Abstract
Across multiple psychiatric conditions, converging evidence links gut microbiome dysbiosis to mental illness via immune activation, barrier dysfunction, neuroendocrine stress signaling, vagal afferents, and microbially derived metabolites (notably short-chain fatty acids and tryptophan–kynurenine metabolites).[1–4] In major depressive disorder (MDD), a recurring signal is reduced abundance of butyrate-producing bacteria and SCFA deficiency, alongside endotoxin-linked immune activation and HPA-axis involvement.[5–8] In anxiety disorders, evidence converges more strongly on stress-induced permeability and inflammatory–metabolite pathways than on a single taxonomic signature, with heterogeneous probiotic trial outcomes and limited FMT clinical evidence.[9, 10] In schizophrenia and first-episode psychosis (FEP), microbial diversity reductions and symptom-severity correlations are frequently described, but taxa-level findings are variable; translational FMT work supports effects on glutamate–glutamine–GABA metabolism and behavior.[11, 12] In autism spectrum disorder (ASD), reviews repeatedly emphasize dysbiosis with barrier dysfunction (“leaky gut”), cytokine-mediated BBB effects, and altered SCFAs and kynurenine-pathway metabolites, with small trials/open-label studies suggesting that FMT/MTT and some probiotics may improve GI symptoms and some behavioral outcomes.[13–17] Bipolar disorder studies repeatedly report decreased Faecalibacterium and suggest links between microbiome state and mood severity, with early evidence for probiotic-associated symptom improvements and FMT-based translational models affecting anxiety-like behavior and sociability.[18–21]
Based on these syntheses, we propose seven falsifiable hypotheses spanning SCFA/butyrate function, immune-driven kynurenine bias, vagal dependence, disorder-specific microbial functional modules (e.g., aspartate degradation in social anxiety), and circadian–microbiome interactions.[5, 6, 22–25]
Introduction
The microbiota–gut–brain axis (MGBA) denotes bidirectional communication between the gut ecosystem and the central nervous system via integrated neural (ENS/ANS), endocrine (HPA axis), immune, and metabolic pathways.[1, 2, 26] Mechanistically, gut-derived signals can reach the brain through vagal afferents that modulate limbic excitability and emotional processing, providing a direct neural route for microbial effects on behavior.[1, 25] In parallel, immune signaling couples peripheral cytokine and inflammatory mediator activity with BBB permeability and microglial activation, thereby extending gut signaling into the CNS inflammatory milieu relevant to multiple psychiatric phenotypes.[3, 4]
Microbial metabolites and precursor pathways provide an additional, transdiagnostic route from gut ecology to neurochemistry. SCFAs can influence serotonergic signaling in the gut and modulate vagal activity and serotonin transporter (SERT) expression, supporting a functional link between fermentation output and affective state regulation.[7, 27] Tryptophan metabolism connects microbiota to host serotonin and kynurenine pathways, especially under inflammatory conditions, offering a plausible bridge between immune activation and mood/psychosis-relevant neuroactive metabolites.[8, 9] Consistent with this multi-pathway model, dysbiosis patterns characterized by reduced diversity and reduced short-chain fatty-acid-producing bacteria have been described as predicting increased depression and anxiety symptom severity across studied cohorts and review-level summaries.[28]
Methods
This review followed a PRISMA-informed workflow using a structured multi-query discovery strategy across major psychiatric conditions and core mechanistic domains (e.g., SCFAs, tryptophan–kynurenine, HPA axis, vagal signaling, neuroinflammation, FMT, and psychobiotics).[1, 8, 12] The search strategy was implemented as ten academic queries with up to 50 results per query (targeting approximately 500 retrieved records), followed by a two-pass screening process prioritizing gut–brain axis focus, psychiatric relevance, human/translational evidence, and substantive study types; this produced the workflow counts summarized in the Results overview (500 retrieved; 448 after first screening; 281 high-quality sources; 105 full texts extracted).[3, 28]
Screening decisions were designed to address known sources of heterogeneity in this field, including methodological and population differences that can drive conflicting microbiome findings, and the need for standardization of sequencing approaches and biomarker definitions.[3, 7] Evidence synthesis prioritized (i) consistent functional themes (e.g., SCFA capacity, permeability/inflammation, kynurenine bias) and (ii) higher-quality human evidence (systematic reviews/meta-analyses, RCTs, and large observational cohorts) while retaining key translational animal findings that directly inform causal hypotheses.[28–30]
Core mechanisms of gut–brain communication
Multiple pathways plausibly mediate microbiome-to-brain effects in mental illness, and the reviewed sources emphasize that immune modulation, neural communication (particularly via the vagus nerve), microbial metabolite production, and neurotransmitter synthesis/metabolism act in combination rather than isolation.[1] Importantly, many studies do not test pathway function directly, so mechanistic inference often relies on convergent patterns across immunologic, endocrine, metabolomic, and behavioral endpoints.[29]
Mechanism overview
The table below summarizes core MGBA pathways and illustrates how they are supported across disorders represented in the extracted full texts.
