From microbes to milestones: Gut bacterial abundances and functional pathways associate with neurodevelopment following preterm birth

From microbes to milestones: Gut bacterial abundances and functional pathways associate with neurodevelopment following preterm birth | medRxiv /* */ /* */ <!-- <!-- /*! * yepnope1.5.4 * (c) WTFPL, GPLv2 */ (function(a,b,c){function d(a){return"[object Function]"==o.call(a)}function e(a){return"string"==typeof a}function f(){}function g(a){return!a||"loaded"==a||"complete"==a||"uninitialized"==a}function h(){var a=p.shift();q=1,a?a.t?m(function(){("c"==a.t?B.injectCss:B.injectJs)(a.s,0,a.a,a.x,a.e,1)},0):(a(),h()):q=0}function i(a,c,d,e,f,i,j){function k(b){if(!o&&g(l.readyState)&&(u.r=o=1,!q&&h(),l.onload=l.onreadystatechange=null,b)){"img"!=a&&m(function(){t.removeChild(l)},50);for(var d in y[c])y[c].hasOwnProperty(d)&&y[c][d].onload()}}var j=j||B.errorTimeout,l=b.createElement(a),o=0,r=0,u={t:d,s:c,e:f,a:i,x:j};1===y[c]&&(r=1,y[c]=[]),"object"==a?l.data=c:(l.src=c,l.type=a),l.width=l.height="0",l.onerror=l.onload=l.onreadystatechange=function(){k.call(this,r)},p.splice(e,0,u),"img"!=a&&(r||2===y[c]?(t.insertBefore(l,s?null:n),m(k,j)):y[c].push(l))}function j(a,b,c,d,f){return q=0,b=b||"j",e(a)?i("c"==b?v:u,a,b,this.i++,c,d,f):(p.splice(this.i++,0,a),1==p.length&&h()),this}function k(){var a=B;return a.loader={load:j,i:0},a}var l=b.documentElement,m=a.setTimeout,n=b.getElementsByTagName("script")[0],o={}.toString,p=[],q=0,r="MozAppearance"in l.style,s=r&&!!b.createRange().compareNode,t=s?l:n.parentNode,l=a.opera&&"[object Opera]"==o.call(a.opera),l=!!b.attachEvent&&!l,u=r?"object":l?"script":"img",v=l?"script":u,w=Array.isArray||function(a){return"[object Array]"==o.call(a)},x=[],y={},z={timeout:function(a,b){return b.length&&(a.timeout=b[0]),a}},A,B;B=function(a){function b(a){var a=a.split("!"),b=x.length,c=a.pop(),d=a.length,c={url:c,origUrl:c,prefixes:a},e,f,g;for(f=0;f<d;f++)g=a[f].split("="),(e=z[g.shift()])&&(c=e(c,g));for(f=0;f<b;f++)c=x[f](c);return c}function g(a,e,f,g,h){var i=b(a),j=i.autoCallback;i.url.split(".").pop().split("?").shift(),i.bypass||(e&&(e=d(e)?e:e[a]||e[g]||e[a.split("/").pop().split("?")[0]]),i.instead?i.instead(a,e,f,g,h):(y[i.url]?i.noexec=!0:y[i.url]=1,f.load(i.url,i.forceCSS||!i.forceJS&&"css"==i.url.split(".").pop().split("?").shift()?"c":c,i.noexec,i.attrs,i.timeout),(d(e)||d(j))&&f.load(function(){k(),e&&e(i.origUrl,h,g),j&&j(i.origUrl,h,g),y[i.url]=2})))}function h(a,b){function c(a,c){if(a){if(e(a))c||(j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}),g(a,j,b,0,h);else if(Object(a)===a)for(n in m=function(){var b=0,c;for(c in a)a.hasOwnProperty(c)&&b++;return b}(),a)a.hasOwnProperty(n)&&(!c&&!--m&&(d(j)?j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}:j[n]=function(a){return function(){var b=[].slice.call(arguments);a&&a.apply(this,b),l()}}(k[n])),g(a[n],j,b,n,h))}else!c&&l()}var h=!!a.test,i=a.load||a.both,j=a.callback||f,k=j,l=a.complete||f,m,n;c(h?a.yep:a.nope,!!i),i&&c(i)}var i,j,l=this.yepnope.loader;if(e(a))g(a,0,l,0);else if(w(a))for(i=0;i (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0];var j=d.createElement(s);var dl=l!='dataLayer'?'&l='+l:'';j.src='//www.googletagmanager.com/gtm.js?id='+i+dl;j.type='text/javascript';j.async=true;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-P4HH5NV'); Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search From microbes to milestones: Gut bacterial abundances and functional pathways associate with neurodevelopment following preterm birth View ORCID Profile Kadi Vaher , View ORCID Profile Aisling Kenny , Paula Lusarreta Parga , View ORCID Profile Lorena Jiménez-Sánchez , View ORCID Profile Helen Turner , View ORCID Profile Rebekah Smikle , View ORCID Profile Amy Corrigan , Hilary Cruickshank , Magda Rudnicka , View ORCID Profile Sue Fletcher-Watson , View ORCID Profile Debby Bogaert , View ORCID Profile James P Boardman doi: https://doi.org/10.1101/2025.09.03.25334925 Kadi Vaher a Centre for Reproductive Health, Institute for Regeneration and Repair, University of Edinburgh , Edinburgh, United Kingdom b Centre for Clinical Brain Sciences, University of Edinburgh , Edinburgh, United Kingdom Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Kadi Vaher Aisling Kenny a Centre for Reproductive Health, Institute for Regeneration and Repair, University of Edinburgh , Edinburgh, United Kingdom Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Aisling Kenny Paula Lusarreta Parga c Centre for Inflammation Research, Institute for Regeneration and Repair, University of Edinburgh , Edinburgh, United Kingdom Find this author on Google Scholar Find this author on PubMed Search for this author on this site Lorena Jiménez-Sánchez b Centre for Clinical Brain Sciences, University of Edinburgh , Edinburgh, United Kingdom d Salvesen Mindroom Research Centre, University of Edinburgh , Edinburgh, United Kingdom Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Lorena Jiménez-Sánchez Helen Turner b Centre for Clinical Brain Sciences, University of Edinburgh , Edinburgh, United Kingdom e Simpson Centre for Reproductive Health, Royal Infirmary Edinburgh , Edinburgh, United Kingdom Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Helen Turner Rebekah Smikle a Centre for Reproductive Health, Institute for Regeneration and Repair, University of Edinburgh , Edinburgh, United Kingdom Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Rebekah Smikle Amy Corrigan a Centre for Reproductive Health, Institute for Regeneration and Repair, University of Edinburgh , Edinburgh, United Kingdom Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Amy Corrigan Hilary Cruickshank e Simpson Centre for Reproductive Health, Royal Infirmary Edinburgh , Edinburgh, United Kingdom Find this author on Google Scholar Find this author on PubMed Search for this author on this site Magda Rudnicka e Simpson Centre for Reproductive Health, Royal Infirmary Edinburgh , Edinburgh, United Kingdom Find this author on Google Scholar Find this author on PubMed Search for this author on this site Sue Fletcher-Watson b Centre for Clinical Brain Sciences, University of Edinburgh , Edinburgh, United Kingdom d Salvesen Mindroom Research Centre, University of Edinburgh , Edinburgh, United Kingdom Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Sue Fletcher-Watson Debby Bogaert c Centre for Inflammation Research, Institute for Regeneration and Repair, University of Edinburgh , Edinburgh, United Kingdom f Department of Pediatric Immunology and Infectious Diseases, Wilhelmina Children’s Hospital and University Medical Center Utrecht , Utrecht, the Netherlands Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Debby Bogaert James P Boardman a Centre for Reproductive Health, Institute for Regeneration and Repair, University of Edinburgh , Edinburgh, United Kingdom b Centre for Clinical Brain Sciences, University of Edinburgh , Edinburgh, United Kingdom Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for James P Boardman For correspondence: james.boardman{at}ed.ac.uk Abstract Full Text Info/History Metrics Supplementary material Preview PDF Abstract The early life gut microbiome has been suggested to be a potential driver of neurocognitive development. Evidence for this relationship in preterm children, who are at increased risk of both gut microbiome disruptions and neurodevelopmental impairment, is scarce. In a sample of 73 very preterm infants drawn from a prospective birth cohort, we assessed associations between the neonatal gut microbiome and neurodevelopment at 9 months and 2 years. The gut microbiome was profiled from stool samples collected prior to NICU discharge using shotgun metagenomics. By taking a consensus-based analytic approach, we found strong evidence for associations between the abundances of several gut bacterial species and measures related to autistic traits (e.g. Klebsiella spp. ), socio-emotional development, including temperament (e.g. Enterobacter cloacae complex, Veillonella parvula ), and executive functioning ( Clostridium perfringens ). The abundances of functional modules involved in gut-brain signalling, particularly those involved in histamine, tryptophan and quinolinic acid metabolism, were associated with measures related to executive functioning and cognitive-behavioural flexibility. This study provides evidence that the neonatal gut microbiome composition may affect longer-term neurodevelopmental profiles following preterm birth, particularly those related to socio-emotional development, autistic traits and executive functioning. 1 Introduction The first 1000 days of life - from conception to two years - constitute a critical period for neurodevelopment, characterised by neurogenesis, synaptogenesis, myelination, rapid brain growth, and the establishment of key neural circuits underlying cognitive, emotional, and behavioural functioning 1 , 2 . During this time, the developing brain demonstrates plasticity in network formation rendering it highly sensitive to both internal and external influences. Physiological perturbations can have long-lasting effects on neurodevelopmental trajectories and impact lifelong cognitive and mental health outcomes. Preterm birth (birth before 37 weeks of gestation), and very preterm birth in particular (birth <32 weeks), can interfere with the dynamic processes occurring during this crucial window of development, shaping neurodevelopment across the lifespan. Preterm children often experience difficulties with cognitive and executive functioning, language abilities, motor skills, social interactions and emotion regulation 3 – 5 , and they are two to four times more likely to be diagnosed with autism 6 , 7 . Nevertheless, considerable heterogeneity exists within this population, and not all preterm infants experience these neurodevelopmental outcomes. This suggests that a lower gestational age (GA) alone is not sufficient to drive neurocognitive differences in children born preterm. Instead, emerging evidence highlights the impact of several, often intertwined exposures associated with preterm birth, such as infections, stress exposure and sub-optimal nutrition, which increase an infant’s likelihood for neurodevelopmental challenges 8 , 9 . Notably, variations in early life gut microbiome composition may contribute to the heterogeneity of neurodevelopmental outcomes observed in preterm children. The term gut microbiome encompasses all microorganisms inhabiting the gastrointestinal tract, along with their collective genetic material. Here, we refer more specifically to the bacterial inhabitants of the gut microbiota (and their functional capacity), which represent the microbiome component that is highest in biomass 10 and has been most extensively studied to date. Gut bacteria are known to play a key role in several essential biological processes such as metabolism and immune functions 11 – 13 , but growing evidence of their contribution to gut-brain communication and, subsequently, their influence on brain development 14 – 17 , may be particularly relevant in the context of preterm birth. Associations between the early life gut microbiome and later cognitive and behavioural outcomes have been observed in typically developing children (see Vaher et al., 2022 18 for review). For example, greater Bacteroides abundance within the first year has been positively correlated with language outcomes at 2–3 years of age 19 – 21 and negatively with motor skills 20 , whereas early life Bifidobacterium abundance has shown positive associations with motor skills 22 , and infant temperament, particularly regulation/orienting and extraversion/surgency dimensions 23 – 26 . However, the specific microbiome features reported to associate with neurodevelopmental outcomes, including alpha (within-sample) and beta diversity (between-sample) and bacterial abundances, vary widely, which is likely due to differences in study methodology, such as the timing of microbiome and outcome assessments, and statistical analyses, as well as potential age- and sex-specific relationships 