shows: The Nikolova cohort follow-up found transient between-group microbial differences and anxiety-associated taxa, with null inflammatory differences.

Does not show: Same NCT03893162 trial as the clinical pilot; within-group richness changes were not superior to placebo and do not prove mediation.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC3.

Shows: Exploratory metabolomics after LP299v showed changes in acylcarnitines and other metabolites.

Does not show: Same underlying trial as the earlier cognition report; altered metabolite panels are neither independent efficacy replication nor proven causal mediators.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC3.

[28] The neuroactive potential of the human gut microbiota in quality of life and depression.

Shows: Population and validation datasets linked butyrate-producing taxa to quality of life and depression after antidepressant adjustment.

Does not show: Replicated association does not establish direction; predicted neuroactive capacity is not measured brain delivery.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1.

[29] Microbiome and tryptophan metabolomics analysis in adolescent depression: roles of the gut microbiota in the regulation of tryptophan-derived neurotransmitters and behaviors in human and mice.

Shows: Roseburia intestinalis rescued stress-related mouse behaviour and tryptophan metabolism alongside adolescent microbial associations.

Does not show: Human first diagnosis and post-sertraline profiles are observational; mouse rescue is not a randomized adolescent treatment result.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC2.

[30] GABA-modulating bacteria of the human gut microbiota.

Shows: Cultured bacteria produced or used GABA, and human Bacteroides abundance correlated with depression-related brain signatures.

Does not show: Bacterial GABA capacity does not show adequate brain exposure, human mediator necessity or clinical benefit.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC2.

[31] Landscapes of bacterial and metabolic signatures and their interaction in major depressive disorders.

Shows: MDD samples differed in bacteria, virome and fecal metabolites and yielded a discriminatory marker panel.

Does not show: Classifier performance and coexpression cannot determine initiation or temporal direction.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1.

[32] Altered fecal microbiota composition in patients with major depressive disorder.

Shows: Active MDD had higher alpha diversity and altered taxa, including lower Faecalibacterium linked to severity.

Does not show: Direction of diversity differs from reduced-diversity reports; treatment-response groups and diagnosis do not isolate microbial causation.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1.

[33] Interplay of Metabolome and Gut Microbiome in Individuals With Major Depressive Disorder vs Control Individuals.

Shows: MDD had energy and lipid metabolic signatures; MR suggested some lipid changes associated with the disease process.

Does not show: The instruments address metabolites rather than a defined microbial mediator; coarse cross-study microbiome links cannot validate dysbiosis-wide causation.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC3.

[34] Unraveling the Microbial Mechanisms Underlying the Psychobiotic Potential of a Bifidobacterium breve Strain.

Shows: CCFM1025 but not a comparator B. breve strain improved stress behaviour and produced distinct microbial/metabolic profiles.

Does not show: Strain comparisons narrow function but genomic predictions and parallel metabolites do not identify every mediator.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2.

[35] Altered metabolome and microbiome features provide clues in understanding irritable bowel syndrome and depression comorbidity.

Shows: IBS discovery/validation cohorts linked tryptophan and serotonin metabolites to depression comorbidity.

Does not show: IBS, diet and transit can influence both mood and stool; this is not primary psychiatric disease randomization.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1.

[36] The change of gut microbiota in MDD patients under SSRIs treatment.

Shows: Microbial composition and diversity shifted after symptom improvement under escitalopram.

Does not show: Follow-up was conditioned on improvement; drug exposure, recovery and time are inseparable without an appropriate comparator.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1.

[37] Multi-Probiotics ameliorate Major depressive disorder and accompanying gastrointestinal syndromes via serotonergic system regulation.

Shows: A three-strain formula improved depression and GI scores relative to placebo; a separate stress model investigated serotonergic effects.

Does not show: The abstract does not supply complete human design or mediator-subtraction details; mouse mechanisms cannot be substituted for human mediation.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2, SC3.

[38] Microbiota modulate behavioral and physiological abnormalities associated with neurodevelopmental disorders.