Vagal signaling
The vagus nerve is repeatedly described as a critical gut–brain communication route, with vagal afferent pathways transmitting microbial signals from the intestine to limbic regions and modulating emotional processing.[1, 36] Translational evidence indicates that subdiaphragmatic vagotomy can abrogate microbiota-induced behavioral and neurogenesis effects, suggesting vagal integrity may be necessary for some microbiome-driven psychiatric phenotypes.[23] SCFAs also modulate vagal activity and SERT expression, linking fermentation output to neurochemical signaling pathways that plausibly scale to anxiety and depression outcomes.[27]
Immune activation and neuroinflammation
Across disorders, a recurring pattern is the coupling of dysbiosis with immune activation through increased permeability and microbial product recognition (e.g., LPS via toll-like receptors), which can drive pro-inflammatory factor secretion and systemic inflammation.[5, 25] Pro-inflammatory cytokines can migrate across the BBB and stimulate microglia-driven neuroinflammatory reactions, providing a mechanistic substrate for mood, anxiety, and neurodevelopmental symptoms.[4] In ASD-focused syntheses, IL-6 and TNF-α are specifically described as mediators that can compromise BBB integrity and interfere with neural signaling associated with behavioral symptoms.[14]
Endocrine stress signaling
The gut microbiome is implicated in regulating HPA-axis responsiveness, with germ-free mice exhibiting heightened ACTH and corticosterone responses to restraint stress compared with microbiota-competent animals.[32] Reviews of depression emphasize activation of the HPA axis and cortisol effects on gut integrity and microbiota alongside endotoxin-driven immune signaling, supporting a bidirectional stress–gut–immune loop in MDD.[7, 8] More broadly, stress hormones can dissolve tight junctions and increase barrier permeability, potentially amplifying endotoxin translocation and inflammatory signaling relevant to anxiety and stress-related disorders.[10]
Microbial metabolites and neurotransmitter precursors
Multiple sources emphasize metabolite signaling as a central MGBA pathway, including SCFAs, tryptophan, and other intermediate products, alongside microbial capacity to produce neurotransmitters such as dopamine, norepinephrine, GABA, serotonin, and histamine.[10, 37] SCFAs can stimulate serotonin release in the gut and influence BBB integrity, providing a mechanistic bridge between fermentation capacity and central neurochemical regulation.[7, 34] Tryptophan metabolism is repeatedly highlighted as linking microbiota to serotonin and kynurenine pathways, particularly under inflammation, consistent with a model where immune activation drives neuroactive metabolite shifts relevant to depression and psychosis.[8, 9]
Intestinal and BBB permeability
Gut dysbiosis and inflammation can produce a “leaky gut” phenotype in which containment of gut contents (including gram-negative LPS) is reduced, eliciting systemic and central inflammatory responses while selecting for taxa that tolerate immune pressure.[38] In ASD-focused syntheses, increased gut permeability is explicitly implicated, with the “leaky gut” allowing bacterial metabolites to cross the intestinal barrier and enter systemic circulation as potentially neuroactive signals.[13] In parallel, immune–BBB coupling is described as dynamically regulated by cytokines and inflammatory mediators, reinforcing the plausibility that peripheral inflammation can alter CNS immune tone and behavior.[3]
Evidence by disorder
Across conditions, the most reproducible evidence tends to be functional (e.g., reduced butyrate-producing capacity, increased permeability/inflammation, kynurenine bias) rather than a single taxonomic “fingerprint,” and many sources explicitly report conflicting human findings due to confounding and methodological variability.[7, 12, 32]
Major depressive disorder
MDD syntheses consistently report dysbiosis patterns that often include increased Actinobacteria and sometimes Fusobacteria, alongside reduced abundance of butyrate-producing bacteria and taxa such as Faecalibacterium (with some studies also noting increased Eggerthella).[5, 7] However, diversity metrics do not show universal agreement, with reviews noting no consensus in alpha- and beta-diversity across studies, consistent with high cross-cohort heterogeneity.[7]