18 , 22 , 23 , 27 , 28 . The microbiota–gut–brain axis has also been increasingly implicated in autism spectrum disorder (ASD), though findings are similarly varied. Several studies have reported compositional differences in the gut microbiota of autistic children, including decreased Bifidobacterium and increased Klebsiella and Clostridium species as early as five months of age, in infants having increased likelihood of later ASD diagnosis 29 . Systematic reviews and meta-analyses have documented altered abundances of Bacteroides , Clostridium , and Bifidobacterium 30 – 34 , while differences in alpha and beta diversity 29 , 30 , 35 and disruptions in short-chain fatty acid (SCFA) metabolism 35 , 36 have also been observed. However, these associations must be interpreted with caution, as it remains unclear whether microbiome alterations precede and contribute to ASD development or arise secondarily due to ASD-related behaviours, dietary preferences, or other environmental factors. Together this increasing body of literature suggests a potential role for the gut microbiome in brain development. Nevertheless, research examining microbiome–neurodevelopmental associations specifically in preterm infants remains limited, though the importance of gut microbes in shaping brain development after preterm birth is increasingly acknowledged 37 , 38 . Knowing these relationships is important given that preterm infants are frequently exposed to microbiome-altering factors, such as caesarean section delivery, antibiotics, complex feeding regimens, and prolonged hospitalisation 39 , which may increase the incidence of microbiome-driven neurodevelopmental differences in this population. The gut microbiome of preterm infants follows a patterned progression, initially with high abundances of Staphylococcus , followed by its gradual replacement with Enterococcaceae, Enterobacteriaceae , and Bifidobacteriaceae, with a strong influence of birth GA 40 – 43 . Compared to their term-born peers, preterm infants are often characterised to have reduced abundance of Bifidobacterium and increased abundance of potentially pathogenic bacteria such as Escherichia and Klebsiella 44 . Many neonatal units routinely prescribe probiotics to prevent necrotising enterocolitis and sepsis in preterm infants 45 . However, the clinical efficacy of these probiotics remains uncertain, particularly in the most vulnerable extremely preterm infants (birth <28 weeks), safety concerns have been raised, and the evidence on the potential impact of probiotics on neurodevelopment in this population is limited 46 – 48 . This variation in microbiome modification in clinical practice underscores the critical importance of understanding how the neonatal gut microbiome contributes to neurodevelopmental outcomes following preterm birth. Some previous studies have suggested that microbiome differences in the preterm population may be associated with neurostructural variation. For example, severe parenchymal brain injury was shown to associate with lower abundances of Enterococcus , Escherichia/Shigella , and Streptococcus in the early days of life, and higher abundances of both Klebsiella —a potentially pathogenic genus—and Bifidobacterium at 4–6 weeks of age 49 . In the larger preterm infant cohort from which this study subset is derived, we previously demonstrated that the abundances of gut bacterial species, especially Escherichia coli and Klebsiella spp. , associated with several MRI features of encephalopathy of prematurity such as deep grey matter microstructure 50 . Associations with neurocognitive outcomes following preterm birth have also been observed in the few available studies. A seminal study utilising data from the EPIFLORE/EPIPAGE cohort found that 4-week-old very preterm infants with gut microbiome profiles characterised by high abundances of Enterococcus or Staphylococcus subsequently had higher odds of death or neurodevelopmental delay compared to infants with microbiome compositions high in Escherichia/Shigella 51 . Furthermore, 2-year neurodevelopmental scores from the Ages and Stages Questionnaire showed weak negative correlations with Staphylococcus abundance and weak positive associations with the abundances of Escherichia . Similarly, Abrahamsson et al 52 reported that neurodevelopmental impairment-associated microbial signatures in the first month after birth mostly included higher abundances of Escherichia/Shigella and Enterococcus together with lower abundances of Enterobacter . In contrast, a smaller study found that reduced Bifidobacterium levels in the first month after birth associated with neurodevelopmental impairment at 2 years of age 53 . Sarkar et al report positive correlations between early gut alpha diversity and neurodevelopment in preterm infants, and distinct bacterial networks in children with typical development versus those referred for neurocognitive evaluation 54 . Two additional studies have shown dynamic changes in the way gut microbiome composition over the course of NICU stay correlates with behaviour at NICU discharge 55 , 56 . Collectively, these findings suggest that early gut microbiome composition in preterm infants may influence later brain and neurocognitive development, though considerable uncertainty remains regarding the specific bacterial biomarkers involved. In this work, we leveraged a richly phenotyped cohort of very preterm infants 57 with the aim to uncover the associations between the neonatal gut microbiome diversity and community composition and neurobehavioural outcomes in the first two years of life 58 . 2 Materials and Methods Given the substantial variability and mixed findings in previous literature, as well as our focus on very preterm infants who have been underrepresented in this type of research, we adopted an exploratory approach to comprehensively assess potential relationships across several neurodevelopmental domains. Given the high analytic freedom, we adopted a consensus-based, multiverse analytical approach, which was preregistered prior to analyses 58 . We have included descriptions of deviations from this analysis plan below. We followed the Strengthening The Organization and Reporting of Microbiome Studies (STORMS) checklist 58 in describing the methodology and reporting results (Supplementary table 1). 2.1 Participants Participants were very preterm infants born at <33 completed weeks’ gestation recruited to the Theirworld Edinburgh Birth Cohort (TEBC), a longitudinal study designed to investigate the effects of preterm birth on brain structure and long-term developmental outcomes 57 . Participants were recruited and faecal samples for gut microbiome profiling were sampled at the Royal Infirmary of Edinburgh, UK, between 2016-2021. Exclusion criteria were death during neonatal period, major congenital malformations, chromosomal abnormalities and congenital infection. The study was conducted according to the principles of the Declaration of Helsinki, and ethical approval was obtained from the UK National Research Ethics Service (South East Scotland Research Ethic Committee 16/SS/0154). Informed consent was obtained from a person with parental responsibility for each participant. All infants were cared for in the Neonatal unit of the Simpson Centre for Reproductive Health, Royal Infirmary of Edinburgh, following standardised feeding and antibacterial guidelines (see previous work 50 , 59 and Supplementary materials for details). Preterm infants admitted to the NICU in the Simpson Centre for Reproductive Health are not routinely administered any pro- or prebiotic supplements. Clinical data was collected from antenatal and neonatal electronic patient records (for definitions see Supplementary table 2). The participants of this work were preterm infants who had their gut microbiome samples obtained prior to discharge from the NICU and who subsequently had any neurodevelopmental outcome data available at 9 months and/or 2 years; see below and in Supplementary materials for details. 2.2 Gut microbiome Faecal sample collection, bacterial DNA extraction and shotgun metagenomic sequencing are described in detail in our previous publication 50 and in Supplementary materials. In brief, bacterial DNA from stool samples was extracted using a protocol that involves phenol/bead beating in combination with the Mag Mini DNA Isolation Kit (LGC genomics, Germany), previously known as ‘ Agowa Mag Mini DNA Isolation Kit’ 60 ; shotgun metagenomic sequencing was performed on the NovaSeq 6000 platform (Illumina); taxonomic and functional profiling were carried out using MetaPhlAn (v3.1) and HUMAnN (v3), respectively 61 . Functional gene family data were grouped to KEGG orthogroups (KOs) from which gut-brain modules (GBM) were calculated 62 . Shotgun sequencing data was available for a total of 97 preterm infants at pre-discharge 50 ; two samples were excluded from the analyses due to collection at a very early postmenstrual age (<33 weeks), prior to infant transfer to another neonatal unit (see Supplementary materials for further info on sample inclusion). In total, 73 preterm infants with microbiome data had at least one neurodevelopmental outcome measure available (see below). After removing species and KOs absent in these 73 infants and those present in <5% of the sample (i.e. species only present in up to 3 samples, a total of 106 species and 3281 KOs), the dataset consists of 71 species and 4384 KOs. From the KOs we calculated a total of 41 GBMs. 2.3 Behavioural outcomes We collected information about infants’ neurodevelopment at 4.5 months, 9 months and 2 years; behavioural assessments were conducted at corrected age for preterm infants. Infant Behaviour Questionnaire (IBQ) at 4.5 months and 9 months and Early Childhood Behavioural Questionnaire (ECBQ) at 2 years These parent reported questionnaires are designed to assess temperament between the ages of 3 and 12 months 63 , or 18 and 36 months 64 , respectively. From the questionnaire data, we calculated three broad dimensions of temperament: negative affectivity, surgency/extraversion, and regulation (equivalent to effortful control domain in ECBQ). Higher scores indicate higher prevalence of behaviours indicative of these traits. Note that data collected at 4.5 months was not included in the statistical analyses but was used to impute data for 10 infants for the 9 month timepoint; please see Supplementary materials for details. Behaviour Rating Inventory of Executive Function, Preschool Version (BRIEF-P) at 2 years BRIEF-P is designed to assess the range of behavioural manifestations of infant’s executive function based on parental report 65 . We calculated the Global Executive Functioning as the composite score for executive functioning, with higher scores indicative of greater executive functioning difficulties. Infants with high negativity or inconsistency ratings (n=1) were not included in the analyses. Quantitative Checklist for Autism in Toddlers (Q-CHAT) at 2 years Q-CHAT is designed to assess frequency of behaviours also observed in autism spectrum conditions in toddlers aged 18 to 30 months based on parental report and identify those who should be referred for diagnostic assessment 66 . We obtained a total Q-CHAT score, with higher scores indicative of more autistic traits. Bayley Scales of Infant and Toddler Development, Third Edition (Bayley-III) at 2 years Bayley-III 67 was conducted as part of the routine National Health Service developmental follow-up appointment at 2 years corrected age; data was extracted from medical records for the purposes of this study. We obtained the standardised composite scores for the five domains measured with Bayley-III: cognitive, motor, language, adaptive, and socio-emotional development, with higher scores indicative of better functioning in these domains. In Bayley-III, the cognitive, motor and language domain scores are determined through clinical assessment, while adaptive and socio-emotional domains are parent reported. 