Shows: B. fragilis ameliorated selected behaviours and barrier abnormalities; administration of an elevated metabolite induced behavioural abnormalities.

Does not show: Antecedent maternal immune activation and neurodevelopmental outcomes are outside the registered human SC3 scope.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2.

[39] Ingestion of Lactobacillus strain regulates emotional behavior and central GABA receptor expression in a mouse via the vagus nerve.

Shows: JB-1 changed stress behaviour and regional GABA receptor expression; effects were absent after vagotomy.

Does not show: Healthy-mouse stress responses and a broad nerve lesion do not prove human psychiatric onset or generalize to all strains.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2.

[40] Bacteroides species differentially modulate depression-like behavior via gut-brain metabolic signaling.

Shows: Specific Bacteroides species recapitulated behaviour and neurogenesis effects of MDD-donor transfer, while B. ovatus did not.

Does not show: Species boundaries are informative; depletion protocols and recipient proxies limit endogenous human inference.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2.

[41] A gut-derived metabolite alters brain activity and anxiety behaviour in mice.

Shows: Defined 4EP biosynthesis increased 4EPS exposure, altered myelination and anxiety-like behaviour; oligodendrocyte-directed treatment prevented behaviour.

Does not show: Engineered exposure establishes a molecule pathway in mice, not its sufficient concentration or necessity in human anxiety disorders.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2.

[42] Stress-sensitive neural circuits change the gut microbiome via duodenal glands.

Shows: Neural activation and gland ablation causally linked psychological-state circuits to Lactobacillus and host defence.

Does not show: This directly supports brain-to-gut direction in a mouse pathway; it does not prove that every human microbial association is reactive.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2.

[43] Oral selective serotonin reuptake inhibitors activate vagus nerve dependent gut-brain signalling.

Shows: Oral SSRI activated gut neural signalling and its tail-suspension effect required intact vagal signalling.

Does not show: Vagal dependence is not specific to microbial mechanisms and does not prove microbes mediate the SSRI benefit.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2.

Shows: L. reuteri rescued social deficits across models through vagus and oxytocin-receptor dependence, without restoring overall community composition.

Does not show: Strong bounded mechanism outside human SC3; therapeutic action without composition restoration challenges a generic normalization narrative.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2.

[45] Short-chain fatty acids: microbial metabolites that alleviate stress-induced brain-gut axis alterations.

Shows: SCFA supplementation alleviated selected anhedonia, stress response and permeability changes while weight and some receptor outcomes remained unaffected.

Does not show: Supplementation does not establish endogenous deficiency as the initiating cause; effects were behavioural-test and stress-context specific.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2.

[46] Discovery and characterization of gut microbiota decarboxylases that can produce the neurotransmitter tryptamine.

Shows: Two microbial decarboxylases produced tryptamine, establishing a concrete biochemical capacity.

Does not show: No psychiatric phenotype, in vivo brain exposure or clinical mediation was demonstrated.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2.

Shows: A triple-blind trial found nonspecific depressive-symptom improvement, a cognitive-reactivity signal and no significant microbiota change.

Does not show: High attrition and subgroup findings constrain inference; vulnerability measures are not a positive primary depression treatment result.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC3.

[48] Immunomodulatory Effects of Probiotic Supplementation in Schizophrenia Patients: A Randomized, Placebo-Controlled Trial.

Shows: Schizophrenia probiotic supplementation altered vWF and selected immune proteins, with several only borderline signals.

Does not show: Same NCT01242371 program as the symptom trial; immune markers do not establish a positive PANSS endpoint or mediator necessity.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC3.

[49] Probiotic normalization of Candida albicans in schizophrenia: A randomized, placebo-controlled, longitudinal pilot study.

Shows: Probiotic allocation reduced Candida antibodies in males, with sex and serostatus dependent bowel and psychiatric trends.

Does not show: Same schizophrenia trial program; trends and subgroup antibody changes cannot replace the null total PANSS result.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC3.

[50] Effect of probiotic supplementation on schizophrenia symptoms and association with gastrointestinal functioning: a randomized, placebo-controlled trial.