Mechanistically, MDD evidence repeatedly emphasizes endotoxin/LPS translocation through compromised gut barrier integrity and downstream immune activation, with associated HPA-axis involvement and cortisol-linked effects on gut integrity and microbiota.[7, 8] SCFAs are highlighted as relevant to serotonergic signaling (e.g., stimulating serotonin release in the gut) and as frequently deficient in depression, with supplementation described as capable of improving depressive symptoms.[6, 7] Tryptophan metabolism is also central, with inflammation-linked shifts toward the kynurenine branch described as leading to accumulation of proinflammatory and neurotoxic metabolites that exacerbate neuroinflammation in the brain.[8, 24]
Intervention evidence is strongest for probiotics as adjunctive therapy, with meta-analytic syntheses suggesting modest improvements in depressive symptoms, particularly when used alongside standard antidepressant therapy, and with some trials reporting biochemical changes such as reduced kynurenine/tryptophan ratio in the probiotic group.[8, 39] Translational causal support includes germ-free mouse experiments receiving FMT from MDD patients that induced depression-like phenotypes, and complementary evidence that transplantation of healthy microbiota reduces depressive- and anxiety-like behaviors in animal models.[30, 40] Despite these signals, medication confounding is substantial, with antidepressant exposure described as imposing diverse alterations that make bacterial-community patterns difficult to predict in medicated depressed patients.[7]
Anxiety disorders
Across the included anxiety evidence, dysbiosis patterns include reports of low butyrate-producing bacteria in individuals with severe anxiety symptoms and decreases in species such as Roseburia intestinalis and Bifidobacterium longum in anxiety-associated contexts, but consistent taxonomic signatures are not yet established across cohorts.[32, 41, 42] Experimental antibiotic perturbation in mice provides mechanistic support that gut disturbance can induce anxiety-like behavior alongside increased Proteobacteria (notably Klebsiella oxytoca), increased fecal and blood LPS, and decreased lactobacilli including Lactobacillus reuteri.[31]
Mechanistic accounts in anxiety disorders emphasize stress-induced increases in intestinal permeability and endotoxin movement leading to low-grade inflammation, alongside cytokine-driven activation of IDO/TDO and diversion of tryptophan metabolism toward the kynurenine pathway with reduced conversion to serotonin and downstream NAS/melatonin.[9] Additional causal-inference approaches include bidirectional Mendelian randomization analyses identifying bacterial genera with putative causal relations to anxiety disorders and suggesting metabolite-dependent mechanisms involving tryptophan, amino acids, and cortisol.[43]
Human intervention evidence is mixed: controlled probiotic trials often report no difference versus placebo, though some analyses report anxiety alleviation, and systematic sources conclude it is too early to translate microbiome modulation into routine anxiety-disorder treatment recommendations.[10] In the anxiety-focused review evidence provided here, no clinical studies examined fecal microbiota transplantation in anxiety disorders, leaving a clear gap for interventional causality testing beyond probiotics/prebiotics and diet.[10]
Generalized anxiety disorder
The GAD-specific evidence provided in this dataset is limited to a single cross-sectional convenience-sample study using self-administered GAD-7 scoring rather than psychiatrist-confirmed diagnosis, constraining inference and generalizability.[44] Within that study, the anxious group showed lower relative abundance of Faecalibacterium and Bifidobacterium (and lower Actinobacteria) and higher abundance of Clostridioides and Bacteroides compared with the non/low-anxiety group.[44] Proposed mechanisms include SCFA-related pathways and immune activation via barrier integrity changes, but the study did not measure SCFAs in blood and did not correct for multiple comparisons in all analyses, underscoring why the evidence strength is best considered inconclusive.[44]
Social anxiety disorder
SAD evidence is emerging and currently limited. A human shotgun metagenomics case-control study reported beta-diversity differences and identified genus-level enrichment of Anaeromassilibacillus and Gordonibacter in SAD, with Parasutterella (including Parasutterella excrementihominis) enriched in controls.[22] The same study reported higher abundance of a microbial functional module (“Aspartate Degradation I”) describing aspartate degradation capacity via aspartate aminotransferase (AspAT), raising the possibility that functional metagenomic signals may be more robust than taxa-only markers in SAD.[22]