2.4 Quantification and statistical analysis Analyses were conducted separately for each outcome measure, given not all 73 infants had data available for every measure. This approach resulted in 13 analysis subsets (one per outcome) to maximise sample size for each measure. For a scheme depicting the quantification and analysis flow, please see Supplementary Figure 1. We used a consensus-based/multiverse approach, using different indices of alpha and beta diversity, and several methods of differential abundance testing. All statistical analyses were performed in R v4.4.2. 2.4.1 Beta diversity and PERMANOVA For both taxonomic and functional data, we calculated both Bray-Curtis dissimilarity (using total-sum-scaled [i.e. relative abundances] species/KO tables) and Aitchison’s distance (using centred log-ratio [CLR] transformed species/KO data) using the vegan package 68 . Then, we used permutational analysis of variance (PERMANOVA), modelled by adonis2 with 1000 permutations, to identify associations between overall microbiome community composition (i.e. beta diversity) and the outcomes of interest. Raw p-values were adjusted for multiple comparisons using the Benjamini-Hochberg method separately for each matrix; we considered raw p-values < 0.05 as nominally significant, and adjusted p-values (q) < 0.05 as statistically significant. In addition to the registered analysis for beta diversity, to assess the agreement of the results when using the two distance/dissimilarity measures, we calculated Pearson’s correlation coefficients and mean [range] absolute differences in the R 2 values. 2.4.2 Dimensionality reduction As in our previous work 50 , principal coordinates analysis (PCoA) was performed on the species-level Bray-Curtis dissimilarity matrix using the function pcoa ( ape package 69 ). Cailliez transformation was applied to correct for negative eigenvalues 70 . We extracted 5 principal coordinates (PCos) each explaining >5% of variance in the data, cumulatively explaining approximately 50% of variance in microbiome composition data. For taxonomic and functional interpretation, Spearman correlation coefficients were calculated between the PCo scores and the relative abundances of the species and GBMs (Supplementary Figures 2-3). The community PCos derived here are similar to those derived from 16S-based sequencing 50 . 2.4.3 Alpha diversity For both taxonomic and functional data, Shannon index and number of observed species/KOs were calculated, before 5% prevalence filtering. As MetaPhlAn outputs species data in relative abundances, estimates of species read counts were first computed by multiplying the relative abundance of a species in a sample by the number of total sequences generated by that sample (following quality checks). For functional data, alpha diversity indices were calculated for the KOs in RPK units rounded to the nearest integer. 2.4.4 Partial Spearman correlation Initially, as outlined in the preregistration, we performed linear regression models to investigate associations for microbiome alpha diversity indices and community composition PCos with outcome measures. However, inspection of model diagnostic plots revealed that there were violations of the assumptions, especially linearity and normal distribution of residuals. We assessed the impact of log transformations on microbiome alpha diversity indices and PCos to improve compliance with model assumptions, however, diagnostic plots continued to indicate violations of certain assumptions Therefore, we performed (partial) Spearman correlation analysis as a nonparametric alternative to associate microbiome alpha diversity indices and PCos with outcomes. As with PERMANOVA, raw p-values were adjusted for multiple comparisons using the Benjamini-Hochberg method separately for each alpha diversity index and PCo; we considered raw p-values < 0.05 as nominally significant, and adjusted p-values (q) < 0.05 as statistically significant. 2.4.5 Differential abundance testing We used three complementary differential abundance methods to investigate associations between individual microbiome features (species and GBMs) and outcomes: MaAsLin2 71 , LinDA 72 , and a multiple linear regression with CLR-transformed abundance data (as recommended in the Bugs as features tutorial 73 ). These methods were chosen as they provide flexible options for covariate adjustment, are widely used in the field, and are heterogenous in their assumptions and data transformation/normalisation approaches. To reduce the number of comparisons, we analysed microbial features which were present in at least 20% of the full sample of 73 infants, resulting in a set of 31 species and 28 GBMs to be analysed using differential abundance testing methods. For MaAsLin2, we used the relative abundance feature table as the input. No further normalisation or filtering was performed, and default options were followed. Outcome measures along with relevant covariates (see next section) were specified as fixed effects. For the linear regression modelling, 5% prevalence-filtered feature relative abundance data was CLR-transformed. Then, linear regression models were performed for each feature present in >20% of the sample and outcome measure, adjusting for relevant covariates. For LinDA, estimates of species read counts were computed by multiplying the relative abundance of a species in a sample by the number of total sequences generated by that sample. Next, the species count data was used as input for the model; the outcome measures along with relevant covariates were specified in the formula. For functional data, RPK-based GBM data was utilised as input. LinDA was chosen to replace the originally preregistered ANCOM-BC2 74 method. This was due to the relatively small sample size, which resulted in an ANCOM-BC2 output with many NA values, especially as model complexity increased with the inclusion of covariates. In all differential abundance methods, Benjamini-Hochberg correction for multiple comparisons was used across the microbiome features separately for each outcome measure. Results with p < 0.05 were considered nominally significant and those with adjusted p-value (q) < 0.25 (as per the default in MaAsLin2) were considered statistically significant. To reflect the consensus-based analysis, results from the three differential abundance testing methods were collated into evidence levels based on concordance across methods. Namely, we defined strong evidence if there were statistically significant associations (q < 0.25) between a bacterial feature and an outcome measure in at least two of the three methods used, and moderate evidence if there was a statistically significant association between a bacterial feature and an outcome measure in at least one of the three methods used along with a nominally significant (p < 0.05) association in at least one or more of the three methods used. Weak evidence was defined as nominally significant (p < 0.05) associations in at least two of the three methods used. Any other associations that only arose in one method are not discussed. To further investigate consensus and alignment between the three different methods, the raw effect sizes were converted to z-scores and Pearson correlation coefficients were computed between the z-scored effect sizes for each pair of the three differential abundance testing methods. 2.4.6 Partial least squares regression To further evaluate the robustness of the microbiome-outcome relationships, in addition to the above-described preregistered analyses, we performed sparse partial least squares regression (sPLS) analysis using the mixOmics package 75 . These analyses were performed separately for all behavioural outcome measures for which at least one microbiome feature was found to be associated at the strong or moderate evidence level in differential abundance analyses. The predictor was the CLR-transformed and 20%-prevalence filtered feature abundance table (i.e. the same 31 species and 28 GBMs that were used in differential abundance analyses). Both the feature table and behavioural variable were subset to those participants for whom this specific behavioural data was available. We used mean absolute error and cross-validation implemented in the tune.spls function to identify the optimal number of components and microbiome features. We report the proportion of variance explained by the resulting components, loadings of the microbiome features to the components, and the correlation between the selected components and the behavioural measure. Finally, we extracted the resulting microbiome components and performed linear regression analyses with the behavioural measure as the outcome and selected microbiome feature component(s) along with all covariates (see next) as the predictors to ensure the correlation was not due to confounding variables impacting either microbiome or behavioural outcomes. 2.4.7 Covariate identification and adjustment In univariable analyses, associations for a range of potential confounders with microbiome alpha diversity indices, community PCos, and outcome measures were evaluated (see section Covariate identification in Supplementary materials for more information). Depending on the distribution of microbiome features and outcome scores, for categorical covariates, we performed two-sample t-tests or Mann-Whitney U-tests, and for continuous variables, we performed Pearson or Spearman correlations. In all analyses described above, we performed multivariable models where we adjusted for any covariates that were suggestively (unadjusted p<0.1) associated with at least one microbiome feature (community PCo or alpha diversity index) and at least one neurodevelopmental outcome measure. The final list of covariates included in all fully adjusted models were: GA at birth, postmenstrual age (PMA) at sample, antibiotics < 72h after birth, birthweight z-score, high vs low proportion of exclusive breast milk days during NICU stay, infant sex, maternal BMI at pregnancy booking, maternal age and the Scottish Index of Multiple Deprivation (SIMD) quintile. For description and coding of the covariates, please see Supplementary Table 2. Maternal education and postnatal depression score also met our definition of confounder, but these variables were omitted because they were missing for >5% of participants. Given that both antibiotic exposure and breastfeeding are strong drivers of infant gut microbiome composition in early life 39 , we performed another, partially adjusted model, excluding the breast milk and antibiotic covariates. This was to assess whether including these strong microbiome determinants in our main fully adjusted model could constitute over-adjustment and potentially mask relevant biological associations. In the main text, we report results that have been adjusted for all covariates. Results arising from univariable (i.e. unadjusted), age-adjusted (i.e. adjusted for GA at birth and PMA at sample only) and partially adjusted models (i.e. without antibiotic and breast milk variables) are presented in Supplementary materials. 2.4.8 Sensitivity analyses In case of any statistically significant associations between the gut microbiome features and 9-month temperament dimensions measured with IBQ, sensitivity analyses were performed excluding the 10 infants with imputed scores (see Supplementary materials) and adjusting for all covariates. 2.4.9 Data availability All data generated in this study are stored at the University of Edinburgh. The data are not part of an open repository due to the terms of the informed parent/guardian consent, which stipulates that the use of anonymised data is for studies of perinatal health that have been approved by regulatory bodies. All raw shotgun data and their derivatives used for analysis alongside participant clinical data, sample metadata, and outcome data used in this work are deposited in Edinburgh DataVault 76 ( https://doi.org/10.7488/e65499db-2263-4d3c-9335-55ae6d49af2b ). Requests for access will be considered under the study’s Data Access and Collaboration policy and governance process ( https://www.ed.ac.uk/centre-reproductive-health/tebc/about-tebc/for-researchers/data-access-collaboration , James.Boardman{at}ed.ac.uk ). Code used for the data analysis in this paper is available on GitLab ( https://git.ecdf.ed.ac.uk/jbrl/microbiome-and-developmental-outcomes ). Specific packages and software used are described in Supplementary table 3. Additional information is available from the corresponding author ( James.Boardman{at}ed.ac.uk ) upon request. 