Shows: A schizophrenia trial found no significant PANSS total benefit but less severe bowel difficulty.

Does not show: Somatic benefit does not establish psychosis disease modification; one trial does not exhaust strain-specific possibilities.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC3.

Shows: An ITT low-mood trial found no probiotic advantage on psychological measures or blood biomarkers.

Does not show: Moderate low mood rather than confirmed MDD constrains scope; suggested responder populations remain untested.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC3.

[52] Multi-Omics Analysis Reveals Aberrant Gut-Metabolome-Immune Network in Schizophrenia.

Shows: Multiomics linked bacteria, metabolites and cytokines and distinguished schizophrenia cases from controls.

Does not show: High AUC does not establish causal direction and can reflect shared confounding or overfitting.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1.

[53] Multiomics Analyses Reveal Microbiome-Gut-Brain Crosstalk Centered on Aberrant Gamma-Aminobutyric Acid and Tryptophan Metabolism in Drug-Naïve Patients with First-Episode Schizophrenia.

Shows: Integrated microbiome, metabolome and imaging data associated neurotransmitter pathways with symptoms and genetic risk.

Does not show: Observational mediation is not randomized microbial causation. Linked erratum PMID37232499 is retained; affected numerical claims were not used.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. ErratumIn: PMID37232499. Subclaims: SC1.

[54] Alteration of the gut microbiome in first-episode drug-naïve and chronic medicated schizophrenia correlate with regional brain volumes.

Shows: Alpha diversity was lower in chronically medicated but not first-episode drug-naive schizophrenia; both groups showed taxonomic differences.

Does not show: Medication, illness duration and severity are not independently randomized; brain correlations do not establish a microbial initiator.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1.

[55] Metagenome-wide association of gut microbiome features for schizophrenia.

Shows: A metagenomic classifier replicated, and schizophrenia-enriched S. vestibularis induced mouse social and neurotransmitter changes.

Does not show: Human direction remains observational and mouse social deficits are not the full schizophrenia syndrome.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC2.

[56] The PROVIT Study-Effects of Multispecies Probiotic Add-on Treatment on Metabolomics in Major Depressive Disorder-A Randomized, Placebo-Controlled Trial.

Shows: PROVIT shifted several stool metabolites, including butyrate, with no significant serum metabolomic changes.

Does not show: Same NCT03300440 trial as the null clinical report; stool target engagement does not establish antidepressant efficacy.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC3.

[57] Extensive impact of non-antibiotic drugs on human gut bacteria.

Shows: Many human-targeted drugs inhibited commensal strains, with antipsychotics overrepresented.

Does not show: In vitro susceptibility validates a plausible medication confound, not the size or direction of every in vivo psychiatric association.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1.

[58] Postnatal microbial colonization programs the hypothalamic-pituitary-adrenal system for stress response in mice.

Shows: Early colonization and defined organisms altered restraint-induced HPA responses, with timing and bacterial-mutant boundaries.

Does not show: HPA responses are physiological endpoints, not diagnosis; late failure maps development rather than adult treatment inevitability.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC2.

[59] Microbiota alterations in proline metabolism impact depression.

Shows: Proline-linked human signals were complemented by dietary, transfer, transporter and microbial perturbations in mice and flies.

Does not show: Cross-system convergence supports a bounded pathway; no randomized human proline/microbial prevention or treatment mechanism was tested.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1, SC2.

Shows: A large metagenomic study supported restricted diet and stool consistency as mediators of ASD microbiome associations, with negligible direct diagnostic associations.

Does not show: Outside registered depression/anxiety/psychosis; informative confounding analogue, not direct refutation of those diseases. Erratum PMID38242089 retained.

Access/status: Complete primary abstract and linked-status metadata examined. No complete-paper or supplement audit claimed. ErratumIn: PMID38242089. Subclaims: SC1.

[61] Medication use is associated with distinct microbial features in anxiety and depression.

Shows: Antidepressant exposure explained more microbial beta-diversity variation than diagnostic status in the studied cohorts.

Does not show: Cross-sectional medication assignment and indication remain confounded; this does not erase diagnosis-associated taxa in other datasets.