Mechanistically, the SAD functional finding is explicitly linked to the tryptophan–kynurenine pathway, with kynurenic acid (KYNA) described as a neuroactive substance elevated by chronic stress and in psychiatric conditions including SAD.[22] Translational causality is supported by a human-to-mouse FMT study showing that mice receiving SAD microbiota had heightened sensitivity to social fear without effects on other tested behaviors, and this phenotype was coupled with immune and oxytocin-related changes (e.g., reduced IL-17A responses and reduced Oxt neurons in BNST).[45] Key limitations include small sample size, single time-point design, psychotropic medication exposure in two-thirds of SAD patients, and lack of symptom–microbiome associations after FDR correction in the human study.[22]
Schizophrenia
Schizophrenia studies frequently report reduced within-sample diversity and richness compared with healthy controls, and across studies, correlations between gut microbiome features and clinical measures are most consistently demonstrated for overall symptom severity and negative symptom severity.[11, 12] Nevertheless, taxa-level directionality is highly variable across cohorts, with reviews attributing variability to regional and methodological differences, while also highlighting that Lactobacilli elevation may be one of the more consistent findings across schizophrenia and increased-risk groups.[12, 35] Individual cohort findings include depletion of Faecalibacterium prausnitzii alongside enrichment of other taxa, illustrating the difficulty of generalizing single taxa as biomarkers without standardized pipelines and careful confound control.[46]
Proposed mechanisms include byproducts of bacterial metabolism crossing the BBB, increased gut permeability, and immune stimulation, together with tryptophan metabolism, HPA-axis effects, and vagal pathways, reflecting a multi-route model rather than a single causal chain.[12] Translational evidence supports neurochemical relevance: germ-free mice receiving schizophrenia microbiome FMT showed lower glutamate and higher glutamine and GABA in the hippocampus and displayed schizophrenia-relevant behaviors consistent with glutamatergic hypofunction models.[11] Metabolic and epigenetic pathways are also highlighted, including tyrosine synthesis modules (dopamine precursor pathways) associated with cognition and butyrate as an HDAC inhibitor that can modulate host epigenetics.[46, 47]
Intervention evidence is mixed. A randomized placebo-controlled probiotic trial (Lactobacillus rhamnosus plus Bifidobacterium lactis Bb12) did not change PANSS scores over 14 weeks, though bowel-movement difficulties improved, while another study reported that Bifidobacterium breve A-1 improved PANSS and anxiety/depression scores alongside changes in cytokines (including reduced TNF-α).[35] Major limitations include cross-sectional human designs and heavy medication confounding, including explicit concerns that microbiome effects observed in FMT models may reflect “medicated microbiome” rather than disease-state microbiome, and the lack of prospective studies needed for causal inference in humans.[11, 35]
First-episode psychosis
Across psychosis and FEP-related evidence, reviews emphasize significant differences in the gut microbiome between psychosis patients and controls, including reported alterations in beta diversity, while also acknowledging a lack of consistent findings regarding specific taxa across studies.[48] One synthesis reports correlations between Lachnospiraceae, Bacteroides spp., and Lactobacillus with symptom severity domains, and notes that FEP cohorts may show increases in Proteobacteria (genus level) and Lactobacillaceae (family level).[33, 48]
Mechanistically, psychosis reviews emphasize neuroimmune and neuroendocrine pathways involving the vagus nerve, with communication mediated by microbially derived molecules including SCFAs and tryptophan metabolites that may cross intestinal and blood–brain barriers.[48] Tryptophan metabolism is again central, with FMT-based “schizophrenic” mice showing increased Kyn–Kyna activity and reduced serotonin-branch activity, supporting a link between gut ecology and neuroactive metabolite bias.[48]