3 Results 3.1 Participant characteristics and neurodevelopmental outcomes 73 preterm participants had available data for shotgun metagenomic sequencing prior to NICU discharge (roughly term-equivalent age [average 36.57 postmenstrual weeks]), and at least one neurodevelopmental outcome measure ( Table 1 ). None of the participants had necrotising enterocolitis. The sample sizes, distributions and intercorrelations of the outcome measures are presented in Figure 1 . As expected, all Bayley-III domains showed positive correlations between one another; Q-CHAT scores negatively correlated with all Bayley-III domains; BRIEF-P had a negative correlation with Bayley-III cognitive domain and with the 2-year effortful control temperament dimension; Bayley-III adaptive behaviour scale also positively correlated with regulation/effortful control temperament dimension at 9 months and 2 years. The temperament measures at the two timepoints appeared less correlated, potentially reflecting a developmental change; yet there were moderate positive correlations in surgency and effortful control between 9 months and 2 years. Download figure Open in new tab Figure 1. Distributions (A) and Pearson correlations (B) of neurodevelopmental outcome measures. Crossed out boxes indicate statistically non-significant correlations. BRIEF-P = Behavior Rating Inventory of Executive Function, Preschool, GEC = global executive composite, Q-CHAT = Quantitative Checklist for Autism in Toddlers; IBQ = Infant Behavior Questionnaire; ECBQ = Early Childhood Behavior Questionnaire; EFFCO-REG = effortful control/regulation, NEGAF = negative affectivity, SURGE = surgency, AB = adaptive behaviour, LANG = language, COG = cognitive, MOT = motor, SE = socio-emotional. View this table: View inline View popup Table 1. Sample characteristics. 3.2 Microbiome correlations with behavioural outcomes 3.2.1 Community level associations Microbiome community compositions of the infants included in this study are visualised in Figure 2A . Prior to NICU discharge, the gut microbiome of these infants were dominated by high levels of Bifidobacterium spp. ( B. breve, B. longum ), E. coli , E. faecalis or Klebsiella spp. (K. oxytoca, K. michiganensis, K. pneumoniae ). Download figure Open in new tab Figure 2. Outcomes in preterm infants were minimally associated with overall microbiome diversity and compositional variance. (A) Relative abundances of the 25 most abundant species identified across the dataset are visualised per sample, with all other species grouped together as residuals. Samples are ordered based on hierarchical clustering of the Bray-Curtis dissimilarity matrix using complete linkage. (B) Heatmap of permutational analysis of variance (PERMANOVA) results showing associations between outcome measures and the gut bacterial beta diversity calculated on species and KEGG ortholog data. Significance of PERMANOVA was based on 1,000 permutations. (C) Heatmap of partial Spearman correlation coefficients showing associations between gut microbiome alpha diversity calculated from species and KEGG ortholog data, and gut microbiome community principal coordinates (PCo) derived from principal coordinates analysis applied on species-level Bray-Curtis dissimilarity matrix. All models were adjusted for GA at birth, PMA at sample, antibiotic exposure < 72h of life, birthweight z-score, high vs low proportion of exclusive breast milk days during NICU stay, infant sex, Scottish Index of Multiple Deprivation quintile, maternal BMI at pregnancy booking, and maternal age. P-values were adjusted for multiple comparisons using the Benjamini-Hochberg (BH) method; asterisks denote statistical significance (^p < 0.05 [nominally significant], *q < 0.05 [statistically significant after adjustment for multiple tests]). We found minimal evidence that the overall gut microbiome structure, including taxonomic and functional alpha or beta diversity, or community composition PCos, correlate with behavioural outcomes ( Figure 2B-C ; Supplementary tables 4-7), regardless of covariate adjustment strategy. Some nominally significant associations were observed but these did not survive adjustment for multiple comparisons. For example, BRIEF-P explained close to 5% of the variance in microbiome taxonomic community composition (when measured with Aitchison’s distance) and had a moderate negative correlation with PCo3 (partial Spearman ρ=-0.46), but these associations were no longer statistically significant following FDR correction. For beta diversity analysis, we observed good concordance/agreement in results obtained using the two different distance/dissimilarity measures: the Pearson correlation of the R 2 values obtained using Aitchison vs Bray-Curtis was 0.832 for species and 0.867 for KOs, and the mean [range] absolute differences between the R 2 values were 0.64% [0.12% – 1.48%] for species and 0.53% [0.07% – 1.27%] for KOs. 3.2.2 Species and functional modules: Associations with autistic traits, socio-emotional development and executive functioning The results of the three differential abundance testing methods were overall well aligned with one another. For species-level analyses, we observed Pearson correlations of r = 0.86, r = 0.90 and r = 0.94 between the effect sizes from LinDA and MaAsLin2, LinDA and linear regression, and MaAsLin2 and linear regression, respectively. For GBM-level analyses, the respective Pearson correlation coefficients were r = 0.92, r = 0.89, r = 0.88. Species-level differential abundance testing showed strong evidence for a total of six species-outcome associations in the fully adjusted model. In particular, the abundances of Klebsiella spp. ( K. pneumoniae, K. quasipneumoniae, K. variicola) were positively, and Cutibacterium avidum and Staphylococcus haemolyticus negatively associated with Q-CHAT scores ( Figure 3 ; see Supplementary tables 8-11 for coefficients, standard errors and p-values of the three methods for unadjusted, age-adjusted, fully adjusted and partially adjusted models). The abundance of Enterobacter cloacae complex was positively associated with the Bayley-III socio-emotional development domain. Download figure Open in new tab Figure 3. Gut microbiome species relative abundances associate with socio-emotional development. (A) Heatmaps of the results of differential abundance analyses showing associations between species abundance and outcome measures of interest. Symbols denote significance level: *statistically significant (q < 0.05 after adjustment for multiple comparisons), ^nominally significant (p < 0.05 prior to adjustment). Boxed cells denote evidence level: solid boxes indicate strong evidence level while dashed boxes indicate moderate evidence level. Species are ordered based on prevalence, with the most prevalent (i.e. missing in the smallest number of infants in the whole sample) at the bottom. (B) Table highlighting strong and moderate associations observed in the differential abundance analyses. Direction of effect is indicated by colour scheme (blue for negative, red for positive). Moderate evidence was observed for another 11 associations between relative abundances of bacterial species and outcome measures. The abundance of Streptococcus parasanguinis was positively and the abundances of Negativicoccus succinicivorans , Veillonella parvula and Enterobacter cloacae complex were negatively associated with the temperament dimension of surgency at 9 months of age. Enterobacter cloacae complex also strongly associated with Bayley-III socio-emotional domain as described above. Negative associations for the abundances of Negativicoccus succinicivorans and Veillonella parvula as well as positive associations for Streptococcus parasanguinis were also observed with surgency at 2 years of age, albeit at the weak evidence level (i.e. nominally significant in at least two methods but did not survive adjustment for multiple comparisons). Negativicoccus succinicivorans again showed moderate evidence for an association with the Bayley language domain and weak evidence for an association with Bayley adaptive behaviour. Differential abundance testing of the GBMs revealed strong evidence for four GBM-outcome associations. We found positive associations for the abundances of histamine and quinolinic acid synthesis with BRIEF-P ( Figure 4A-B ; see Supplementary tables 12-15 for coefficients, standard errors and p-values of the three methods for unadjusted, age-adjusted, fully adjusted and partially adjusted models). Quinolinic acid synthesis and 17-β-estradiol degradation showed negative associations with the effortful control/regulation trait at 9 months and 2 years, respectively. Five associations showed moderate evidence for GBM abundance and outcome correlations. Notably, quinolinic acid degradation negatively (and quinolinic acid synthesis positively) associated with the BRIEF-P measure. Clostridium perfringens was the strongest contributor to the histamine synthesis module and it was among the top six contributors to both quinolinic acid degradation and synthesis modules ( Figure 4C ), all of which showed significant associations with BRIEF-P. BRIEF-P was additionally negatively associated with the menaquinone (Vitamin K2) synthesis module, and Bayley-III adaptive behaviour negatively associated with 17-β-estradiol degradation module. With regards to the GBM-outcome associations identified at the weak evidence level (Supplementary table 14), it is noteworthy that the Klebsiella spp. associated with Q-CHAT, are among the top contributors to several GBMs associated with Q-CHAT (Supplementary figure 7): acetate degradation, GABA synthesis, propionate synthesis, and the isovaleric acid synthesis (KADH pathway); whereas Staphylococcus aureus is the top contributor to nitric oxide synthesis GBM, both of which were positively correlated with Bayley language domain at the moderate and weak evidence levels, respectively. Download figure Open in new tab Figure 4. Gut-brain modules associate with executive functioning and adaptive behaviour. (A) Heatmaps of the results of the differential abundance analyses showing associations between gut-brain module abundances and outcome measures of interest. Symbols denote significance level: *statistically significant (q < 0.05 after adjustment for multiple comparisons), ^nominally significant (p < 0.05 prior to adjustment). Boxed cells denote evidence level: solid boxes indicate strong evidence level while dashed boxes indicate moderate evidence level. Modules are ordered based on prevalence, with the most prevalent (i.e. missing in the smallest number of infants in the whole sample) at the bottom. (B) Table highlighting strong and moderate associations observed in the differential abundance analyses. Direction of effect is indicated by colour scheme (blue for negative, red for positive). (C) Relative abundance of the four gut-brain modules associated with neurodevelopmental outcomes stratified by species to show the contributions from known and unknown bacteria; for each module, only the top seven species with the highest abundance of genes present in that module are shown. Depth of the blue colour represents the mean relative abundance of the species in the dataset, with darker colours corresponding to higher overall relative abundance. Given the observed significant associations between 9-month temperament dimensions and abundances of bacterial species and GBMs, as a sensitivity analysis we repeated the models excluding the 10 infants with imputed scores (see Methods and Supplementary materials). Overall, the pattern of results aligned well with those reported above, both for species (Supplementary