Access/status: Complete primary abstract and linked-status metadata examined. Selected body sections: 3:Discussion:paragraph7, 3:Discussion:paragraph10. No complete-paper or supplement audit claimed. UpdateOf: PMID38562901. Subclaims: SC1.

[62] Gut feelings: the relations between depression, anxiety, psychotropic drugs and the gut microbiome.

Shows: Most internalizing-disorder microbial associations survived psychotropic and further GI/diet adjustments, while alpha diversity was not significantly associated.

Does not show: Adjustment does not establish temporality or eliminate residual confounding; the result counters a medication-only explanation without proving a microbial initiator.

Access/status: Complete primary abstract and linked-status metadata examined. Selected body sections: 1:Results:paragraph0, 1:Results:paragraph1, 1:Results:paragraph2, 1:Results:paragraph3. No complete-paper or supplement audit claimed. none_found. Subclaims: SC1.

HSEF level	Β — Associational
Confidence	Moderate
Role of mechanism	reinforcing / stabilizing factor
Calibration role of this review	Not a calibration entry; the separate five-case Gemini/Abacus review and PKU calibration record are documented in the framework record
Schema	5.3
Posted	2026-10-06

ID	Sub-claim	Type	Level	Pivotal card
SC1	Gut microbial composition correlates reproducibly with specific human psychiatric states.	existence	Β	Card ↗
SC2	Defined microbial manipulation changes behavioral phenotypes through identified mediators.	mechanism	Ε	Card ↗
SC3	These microbiota-to-brain mechanisms make a clinically meaningful causal contribution to human psychiatric disease.	generalization	Γ	Card ↗

Gut–Brain Axis Drives Psychiatric Disease

In plain terms

What decisive evidence would look like

What this review means for Gut–Brain Axis Drives Psychiatric Disease

Which explanation best fits the evidence reviewed?

How does this explanation compare with alternatives?

Bounded microbial amplification

Stress-to-brain and stress-to-gut feedback

General upstream microbial psychiatric cause

Review summary

Key points

Verdict and review details

At a glance

Sub-claims assessed separately

Pivotal subclaim illustrations

Figures

Full assessment

AT A GLANCE

COMPETING MECHANISTIC MODELS

FULL REVIEW

Registered claim and scope

Evidence and causal interpretation

HYPOTHESIS STATE EVALUATION FRAMEWORK (HSEF v5.3)

Subclaim matrix and aggregation

Decisive-perturbation inventory and causal tally

Cap audit, precedence and Β/Γ/Δ boundary

Upward threshold ladder

Falsification and pitfalls

Adversarial steelman: above, then below

Final score, causal direction and M/S/D

PROPOSED DECISIVE EXPERIMENTS

RESEARCH TIMELINE

KNOWLEDGE GAPS

METHODS AND CONSENSUS CONTRIBUTION

REVIEWER NOTES

ADVERSARIAL REVIEW AND APPROVAL

REFERENCES

Source-specific evidence cards

[1] Transferring the blues: Depression-associated gut microbiota induces neurobehavioural changes in the rat.

[2] Gut microbiome remodeling induces depressive-like behaviors through a pathway mediated by the host's metabolism.

[3] Microbial Reconstitution Reverses Maternal Diet-Induced Social and Synaptic Deficits in Offspring.

[4] Involvement of the microbiota-gut-brain axis in chronic restraint stress: disturbances of the kynurenine metabolic pathway in both the gut and brain.

[5] Gut dysbiosis induces the development of depression-like behavior through abnormal synapse pruning in microglia-mediated by complement C3.

[6] Changes in Gut Microbiota by Chronic Stress Impair the Efficacy of Fluoxetine.

[7] Gut microbiota changes require vagus nerve integrity to promote depressive-like behaviors in mice.

[8] Regulation of prefrontal cortex myelination by the microbiota.

[9] The microbiota regulate neuronal function and fear extinction learning.

[10] Effect of gut microbiota on depressive-like behaviors in mice is mediated by the endocannabinoid system.