Intervention evidence remains preliminary. A randomized controlled trial reported that a probiotic supplement containing Lactobacilli and Bifidobacterium bifidum (with vitamin D) decreased CRP and improved general and total PANSS scores, though the active component driving benefit was unclear.[48] Reviews simultaneously caution that much MGBA research remains animal-model based and that extrapolation from rodent FMT studies can overstate the role of the microbiome in human disease, reinforcing the need for larger, better-controlled early-psychosis cohorts.[48]
Autism spectrum disorder
ASD reviews commonly report decreased bacterial diversity and altered phylum ratios (e.g., decreased Bacteroidetes-to-Firmicutes ratio in some cohorts) as recurring dysbiosis themes, alongside elevations of specific Clostridium groups in some studies.[15, 16] Nonetheless, syntheses emphasize that the exact microbial composition associated with ASD remains undetermined, with contradictory findings at phylum, genus, species, and diversity levels.[49]
Mechanistically, ASD-focused literature emphasizes increased gut permeability (“leaky gut”) allowing bacterial metabolites and endotoxins such as LPS into systemic circulation, with downstream pro-inflammatory mediators (including IL-6 and TNF-α) capable of compromising BBB integrity and initiating neuroinflammatory cascades linked to behavioral symptoms.[13, 14] Metabolite-level evidence includes reports of altered SCFA concentrations (including lower total SCFAs in some syntheses) and kynurenine-pathway shifts toward metabolites such as xanthurenic and quinolinic acids with reduced serotonin/melatonin pathway products.[15, 16]
Intervention evidence includes small/open-label studies and limited controlled trials suggesting that microbiota transplantation and some probiotic interventions can improve GI symptoms and sometimes ASD-related behavioral scores. For example, 8-week FMT treatment was reported to improve GI and ASD-related symptoms in 16 of 18 children, with additional syntheses noting sustained GI improvement and behavioral improvements in parental impression measures, while also highlighting observed adverse effects and the need for larger studies to clarify long-term safety and tolerability.[13, 15, 17] A randomized placebo-controlled trial of L. plantarum WCSF1 reported amelioration of gut symptoms and improvements in behavioral scores, yet other studies report no significant differences in autism severity or inflammatory markers, illustrating heterogeneity and the need for large-sample, multicenter RCTs.[50, 51]
Bipolar disorder
Bipolar disorder evidence repeatedly highlights reduced representation of Faecalibacterium in BD cohorts and identifies Faecalibacterium as a discriminating feature between BD individuals and controls, supporting a plausible role for reduced butyrate-associated taxa in BD pathophysiology.[18, 37] Mood symptom severity associations are also reported, including negative correlations between MADRS scores and Faecalibacterium abundance in at least one dataset.[19]
Mechanistic syntheses emphasize gut inflammation and leaky-gut signaling involving LPS leakage into circulation and central/systemic inflammatory immune responses, and describe that inflammatory factors and neuroactive substances produced by gut microbiota can cross the BBB, activate the HPA axis, and disrupt brain function.[20, 38] Translational models indicate that BD donor microbiota can induce anxiety-like behavior and decreased sociability in recipient mice, consistent with a causal plausibility for microbiome-mediated behavioral modulation, albeit without proving the directionality in humans.[21]
Clinical intervention evidence remains preliminary. A review-level citation reports that an 8-week probiotic supplement reduced severity of depression and mania in type I BD and that a 3-month probiotic treatment improved attention and executive function in euthymic BD participants, while other work reports that depressive symptom improvement during quetiapine treatment coincided with increases in Eubacterium rectale and Bifidobacteria, leaving open whether microbiome changes are causal or medication-driven.[20, 52] Gaps include cross-sectional designs and inability to control for medication use or standardized diet, supporting the need for longitudinal, episode-stratified trials and mechanistic measurement of metabolite outputs.[18, 37]
OCD
OCD-specific evidence in the extracted text base is limited but consistent with broader “butyrate-producer depletion” themes: one study is described as finding lower alpha diversity in OCD and lower relative abundance of three butyrate-producing genera (Oscillospira, Odoribacter, Anaerostipes) known to be anti-inflammatory, with additional discussion proposing that low Odoribacter (a butyrate producer) may increase inflammation and potentially relate to OCD onset.[9, 53]