figure 4) and GBMs (Supplementary figure 5); the exceptions were the associations for surgency with Streptococcus parasanguinis and Negativicoccus succinicivorans which were no longer statistically significant. For the three differential abundance methods, the correlations of the model coefficients resulting from the analyses performed with imputed vs non-imputed IBQ data were between 0.87 and 0.97 for species-level and 0.95 and 0.96 for GBM-level analyses, suggesting that the main results were not driven by data imputation. Regarding the different covariate adjustment strategies, overall, the top findings were similar for the species-level differential abundance analyses in the age-adjusted and fully adjusted models, with some species-outcome associations reducing in magnitude with the inclusion of more covariates as expected (Supplementary tables 8-11). We observed the greatest differences in the top findings for the unadjusted univariable models (Supplementary table 8), with several associations only appearing with the inclusion of age and other covariates. This could be due to noise in the effect estimates given the relatively wide sampling age range. Excluding antibiotic and breast milk covariates (partially adjusted model, Supplementary table 11) also resulted in very similar top results compared to the fully adjusted model, suggesting that the majority of the species-outcome associations are independent of these previously identified early life drivers of microbiota development. However, a few associations changed evidence strength from moderate to strong (e.g. between Finegoldia magna and negative affectivity at 2 years ) or from moderate to weak (e.g. between K. oxytoca and negative affectivity at 9 months), or only appeared in one or the other model (e.g. moderate evidence for a negative association between Enterobacter cloacae complex and Q-CHAT in the partially adjusted model), reflecting that some effects may be sensitive to these covariates. For the GBM-outcome relative abundance analysis, the top findings were also consistent across the unadjusted, age-adjusted and fully adjusted models (Supplementary tables 12-14), though several associations expectedly changed in evidence strength level from stronger to weaker whilst some only appeared as relevant using our evidence level thresholding with the additional adjustment for potential confounders. As with the species-level differential abundance analyses, with the exclusion of antibiotic and breast milk covariates (Supplementary table 15), some associations changed evidence levels between the fully vs partially adjusted models (e.g. a change from moderate to strong for the negative association between 17-β-estradiol degradation and Bayley adaptive behaviour, and a change from strong to moderate for the negative association between quinolinic acid synthesis and effortful control at 2 years), while some only appeared in one or the other (e.g. strong evidence for an association between quinolinic acid degradation and effortful control at 2 years in the partially adjusted model, and moderate evidence for BRIEF-P associations with quinolinic acid degradation, tryptophan synthesis and menaquinone synthesis in the fully adjusted model). In sum, the main results were broadly consistent across the covariate adjustment strategies, with only a minority of associations changing in strength of evidence or appearing uniquely in one model, highlighting the robustness of the main findings and indicating that the inclusion of strong determinants of early life microbiome (breast milk and antibiotic variables) did not result in a substantial over-adjustment. In previous work 50 , using 16S-based data, there was a good agreement between community-level PCo-based and differential abundance analyses for interpreting the microbiome-brain MRI relationships 50 ; however, this agreement was not observed here for neurobehavioral outcomes. Because of this, we conducted an additional supervised data reduction analysis by performing sPLS to identify potential latent components in the microbiome features data that explain variance in the behavioural outcomes and to explore if the species/GBMs contributing to the latent components overlap with those associated with the outcomes in differential abundance analyses. To a large degree, sPLS results were aligned with the main observations from differential abundance testing. For example, the first latent component, which significantly correlated with Q-CHAT scores (r=0.589, explaining 34.7% of the variation) contained all 5 species that showed strong or moderate evidence for association with Q-CHAT in differential abundance analyses; four of these species (all except S. haemolyticus ) were among the top 5 contributors to this 20-species component ( Figure 5A ). Similar trends were observed for the Bayley socio-emotional development ( Figure 5B ) and 2-year negative affectivity measures ( Figure 5C ), where the species significantly associated with these outcomes in differential abundance analyses, e.g. Enterobacter cloacae complex and Finegoldia magna, were also selected as the top contributors to the sPLS component. These components explained 28.06% (r=0.530) and 44.9% (r=0.670) of variance in these outcomes, respectively (see Supplementary figure 8 for the loadings of species to the sPLS components and their correlations with the rest of the outcome measures). Download figure Open in new tab Figure 5. Results of the sparse partial least squares analysis demonstrating the loadings of species or gut-brain modules to the first component (left) and correlation of the first component with the outcome measures (right) for (A) Q-CHAT, (B) Bayley socio-emotional composite score, (C) ECBQ negative affectivity, (D) BRIEF-P and (E) Bayley motor score. Unadjusted β notes the standardised coefficient from the linear regression model associating the sPLS components with the outcome measures, adjusted β notes standardised coefficient from a linear regression model that additionally included GA at birth, PMA at sample, antibiotic exposure < 72h of life, birthweight z-score, high vs low proportion of exclusive breast milk days during NICU stay, infant sex, Scottish Index of Multiple Deprivation quintile, maternal BMI at pregnancy booking, and maternal age as covariates. The components resulting from sPLS with GBMs also appeared mostly in line with the results of differential abundance testing. For example, 4 out of 5 GBMs correlated with BRIEF-P at strong or moderate evidence level were selected as the contributors to the 16-feature component associated with BRIEF-P with Pearson r=0.724 ( Figure 5D ). For the other four outcome measures that each had one GBM associated at the strong/moderate evidence level, sPLS similarly selected those GBMs among top 3 strongest contributors to the component associated with those developmental outcomes (Supplementary figure 9). However, sPLS also selected some other features that were not associated with the outcome measures in the fully adjusted differential abundance analyses, potentially due to the underlying correlational nature of the microbiome data itself. For example, for Q-CHAT, the second greatest contributor to the sPLS component was Enterobacter cloacae complex ( Figure 5A ), which was not identified in the fully adjusted differential abundance analyses, even at the weak evidence threshold, but this association did appear at the moderate evidence level in the partially adjusted model. Another example is the latent component identified for 9-month surgency: in addition to the Veillonella parvula and Enterobacter cloacae complex, which we found significantly associated in differential abundance analyses, the component had additionally high loadings from Bifidobacterium longum , Enterococcus faecalis and K. pneumoniae, among others (Supplementary figure 8E). For GBMs, a contrasting result is that other modules were identified as relatively stronger contributors to the component associated with the Bayley motor domain, such as p-Cresol synthesis and isovaleric acid synthesis ( Figure 5E ). Conversely, some other diverging results compared to differential abundance testing are noteworthy: in light of the result that tryptophan synthesis was an identified module both in differential abundance analyses as well as sPLS in association with BRIEF-P, it is interesting that tryptophan degradation was also the third strongest contributor to the BRIEF-P-associated component ( Figure 5D ), further strengthening the evidence that tryptophan metabolism may be important for infant development. 4 Discussion In a sample of very preterm children we report associations between the neonatal abundances of specific gut bacterial species and functional metabolic modules implicated in gut-brain signalling, and neurodevelopmental outcomes measured during the first two years of life. The associations included autistic traits, executive functioning, socio-emotional development, language skills, and temperament. The specific microbial features associated varied across neurodevelopmental measures, suggesting that gut microbiome–neurodevelopment interactions may be domain-specific and/or temporally dependent. Together, these findings indicate that distinct microbial species and associated metabolic functions may exert selective effects on specific neurodevelopmental pathways. This contributes to a growing body of evidence that the microbiome-gut-brain axis may be important for the neurodevelopment of preterm infants 4.1 No association between microbial diversity or community-level variance and neurodevelopmental outcomes We found limited evidence that general microbiome features, such as alpha and beta diversity, associate with developmental outcomes. Inconsistent results for these measures have been observed previously 18 and we further add to this literature that microbiome diversity indices may not be adequate markers of gut health or carry prognostic value for future outcomes, especially in early life, although some studies in preterm infants have shown evidence for alpha diversity-outcome relationships 52 , 54 . Despite observing several associations between overall gut microbiome community composition differences between infants, measured via community composition PCos, and brain MRI features in our previous work 50 , here we do not find evidence that these measures capturing microbiome compositional variance associate with neurodevelopmental outcomes. Instead, it appears that the abundances of specific species and their functional capacity, as will be discussed next, may carry more information for future outcomes. Our supervised data reduction analyses using sPLS further support this, indicating that distinct microbial signatures in early life may be predictive of neurodevelopment, even when general compositional differences are less apparent. In other words, there is limited evidence for a global “dysbiotic state” associated with outcomes, but there may be more specific associations between neonatal gut microbial features and different neurodevelopmental outcomes across infancy and toddlerhood. 