[11] Role of Tryptophan in Microbiota-Induced Depressive-Like Behavior: Evidence From Tryptophan Depletion Study.

[12] Roseburia hominis Alleviates Neuroinflammation via Short-Chain Fatty Acids through Histone Deacetylase Inhibition.

[13] Clinical, gut microbial and neural effects of a probiotic add-on therapy in depressed patients: a randomized controlled trial.

[14] Acceptability, Tolerability, and Estimates of Putative Treatment Effects of Probiotics as Adjunctive Treatment in Patients With Depression: A Randomized Clinical Trial.

[15] PROVIT: Supplementary Probiotic Treatment and Vitamin B7 in Depression-A Randomized Controlled Trial.

[16] Bifidobacterium breve CCFM1025 attenuates major depression disorder via regulating gut microbiome and tryptophan metabolism: A randomized clinical trial.

[17] Effect of probiotic and prebiotic vs placebo on psychological outcomes in patients with major depressive disorder: A randomized clinical trial.

[18] Probiotic Bifidobacterium longum NCC3001 Reduces Depression Scores and Alters Brain Activity: A Pilot Study in Patients With Irritable Bowel Syndrome.

[19] Adjunctive fecal microbiota transplantation for major depressive disorder: A randomized, double-blind, placebo-controlled trial.

[20] Feasibility, Acceptability, and Safety of Faecal Microbiota Transplantation in the Treatment of Major Depressive Disorder: A Pilot Randomized Controlled Trial.

[21] Effect of short-term, high-dose probiotic supplementation on cognition, related brain functions and BDNF in patients with depression: a secondary analysis of a randomized controlled trial.

[22] Effect of Lactobacillus rhamnosus HN001 in Pregnancy on Postpartum Symptoms of Depression and Anxiety: A Randomised Double-blind Placebo-controlled Trial.

[23] Clinical and metabolic response to probiotic administration in patients with major depressive disorder: A randomized, double-blind, placebo-controlled trial.

[24] The effects of Lactobacillus plantarum PS128 in patients with major depressive disorder: an eight-week double-blind, placebo-controlled study.

[25] Probiotic Lactobacillus Plantarum 299v decreases kynurenine concentration and improves cognitive functions in patients with major depression: A double-blind, randomized, placebo controlled study.

[26] Exploring the mechanisms of action of probiotics in depression: Results from a randomized controlled pilot trial.

[27] Probiotic Lactobacillus plantarum 299v supplementation in patients with major depression in a double-blind, randomized, placebo-controlled trial: A metabolomics study.

[28] The neuroactive potential of the human gut microbiota in quality of life and depression.

[29] Microbiome and tryptophan metabolomics analysis in adolescent depression: roles of the gut microbiota in the regulation of tryptophan-derived neurotransmitters and behaviors in human and mice.

[30] GABA-modulating bacteria of the human gut microbiota.

[31] Landscapes of bacterial and metabolic signatures and their interaction in major depressive disorders.

[32] Altered fecal microbiota composition in patients with major depressive disorder.

[33] Interplay of Metabolome and Gut Microbiome in Individuals With Major Depressive Disorder vs Control Individuals.

[34] Unraveling the Microbial Mechanisms Underlying the Psychobiotic Potential of a Bifidobacterium breve Strain.

[35] Altered metabolome and microbiome features provide clues in understanding irritable bowel syndrome and depression comorbidity.

[36] The change of gut microbiota in MDD patients under SSRIs treatment.

[37] Multi-Probiotics ameliorate Major depressive disorder and accompanying gastrointestinal syndromes via serotonergic system regulation.

[38] Microbiota modulate behavioral and physiological abnormalities associated with neurodevelopmental disorders.

[39] Ingestion of Lactobacillus strain regulates emotional behavior and central GABA receptor expression in a mouse via the vagus nerve.

[40] Bacteroides species differentially modulate depression-like behavior via gut-brain metabolic signaling.

[41] A gut-derived metabolite alters brain activity and anxiety behaviour in mice.

[42] Stress-sensitive neural circuits change the gut microbiome via duodenal glands.