PTSD
Within the provided PTSD-focused source, the gut microbiota is discussed as a plausible mediator of stress- and trauma-related immune and HPA-axis dysregulation, with enhanced peripheral proinflammatory cytokines and low cortisol described as predisposing individuals to develop PTSD after trauma, and stress described as a major factor altering the gut microbiota and barrier function.[54] However, that source explicitly states that, as of its publication, the role of the gut microbiota in PTSD development had never been investigated, and it cautions that cause-and-effect relationships remain difficult to establish and that germ-free study results must be cautiously transposed to human health and disease.[54]
ADHD
ADHD-specific microbiome signatures are not established in the extracted evidence set used here, but the core MGBA pathways emphasized across disorders—microbial production of neurotransmitters and neuroactive intermediates, immune–BBB–microglia coupling, and HPA-axis modulation—provide a plausible mechanistic substrate for investigating attention and executive-function domains in future ADHD-focused work.[3, 10, 25] Consistent with the plausibility of cognitive-domain sensitivity to microbiome interventions, probiotic treatment has been reported to enhance attention and executive function in a small euthymic bipolar disorder study, supporting the feasibility of cognitive endpoints as microbiome-responsive measures even when diagnostic categories differ.[20]
Other conditions
Stress-related mood disorder syntheses emphasize that chronic circadian rhythm disturbances, sleep loss, and depression can alter indigenous gut bacteria composition (e.g., reducing Lactobacillaceae and increasing taxa such as Enterococci and Lachnospiraceae), with human findings often conflicting due to confounding factors.[25, 32] Subthreshold and recurrent depressive symptom evidence in the extracted texts highlights metabolite-level signals, including improvements in mental-health-related SF-36 domains after successful FMT for recurrent C. difficile infection accompanied by increased circulating butyrate and related short-chain/carboxylic acids, and urinary 3-indoxylsulfate associations with recurrent depressive symptoms, supporting a hypothesis that microbial SCFA and indole/tryptophan metabolism pathways may be clinically relevant even outside formal MDD cohorts.[55, 56]
Cross-cutting themes
Across disorders, a recurring transdiagnostic theme is that functional signatures (reduced butyrate-producing capacity, altered SCFA output, permeability/inflammation, kynurenine bias) appear more reproducible than specific taxa lists, consistent with the broader observation that there is often “no consensus” on alpha/beta diversity and that findings vary with methodology and population differences.[7] This functional convergence is consistent with multiple lines of evidence linking dysbiosis to increased permeability and endotoxin translocation and subsequent immune activation, with cytokine-mediated effects on BBB permeability and microglial activation described as a continuation of gut–brain communication.[3, 7]
A second cross-cutting theme is the centrality of immune–metabolite coupling, where stress-related permeability and cytokines can drive IDO/TDO activation and divert tryptophan metabolism toward kynurenine, while disorder-specific consequences may differ (e.g., neurotoxic metabolites in depression; KYNA/NMDAR-related modulation in schizophrenia; altered kynurenine metabolites in ASD).[9, 16, 24, 35] A third theme is causality asymmetry: while FMT and germ-free models provide supportive causal signals for depression- and psychosis-relevant phenotypes, sources caution that translation from animal models to human disease can overstate effects and that many human studies remain cross-sectional.[30, 35, 48]
Finally, intervention heterogeneity is a pervasive theme. Controlled probiotic trials frequently show mixed efficacy in anxiety and psychobiotic studies broadly, and reviews highlight that many studies are underpowered, heterogeneous, and limited in duration, motivating biomarker-guided and mechanistically anchored trial designs.[1, 10, 28]
Justified hypotheses
The hypotheses below are designed to be testable and falsifiable, and each is grounded in convergent mechanistic and/or interventional signals from the extracted full texts.