4.2 Species abundances associate with autistic traits and socio-emotional measures We found the abundance of several species to be associated with Q-CHAT scores, a screening tool designed to measure autistic traits in toddlers. Both sPLS and differential abundance analyses suggested that higher abundance of multiple Klebsiella spp . are positively, and Cutibacterium avidum negatively, associated with higher Q-CHAT scores. Klebsiella spp . are common colonisers of the preterm infant gut 41 , 42 , with prior work showing higher abundance of Klebsiella in infants exposed to antibiotics during their NICU stay 77 – 79 , 50 . Preterm infants are more likely to be diagnosed with autism compared to their term born peers 7 , and indeed preterm toddlers compared to term-born peers show higher Q-CHAT scores in the broader TEBC cohort 80 . Klebsiella overgrowth, together with a pro-inflammatory profile in the blood, has been associated with severe brain injury after preterm birth 49 and we previously demonstrated that several Klebsiella spp. , including K. pneumoniae , as well as Cutibacterium avidum , associated with diffusion MRI microstructural measures of encephalopathy of prematurity in this sample 50 . Interestingly, Q-CHAT-associated Klebsiella spp . were top contributors to the GBM for propionate synthesis, a SCFA shown to be associated with ASD 26 , 81 . Propionate synthesis showed nominal associations with Q-CHAT here and was associated with deep grey matter microstructure in our previous work 50 , providing corroborative evidence that propionate metabolism may mediate gut-brain interactions following preterm birth. We observed different microbiome-autistic trait associations compared to those reported by Laue et al 27 , which could be due to differences in populations (term vs preterm-born), behavioural assessments (Social Responsiveness Scale vs Q-CHAT) as well as timings of microbiome profiling and behavioural measurement. Although high scores on Q-CHAT do not equate to a diagnosis of ASD, and may reflect more general behavioural or socio-emotional differences 66 , 82 , previous studies have suggested that increased abundance of Klebsiella is present in autistic children. A study of 71 autistic boys and 18 neurotypical controls in China found that autistic children had significantly greater relative abundance of Klebsiella species compared to controls 83 and a small study by Abuljadayel et al (2024) 84 also reported higher abundances of the phylum Proteobacteria , including a Klebsiella species ( K. oxytoca ) in autistic children compared to their neurotypical siblings. The extent to which our findings can be directly compared to these previous studies is limited, given that we assess early autistic traits using a proxy measure rather than a formal ASD diagnosis, and the microbiome data utilised here were collected in the neonatal period, whereas many prior studies have focused on older children. Nonetheless, these results suggest temporal associations between early-life Klebsiella spp. abundance and brain and behavioural outcomes following preterm birth. From differential abundance testing and sPLS analyses, we also found consistent evidence linking the Enterobacter cloacae complex to socio-emotional behaviours: the abundance of this species group was positively associated with the Bayley socio-emotional domain at 2 years and negatively associated with surgency and negative affectivity at 9 months. It also negatively associated with Q-CHAT in sPLS analyses and in differential abundance testing using the model without antibiotic and breast milk exposure covariates, further strengthening the evidence that this species group may be relevant in the context of behavioural outcomes in preterm-born children. Enterobacter cloacae complex species are common gut commensals, including in preterm infants 40 , and are frequent causes of hospital-acquired infections, including neonatal sepsis in intensive care settings 85 . To our knowledge, there is little previous evidence directly linking the Enterobacter cloacae complex to neurodevelopment. One study reported that, on the genus level, the abundance of Enterobacter negatively associated with thalamic volume in two-week-old infants 86 . Moreover, recent research in extremely preterm and extremely low– birth-weight infants found that relatively lower levels of the Enterobacter genus, together with higher abundances of Enterococcus and Escherichia , were observed in neonates who later exhibited poorer neurodevelopmental outcomes at age two 52 . However, as these findings are at the genus level and do not include measures of socio-emotional behaviour, comparisons with the Enterobacter cloacae complex specifically are limited. Yet, given we observed several lines of evidence for an association of this group of bacteria with different outcome measures, particularly those related to socio-emotional behaviours, further investigations into the role of Enterobacter cloacae complex in modulating the gut-brain axis in preterm infants are warranted. Moderate evidence was observed for a negative association between Veillonella parvula and surgency at 9 months, and weaker evidence for an association with the same temperament domain at 2 years of age. Veillonella is a lactate-fermenting bacteria which has been considered a signature taxon in the 4-month-old term infant microbiota 87 and whose abundance thereafter gradually decreases over the first five years of life 88 (similar to the trajectory observed for Bifidobacterium ). While specific associations between V. parvula and surgency have not been previously evidenced, associations with other temperament dimensions and affective behaviours have been observed. Aatsinki and colleagues 23 examined faecal microbiome composition at 6 months and found that, for female infants, higher levels of V. parvula and V. dispar were associated with lower fear reactivity, and across both sexes, infants with gut microbiota profiles dominated by V. dispar had lower surgency scores. On the genus level, a small study of 34 infants found that Veillonella was a significant negative driver of a non-social fear related PCo 89 , and another study reported a nominally significant positive association between Veillonella abundances at two weeks of life and negative affect at 30 months of age 90 . Furthermore, at two weeks of life, the abundance of Veillonella negatively correlated with insula volume 86 . In our previous work in preterm neonates 50 , V. parvula also associated with several microstructural indices in the deep grey matter. Collectively, our findings therefore add to the growing body of evidence linking the abundances of Veillonella spp. in early life with socio-emotional development, including the structure of brain regions involved in affective and sensory processing in the neonatal period as well as later behavioural manifestations related to affect and temperament. However, further research is required to elucidate the mechanisms through which Veillonella spp. may influence brain development, and in turn, infant temperament. Another noteworthy result is a positive correlation between Finegoldia magna and negative affectivity at 2 years of age; this species was also correlated with deep grey matter microstructure in our prior work 50 . F. magna , a commensal and opportunistic pathogen which commonly colonises the skin and mucosal surfaces, including the female urogenital tract, has been associated with preterm birth at the genus level 91 , and at the species level with chorioamnionitis 92 (a common cause of preterm birth). Although, to our knowledge, associations between F. magna and negative affectivity have not been reported previously, higher abundances of F. magna have been identified in infants of mothers with postpartum depressive symptoms, where it was also negatively associated with fine motor skills 93 . The association observed here between F. magna and negative affectivity in preterm infants is novel and highlights a potential avenue for future research. 4.3 Gut-brain modules associate with executive functioning, effortful control and regulation Analyses of the functional potential of the metagenome, both from differential abundance testing as well as sPLS models, showed strong evidence for associations between several gut-brain modules, including histamine synthesis and modules related to quinolinic acid and tryptophan metabolism, and executive functioning (EF) difficulties measured with BRIEF-P. EF, including working memory, cognitive flexibility, attention and inhibitory control, are frequently affected in children born preterm 94 – 96 , even after controlling for IQ 97 . These differences are important, given EF abilities have been consistently linked to academic success across childhood, adolescence, and adulthood 98 – 101 . Lower EF scores, including assessment using the BRIEF instrument which we used here, have also been associated with later risk for psychopathology, including depression, anxiety, attention-deficit/hyperactivity disorder (ADHD), and obsessive-compulsive disorder, as well as with alterations in resting-state functional brain connectivity 102 – 104 . This suggests that neonatal microbiome, through modulating EF in toddlerhood, may have wide-ranging implications for preterm children. To our knowledge, only one study has specifically assessed correlations between gut microbiota composition and EF in childhood, finding no evidence for associations between genus level bacterial relative abundances and EF 105 . Our use of shotgun sequencing, which allowed probing of the functional capacity, and focus on the high-risk preterm population, may have allowed us to detect previously uncovered relationships. Histamine, a neurotransmitter with several nervous system functions including synaptic plasticity, neurotransmission and modulation of the immune system 106 , 107 , has been associated with attention, wakefulness, memory and cognition 108 , 109 , and peripherally (including gut) derived histamine can impact the brain indirectly, particularly under conditions of inflammation 75 , 76 . Histamine also plays an important role in brain development 110 , with early developing histamine immunoreactive cells observed in the human foetal brain at 19 weeks and more mature-like histaminergic neurons observed by 36 weeks gestation 111 , indicating that the histaminergic system develops progressively throughout mid-to-late gestation. This suggests that in very early or extreme preterm birth, histamine reactive cells and circuits may be mid-development at the time of birth. Given this immaturity, the developing brain may be more vulnerable to increased peripheral histamine production, particularly in the context of increased inflammation. Interestingly, Clostridium perfringens , the species which was the strongest contributor to the histamine synthesis GBM, and which was positively associated with BRIEF-P, has been highlighted previously for its ability to produce high concentrations of histamine in vitro 112 . In addition, it was more frequently identified in patients with histamine intolerance (a result of histamine build-up in the body) vs healthy individuals 113 , 114 . Together, these findings suggest that Clostridium perfringens may contribute to increased activation of microbial histamine synthesis pathways, and this may represent one mechanism contributing to differences in cognitive outcomes, in this case, BRIEF-P scores. Relationships were also observed between quinolinic acid metabolism and EF difficulties in multiple measures including BRIEF-P and ECBQ effortful control domain. Quinolinic acid is an excitotoxic NMDA receptor agonist and a downstream metabolic product of tryptophan degradation through the kynurenine pathway 115 . Elevated levels of quinolinic acid have been associated with several neurocognitive and neurodegenerative conditions including Alzheimer’s disease 116 , 117 , ASD 118 and schizophrenia 119 , while imbalances in the ratio of kynurenic acid to quinolinic acid have been observed in mood disorders 120 . sPLS analysis confirmed the differential abundance findings, with tryptophan and quinolinic acid metabolism among the top GBMs contributing to the BRIEF-P- and effortful-control-associated components. Interestingly, we previously showed that tryptophan degradation GBM associated with deep grey matter microstructure 50 . These observations further support the relevance of tryptophan-related metabolism, particularly the kynurenine pathway, to neurodevelopmental outcomes involving EF, as suggested previously 121 . As with histamine synthesis, Clostridium perfringens was among the top 7 species with the highest abundance of genes involved in quinolinic acid synthesis and degradation. Although seemingly contradictory, it is not unusual for microbial species to carry genes for both the synthesis and degradation of a given compound, with functional expression depending on multiple factors, including environmental conditions. For