H1
H1 proposes that reduced gut-derived butyrate/SCFA signaling causally increases depressive and anxiety symptom severity and that restoring SCFA output will reduce symptoms across MDD and comorbid anxiety phenotypes.[5, 6] This is mechanistically justified by evidence linking dysbiosis to LPS translocation and inflammation, and by stress-related permeability pathways that increase proinflammatory cytokines and divert tryptophan metabolism toward kynurenine, alongside trial evidence that probiotic supplementation can decrease depression and anxiety scores and is accompanied by favorable stress/inflammatory biomarker changes.[5, 9, 57] A key caveat is that many controlled probiotic trials in anxiety show no placebo difference and that diversity/taxa findings remain inconsistent across depression cohorts, implying that SCFA-based interventions should be stratified by baseline inflammation, SCFA output, and confounds such as medication exposure.[7, 10]
H2
H2 proposes that increased microbial aspartate degradation capacity (AspAT; “Aspartate Degradation I”) in SAD drives downstream tryptophan–kynurenine imbalance (including KYNA elevation) contributing to social fear, and that modulating this microbial function will reduce SAD severity and normalize kynurenine-pathway biomarkers.[22] Support comes from functional and taxonomic differences observed in SAD and from causal FMT-to-mouse evidence showing heightened social fear sensitivity after SAD microbiota transfer, coupled with immune and oxytocin-related changes, although the human study’s small sample and lack of symptom associations after FDR correction temper causal confidence.[22, 45]
H3
H3 proposes that microbiome effects on psychiatric symptoms depend on intact vagal signaling, predicting larger antidepressant/anxiolytic effects of SCFA-enhancing interventions in individuals with preserved vagal function and tone.[1, 23] This is supported by vagotomy experiments abolishing microbiota-induced behavioral/neurogenesis effects and by evidence that SCFAs modulate vagal activity and SERT expression, though direct human tests of vagal moderation are not provided and many studies do not directly test pathway function.[23, 27, 29]
H4
H4 proposes that immune-driven kynurenine-pathway bias is a unifying mechanism across depression, social anxiety, and schizophrenia/FEP, such that gut-linked immune activation shifts tryptophan metabolism toward kynurenine/KYNA with symptom consequences, and that interventions reducing gut-driven immune activation will normalize kynurenine markers and improve outcomes.[9, 24, 48] Supporting evidence includes depression-specific descriptions of kynurenine-branch neurotoxic metabolite accumulation, probiotic-associated reduction in kynurenine/tryptophan ratio, KYNA relevance in SAD and schizophrenia, and mechanistic linkage of kynurenate to NMDAR hypofunction in schizophrenia models, while caveats include heterogeneity of taxa findings and cross-sectional limitations in schizophrenia studies.[7, 22, 24, 35, 39]
H5
H5 proposes that reduced butyrate-associated taxa (especially Faecalibacterium-related function) increase BBB permeability and neuroinflammation, worsening negative symptoms and cognition in schizophrenia spectrum disorders, and that restoring butyrate-producing capacity will improve negative and cognitive endpoints.[12] Mechanistic plausibility is supported by the role of butyrate as an HDAC inhibitor and by schizophrenia FMT evidence showing altered hippocampal glutamate–glutamine–GABA profiles with schizophrenia-relevant behaviors, while an important caveat is that BBB permeability measures are not directly provided in the schizophrenia excerpts and medication confounding remains substantial.[11, 47]
H6
H6 proposes that ASD behavioral improvements from microbiota-targeting therapies are mediated by reduced LPS/TLR-driven inflammation that compromises the BBB and by normalization of SCFA-dependent barrier integrity and tryptophan metabolism balance away from kynurenine metabolites toward serotonin/melatonin pathways.[14–16] This is supported by “leaky gut” descriptions in ASD, cytokine-mediated BBB compromise mechanisms, and multiple reports of FMT/MTT-associated improvements in GI symptoms alongside behavioral improvements, while counterevidence includes inconsistent ASD microbial signatures, reports of null probiotic effects on autism severity or inflammatory markers, and safety/tolerability concerns for FMT requiring larger trials.[13–15, 17, 49, 50]
H7
H7 proposes that circadian disruption and sleep loss drive dysbiosis that increases HPA-axis hyperreactivity and inflammatory activation, and that combining chronobiotic alignment with microbiome modulation will outperform microbiome modulation alone in stress-related mood disorders.[25, 32] Support comes from evidence that circadian/sleep disruption changes microbiota composition, germ-free stress responses show heightened ACTH/corticosterone dynamics, and probiotics can reduce HPA hyperreactivity in stress models, while key limitations include the lack of accurate sleep-quality tests emphasized in reviews and heterogeneity of intervention methods and outcomes.[10, 25, 32, 58]
Limitations of the current evidence base
Across disorders, inconsistent alpha/beta diversity results and variable taxa findings remain major barriers to translation, with explicit statements that no consensus in diversity is evident in depression and that human findings are often conflicting due to confounding factors and methodological differences.[7, 32] Medication effects are a central confound, including antidepressant- and antipsychotic-associated microbiome alterations that complicate inference about disease-state signatures and causal directionality.[7, 11] Many schizophrenia studies are cross-sectional and not prospective, making causal relationships unfeasible, and similar cross-sectional limitations constrain SAD and GAD evidence bases.[22, 35, 44]
Intervention trials are frequently underpowered and heterogeneous, with mixed methodological quality and short durations that may miss longer-term effects, and reviewers explicitly call for standardized sequencing methods and biomarker determination to improve reproducibility.[1, 3, 28] For FMT specifically, evidence is described as limited in humans in some domains (including anxiety), and ASD syntheses emphasize safety/tolerability uncertainties and donor/protocol variability, reinforcing why standardized, well-characterized microbiota products and rigorous monitoring are needed.[10, 17, 32]
Future directions
A recurring recommendation across conditions is standardized, longitudinal, and mechanistically deep studies that combine microbiome composition with measured functional outputs (SCFAs, tryptophan/kynurenine metabolites), permeability markers, immune phenotyping, and symptom-domain outcomes, rather than relying solely on cross-sectional taxa associations.[3, 8, 59] Disorder-specific needs include: larger SAD cohorts with longitudinal symptom tracking and functional metagenomics to validate AspAT-related signals; early-onset psychosis/FEP studies to enable targeted interventions closer to disease onset; and ASD trials with multicenter, large-sample randomized designs to clarify efficacy and long-term safety of microbiota-directed therapies.[22, 33, 51]
Intervention development would benefit from biomarker-guided stratification (e.g., baseline inflammation, kynurenine/tryptophan ratio, or functional SCFA deficiency), given that probiotic effects are inconsistent across trials and may depend on baseline microbiome state and pathway activation.[8, 10, 39] Finally, mechanistic testing should explicitly quantify pathway function (vagal, immune, endocrine, metabolite) because most studies do not test each pathway directly, which currently limits the ability to falsify competing causal models.[29]
Conclusion
Across MDD, anxiety disorders, schizophrenia/FEP, ASD, and bipolar disorder, the extracted evidence supports a multi-pathway MGBA model in which dysbiosis, reduced SCFA/butyrate-associated capacity, barrier dysfunction, immune activation, stress-axis modulation, and tryptophan–kynurenine shifts interact to influence psychiatric symptoms, albeit with substantial heterogeneity at the taxa level and strong confounding by medications and study design differences.[5, 7, 9, 11, 49] Translational FMT and germ-free models provide important causal support for microbiome-driven behavioral and neurochemical changes, while human interventional evidence is most suggestive (but not definitive) for strain- and context-dependent probiotics/psychobiotics and for FMT/MTT in selected ASD or GI-associated contexts.[8, 11, 17, 30, 32] The hypotheses proposed here prioritize function-first, pathway-anchored targets—SCFAs/butyrate, kynurenine bias, vagal dependence, barrier-immune signaling, and circadian–microbiome coupling—as actionable and falsifiable directions for the next generation of biomarker-guided clinical trials.[6, 23, 25, 39]