example, a murine clostridial myonecrosis infection model showed that under inflammation-like conditions, Clostridium perfringens exhibits a significantly altered gene expression profile 122 , highlighting how environmental cues may potentially influence which metabolic pathways are favoured. Indeed, we also found several Bifidobacterium spp . among the top contributors to these quinolinic acid metabolism pathways, yet the abundance of these Bifidobacterium spp . themselves was not significantly correlated with outcomes, further highlighting the differences between the abundance of a certain bacterial species, the functional capacity of the microbes, and the expression of this function. Future studies with metatranscriptomic or targeted metabolomics analyses are needed to clarify the role of quinolinic acid metabolism in neurodevelopment, and the bacteria involved in this process. BRIEF-P also showed a negative association with menaquinone (Vitamin K2) synthesis in the fully adjusted model. Vitamin K2 is synthesised by gut bacteria 123 , 124 . Supplementation with Vitamin K2 has been proposed as a potential intervention to mitigate cognitive decline associated with Alzheimer’s disease 125 , 126 and it has shown cognitive benefits in rodent models. For example, a study in neonatal mice exposed to repeated anaesthesia—which induced “ADHD-like” behaviours such as impulsivity, hyperactivity, and cognitive impairment—found that administering Vitamin K2 30 minutes prior to each anaesthetic procedure significantly reduced these cognitive deficits 127 . Interestingly, several Klebsiella spp., including K. variicola and K. oxytoca , which showed weak levels of evidence for associations with BRIEF-P and effortful control, respectively, were among the top contributors to this functional module. Our results therefore add to the literature supporting the role of Vitamin K2 in neural functions and highlight the need of future studies investigating gut microbiome-produced Vitamin K2 in neurodevelopment. This study is among the first to provide evidence that neonatal gut microbiome, and especially it’s functional capacity, may correlate with EF and effortful control in 2-year-old preterm toddlers, motivating future studies to investigate this question in more detail. To reduce the number of tests performed in our relatively small sample, we only used the global executive composite of BRIEF-P. Yet, this rating scale consists of multiple subdomains targeting different aspects of EF, such as inhibitory control and impulsivity, task shifting ability, emotion regulation/control, planning, and working memory. The ECBQ effortful control dimension similarly captures a range of behaviours related to inhibitory control, attentional focussing and shifting. Future work should clarify the specific subdomains of EF most strongly associated with variation in early life microbiome. Beyond BRIEF-P and effortful control, the abundance of 17β-estradiol degradation GBM negatively associated with the temperament domain of regulation at 9 months of age, and the Bayley adaptive behaviour domain at 2 years. Additionally, it was amongst the top component contributors for both of these outcome measures in sPLS analysis. 17β-estradiol is a steroid hormone and the primary form of oestrogen in humans. 17β-estradiol, produced by the placenta 128 – 130 , plays a key role in foetal neurodevelopment, including sexual differentiation of the brain, synaptogenesis, myelination, and neuroprotection 131 – 133 . Several animal studies have demonstrated that estradiol treatment can reverse or mitigate the neurodevelopmental disruptions caused by prematurity, supporting neurogenesis, synapse formation, and cognitive outcomes 134 – 136 , though clinical evidence remains limited 137 . Regarding the metabolism of 17β-estradiol by the gut microbes, a recent study found that the gut microbiota of premenopausal women with depression degraded nearly 80% of serum 17β-estradiol within two hours, compared to just 19% degradation by microbiota from age-matched, non-depressed controls 138 . The researchers identified specific oestradiol-degrading bacteria, such as Klebsiella aerogenes , and demonstrated that transplantation of these microbes into recipient mice led to reduced serum estradiol levels and the onset of depressive-like behaviours. Interestingly, in our work, Klebsiella pneumoniae was among the top species with the highest abundance of genes for 17β-estradiol degradation. Although this species was not significantly associated with either regulation at 9 months or Bayley adaptive behaviour, it associated with higher Q-CHAT scores. The association between 17β-estradiol and regulation and adaptive behaviour observed in this work is novel and may warrant further investigation, particularly in light of the well-documented sex-related differences in neurodevelopmental outcomes amongst preterm infants 139 , 140 . Finally, it is important to highlight that for some outcomes which have previously been associated with gut microbiome features, we did not observe significant associations. For example, motor skills were not associated with any species’ abundances, and only weak to moderate evidence for associations with the abundances of two GBMs was observed. Furthermore, the Bayley cognitive domain only showed weak evidence for an association with the abundances of a single bacterium. This could result from several methodological differences such as the study of exclusively preterm-born individuals, the specific timepoint selected for microbiome profiling, use of shotgun rather than 16S-based sequencing and our consensus-based statistical analyses. 4.4 Cross-method concordance strengthens confidence in differential abundance results The high dimensionality of gut microbiome data, combined with substantial analytical flexibility available, has likely contributed to inconsistent findings in the literature. Therefore, in this preregistered work we opted for a consensus-based approach. We employed multiple data modelling techniques, including three differential abundance testing methods with distinct data transformations and underlying assumptions alongside sPLS analysis, and focussed on results that converge across the methods. Overall, we observed strong concordance across methods, which strengthens confidence in the robustness of our findings. However, the slight variation in statistically significant results across individual differential abundance testing methods suggests that relying on a single approach could yield divergent conclusions. This underscores the value of using a consensus approach in microbiome analyses, as has been previously suggested 141 . 4.5 Strengths and limitations This study has several strengths. Neonatal gut microbiome linked to longitudinal behavioural data in very preterm infants, who have a high likelihood for atypical neurodevelopment, is a rare resource. Shotgun metagenomic sequencing allowed us to comprehensively characterise the associations both for the taxonomic composition at the species level and as well as the functional potential of the metagenome. The latter was profiled by specifically focussing on the GBMs – curated groups of functional processes involved in gut-brain signalling – which allowed for a principled approach to data reduction. We included a wide battery of assessments, incorporating both parental reports as well as direct observational measures, to assess development comprehensively. We adopted multiple covariate adjustment strategies and were able to show that the main findings were broadly consistent across models, highlighting the robustness of the main findings. Lastly, the consensus-based approach adopted here provides a particular strength to this work, adding robustness and confidence in the reported results. Several limitations of the current study were also identified. The relatively small sample size for microbe-outcome associations, partially due to some loss of follow-up from the neonatal period to toddlerhood during the pandemic, may have limited statistical power to detect smaller effect sizes and precluded detailed analyses of sex- or GA-specific microbiome–outcome associations. Next, whilst the microbiome sampling timepoint was chosen to reflect the cumulative effects of preterm birth and NICU exposures, the single timepoint design limited our ability to examine whether specific developmental windows are particularly sensitive to microbiome influences; future studies with longitudinal sampling over the course of NICU stay are needed to further clarify the critical windows of microbial influence on neurodevelopment. Additionally, although the cohort is representative of most survivors of preterm birth in high income settings in terms of clinical features and brain MRI findings 142 , the single-centre nature of recruitment may impact the degree to which these findings can be generalised to infants cared for in other neonatal units where practices may differ. Future studies are also encouraged to include a term comparator group in microbiome-outcome investigations to determine whether the observed associations are specific to preterm infants or also apply to those born at term, and to explore if microbiome compositions more closely resembling those of term infants are associated with improved neurodevelopmental outcomes following preterm birth. Given the observational nature of this study, causal inferences cannot be drawn as the relationships observed could arise due to other shared processes affecting both the gut microbiome and behavioural development. 5 Conclusion This study identified specific features of the neonatal gut microbiome associated with neurodevelopmental outcomes during the first 2 years of life in preterm-born children. These include bacterial species linked to autistic traits, socio-emotional development, and temperament. Notably, novel links were observed between microbial functional potential and early EF in this population. Distinct associations between microbiome features and outcome domains suggest that microbiome– neurodevelopment relationships may be domain-specific, and the use of multiple analytical approaches strengthened the robustness of findings. Collectively, these results contribute to growing evidence that early gut microbiome composition and functional capacity may influence later neurodevelopment. Our findings may carry importance for the development of future microbiome-based risk stratification tools or microbiome-modification based therapies, including probiotics, to support development following preterm birth, and highlight the need to incorporate longitudinal neurodevelopmental assessments, including those related to autism, socio-emotional development, and executive functioning, into future studies evaluating such tools and therapies in preterm infants. 7 Author contributions Conceptualisation: K.V., A.K., D.B., J.P.B. Methodology: K.V., A.K., D.B., J.P.B. Investigation: A.C. (participant recruitment, sample collection and storage); A.C., K.V. (clinical data collection); K.V., P.L.P. (laboratory work, execution and quality control); K.V., L.J-S., H.T., R.S., A.C., H.C., M.R., S.F.-W. (neurobehavioural outcome data collection and quality checking) Data Curation: K.V. Formal Analysis: K.V., A.K. Visualisation: K.V., A.K. Writing – Original Draft: K.V., A.K. Writing – Review & Editing: D.B., J.P.B., all authors Supervision: D.B., J.P.B. Funding Acquisition: K.V., A.K., D.B., J.P.B All authors contributed to the design of the analyses prior to preregistration, interpretation of results, critical revision of the manuscript, and approved the final version. 8 Declaration of interest statement DB declares having received funding from OM Pharma for unrelated work. The other authors declare no competing interests. 6 Acknowledgments This research was funded in part by the Wellcome (no. 108890/Z/15/Z, 228326/Z/23/Z, 218493/Z/19/Z). For the purpose of open access, the author has applied a CC BY public copyright license to any Author Accepted Manuscript version arising from this submission. This work was supported by Theirworld ( www.theirworld.org ). K.V., A.K., L.J-S. and R.S. were supported by the Translational Neuroscience PhD Program at the University of Edinburgh, funded by Wellcome (grant numbers 108890/Z/15/Z, 228326/Z/23/Z and 218493/Z/19/Z). J.P.B. is supported by a UKRI MRC programme grant (MR/X003434/1). H.T. is a PhD student in Clinical Brain Sciences and receives funding support from Simpson Special Care Babies charity. Collection of Bayley-III data was funded by NHS Lothian. The authors are grateful to the families who consented to take part in the study. We acknowledge Heleen de Weerd and Edinburgh Genomics for executing the bioinformatic processing of the shotgun data; shotgun metagenomic sequencing was performed at Novogene. ChatGPT-5 was utilised as an aid to improve conciseness, clarity and flow while writing this manuscript. Its use was restricted to linguistic editing, and the lead authors thoroughly reviewed, adapted and edited the outputs where appropriate. The authors take full responsibility for the final text. 9 References 1. ↵ Dan , B. The first 1000 days: A critical window for neurodevelopmental trajectories and interventions – and for parents . Dev Med Child Neurol 67 , 1108 – 1109 ( 2025 ). OpenUrl PubMed 2. ↵ Cusick , S. E. & Georgieff , M. K . The Role of Nutrition in Brain Development: The Golden Opportunity of the “First 1000 Days” . J Pediatr 175 , 16 ( 2016 ). OpenUrl CrossRef PubMed 3. ↵ Johnson , S. & Marlow , N . Early and long-term outcome of infants born extremely preterm . Arch Dis Child 102 , 97 – 102 ( 2017 ). OpenUrl Abstract / FREE Full Text 4. Johnson , S. & Marlow , N. Preterm Birth and Childhood Psychiatric Disorders . Pediatric Research 2011 69:8 69 , 11 – 18 ( 2011 ). OpenUrl CrossRef 5. ↵ Wolke , D. , Johnson , S. & Mendonça , M . The Life Course Consequences of Very Preterm Birth . Annu Rev Dev Psychol 1 , 69 – 92 ( 2019 ). OpenUrl CrossRef 6. ↵ Crump , C. , Sundquist , J. & Sundquist , K . Preterm or early term birth and risk of autism . Pediatrics 148 , 2020032300 ( 2021 ). OpenUrl 7. ↵ Laverty , C. et al. The prevalence and profile of autism in individuals born preterm: a systematic review and meta-analysis . J Neurodev Disord 13 , 1 – 12 ( 2021 ). OpenUrl CrossRef PubMed 8. ↵ Boardman , J. P. & Counsell , S. J . Invited Review: Factors associated with atypical brain development in preterm infants: insights from magnetic resonance imaging . Neuropathol Appl Neurobiol 46 , 413 – 421 ( 2020 ). OpenUrl PubMed 9. ↵ Twilhaar , E. S. et al. Cognitive Outcomes of Children Born Extremely or Very Preterm Since the 1990s and Associated Risk Factors: A Meta-analysis and Meta-regression . JAMA Pediatr 172 , 361 – 367 ( 2018 ). OpenUrl PubMed 10. ↵ Sender , R. , Fuchs , S. & Milo , R . Are We Really Vastly Outnumbered? Revisiting the Ratio of Bacterial to Host Cells in Humans . Cell 164 , 337 – 340 ( 2016 ). OpenUrl CrossRef PubMed 11. ↵ Fan , Y. & Pedersen , O . Gut microbiota in human metabolic health and disease . Nature Reviews Microbiology 2020 19:1 19 , 55 – 71 ( 2020 ). OpenUrl PubMed 12. Neuman , H. , Debelius , J. W. , Knight , R. & Koren , O . Microbial endocrinology: the interplay between the microbiota and the endocrine system . FEMS Microbiol Rev 39 , 509 – 521 ( 2015 ). OpenUrl CrossRef PubMed 13. ↵ Zheng , D. , Liwinski , T. & Elinav , E . Interaction between microbiota and immunity in health and disease . Cell Res doi: 10.1038/s41422-020-0332-7 . OpenUrl CrossRef PubMed 14. ↵ Cowan , C. S. M. , Dinan , T. G. & Cryan , J. F . Annual Research Review: Critical windows – the microbiota–gut–brain axis in neurocognitive development . Journal of Child Psychology and Psychiatry 61 , 353 – 371 ( 2020 ). OpenUrl CrossRef PubMed 15. Cryan , J. F. et al. The Microbiota-Gut-Brain Axis . Physiol Rev 99 , 1877 – 2013 ( 2019 ). OpenUrl CrossRef PubMed 16. Heijtz , R. D. et al. Normal gut microbiota modulates brain development and behavior . Proc Natl Acad Sci U S A 108 , 3047 – 3052 ( 2011 ). OpenUrl Abstract / FREE Full Text 17. ↵ Sampson , T. R. & Mazmanian , S. K . Control of brain development, function, and behavior by the microbiome . Cell Host Microbe 17 , 565 – 576 ( 2015 ). OpenUrl CrossRef PubMed 18. ↵ Vaher , K. , Bogaert , D. , Richardson , H. & Boardman , J. P . Microbiome-gut-brain axis in brain development, cognition and behavior during infancy and early childhood . Developmental Review 66 , 101038 ( 2022 ). OpenUrl 19. ↵ Carlson , A. L. et al. Infant Gut Microbiome Associated With Cognitive Development . Biol Psychiatry 83 , 148 – 159 ( 2018 ). OpenUrl CrossRef PubMed 20. ↵ Sordillo , J. E. et al. Association of the Infant Gut Microbiome With Early Childhood Neurodevelopmental Outcomes: An Ancillary Study to the VDAART Randomized Clinical Trial . JAMA Netw Open 2 , Article e190905 ( 2019 ). 21. ↵ Kort , R. et al. Model Selection Reveals the Butyrate-Producing Gut Bacterium Coprococcus eutactus as Predictor for Language Development in 3-Year-Old Rural Ugandan Children . Front Microbiol 12 , 681485 ( 2021 ). OpenUrl CrossRef PubMed 22. ↵ Acuña , I. gut microbiota associated with fine motor skills et al. Infant gut microbiota associated with fine motor skills . Nutrients 13 , 1673 ( 2021 ). OpenUrl PubMed 23. ↵ Aatsinki , A.-K. et al. Gut microbiota composition is associated with temperament traits in infants . Brain Behav Immun 80 , 849 – 858 ( 2019 ). OpenUrl CrossRef PubMed 24. Fox , M. et al. Development of the infant gut microbiome predicts temperament across the first year of life . Dev Psychopathol 34 , 1914 – 1925 ( 2022 ). OpenUrl CrossRef PubMed 25. Kelsey , C. M. et al. Gut microbiota composition is associated with newborn functional brain connectivity and behavioral temperament . Brain Behav Immun 91 , 472 – 486 ( 2021 ). OpenUrl CrossRef PubMed 26. ↵ Wang , Y. et al. Association between Gut Microbiota and Infant’s Temperament in the First Year of Life in a Chinese Birth Cohort . Microorganisms 8 , 753 ( 2020 ). OpenUrl PubMed 27. ↵ Laue , H. E. , Coker , M. O. & Madan , J. C . The Developing Microbiome From Birth to 3 Years: The Gut-Brain Axis and Neurodevelopmental Outcomes . Front Pediatr 10 , ( 2022 ). 28. ↵ Rothenberg , S. E. et al. Neurodevelopment correlates with gut microbiota in a cross-sectional analysis of children at 3 years of age in rural China . Sci Rep 11 , Article 7384 ( 2021 ). OpenUrl PubMed 29. ↵ Zuffa , S. et al. Early-life differences in the gut microbiota composition and functionality of infants at elevated likelihood of developing autism spectrum disorder . Transl Psychiatry 13 , 257 ( 2023 ). OpenUrl PubMed 30. ↵ Ho , L. K. H. et al. Gut microbiota changes in children with autism spectrum disorder: A systematic review . Gut Pathog 12 , 1 – 18 ( 2020 ). OpenUrl CrossRef PubMed 31. Iglesias-Vázquez , L. , Riba , G. V. G. , Arija , V. & Canals , J . Composition of Gut Microbiota in Children with Autism Spectrum Disorder: A Systematic Review and Meta-Analysis . Nutrients 12 , 792 ( 2020 ). OpenUrl CrossRef PubMed 32. Coretti , L. et al. Gut microbiota features in young children with autism spectrum disorders . Front Microbiol 9 , 417648 ( 2018 ). OpenUrl 33. Bezawada , N. , Phang , T. H. , Hold , G. L. & Hansen , R . Autism Spectrum Disorder and the Gut Microbiota in Children: A Systematic Review . Ann Nutr Metab 76 , 16 – 29 ( 2020 ). OpenUrl 34. ↵ Andreo-Martínez , P. , Rubio-Aparicio , M. , Sánchez-Meca , J. , Veas , A. & Martínez-González , A. E . A Meta-analysis of Gut Microbiota in Children with Autism . J Autism Dev Disord 52 , 1374 – 1387 ( 2022 ). OpenUrl PubMed 35. ↵ Zhao , Y. et al. Altered Gut Microbiota as Potential Biomarkers for Autism Spectrum Disorder in Early Childhood . Neuroscience 523 , 118 – 131 ( 2023 ). OpenUrl PubMed 36. ↵ Liu , S. et al. Altered gut microbiota and short chain fatty acids in Chinese children with autism spectrum disorder . Sci Rep 9 , 1 – 9 ( 2019 ). OpenUrl CrossRef PubMed 37. ↵ He , Y. , Zhang , Y. , Li , F. & Shi , Y . White Matter Injury in Preterm Infants: Pathogenesis and Potential Therapy From the Aspect of the Gut–Brain Axis . Front Neurosci 16 , 849372 ( 2022 ). OpenUrl PubMed 38. ↵ Lu , J. & Claud , E. C . Connection between gut microbiome and brain development in preterm infants . Dev Psychobiol 61 , 739 – 751 ( 2019 ). OpenUrl PubMed 39. ↵ Aguilar-Lopez , M. , Dinsmoor , A. M. , Ho , T. T. B. & Donovan , S. M . A systematic review of the factors influencing microbial colonization of the preterm infant gut . Gut Microbes 13 , 1 – 33 ( 2021 ). OpenUrl CrossRef PubMed 40. ↵ Korpela , K. et al. Intestinal microbiota development and gestational age in preterm neonates . Sci Rep 8 , 1 – 9 ( 2018 ). OpenUrl CrossRef PubMed 41. ↵ Rao , C. et al. Multi-kingdom ecological drivers of microbiota assembly in preterm infants . Nature 591 , 633 – 638 ( 2021 ). OpenUrl CrossRef PubMed 42. ↵ Beck , L. C. et al. Strain-specific impacts of probiotics are a significant driver of gut microbiome development in very preterm infants . Nat Microbiol 7 , 1525 – 1535 ( 2022 ). OpenUrl PubMed 43. ↵ Grier , A. et al. Impact of prematurity and nutrition on the developing gut microbiome and preterm infant growth . Microbiome 5 , 158 ( 2017 ). 44. ↵ Healy , D. B. , Ryan , C. A. , Ross , R. P. , Stanton , C. & Dempsey , E. M . Clinical implications of preterm infant gut microbiome development . Nat Microbiol 7 , 22 – 33 ( 2022 ). OpenUrl PubMed 45. ↵ Aveline , A. et al. Description of probiotic use in preterm infants in England and Wales 2016-2022 . BMJ Paediatr Open 9 , ( 2025 ). 46. ↵ Sharif , S. , Meader , N. , Oddie , S. J. , Rojas-Reyes , M. X. & McGuire , W . Probiotics to prevent necrotising enterocolitis in very preterm or very low birth weight infants . Cochrane Database of Systematic Reviews 2023 , ( 2023 ). 47. Jacobs , S. E. et al. Probiotics, prematurity and neurodevelopment: follow-up of a randomised trial . BMJ Paediatr Open 1 , 176 ( 2017 ). OpenUrl 48. ↵ Upadhyay , R. P. , Taneja , S. , Chowdhury , R. , Strand , T. A. & Bhandari , N . Effect of prebiotic and probiotic supplementation on neurodevelopment in preterm very low birth weight infants: findings from a meta-analysis . Pediatr Res 87 , 811 – 822 ( 2020 ). OpenUrl PubMed 49. ↵ Seki , D. et al. Aberrant gut-microbiota-immune-brain axis development in premature neonates with brain damage . Cell Host Microbe 29 , 1558 – 1572 .e6 ( 2021 ). OpenUrl CrossRef PubMed 50. ↵ Vaher , K. et al. The neonatal gut microbiota: A role in the encephalopathy of prematurity . Cell Rep Med 5 , 101845 ( 2024 ). OpenUrl PubMed 51. ↵ Rozé , J. C. et al. Assessment of Neonatal Intensive Care Unit Practices and Preterm Newborn Gut Microbiota and 2-Year Neurodevelopmental Outcomes . JAMA Netw Open 3 , e2018119 ( 2020 ). OpenUrl 52. ↵ Abrahamsson , T. , Wejryd , E. , Sverremark-Ekström , E. , Jenmalm , M. C. & Martí , M . Dynamics of early gut microbiota maturation in extremely preterm infants in relation with neurodevelopment at 2 years . medRxiv 2025.03.17.25324095 ( 2025 ) doi: 10.1101/2025.03.17.25324095 . OpenUrl Abstract / FREE Full Text 53. ↵ Beghetti , I. et al. Early-life gut microbiota and neurodevelopment in preterm infants: any role for Bifidobacterium? Eur J Pediatr 181 , 1773 – 1777 ( 2022 ). OpenUrl PubMed 54. ↵ Sarkar , A. et al. Relationships of the very low birth weight infant microbiome with neurodevelopment at 2 and 4 years of age . Dev Psychobiol 64 , ( 2022 ). 55. ↵ Liu , X. , Cong , X. , Li , G. , Maas , K. & Chen , K . Multivariate log-contrast regression with sub-compositional predictors: Testing the association between preterm infants’ gut microbiome and neurobehavioral outcomes . Stat Med 41 , 580 – 594 ( 2022 ). OpenUrl CrossRef PubMed 56. ↵ Sun , Z. , Xu , W. , Cong , X. , Li , G. & Chen , K . Log-contrast regression with functional compositional predictors: Linking preterm infants’ gut microbiome trajectories to neurobehavioral outcome . Annals of Applied Statistics 14 , 1535 – 1556 ( 2020 ). OpenUrl PubMed 57. ↵ Boardman , J. P. et al. Impact of preterm birth on brain development and long-term outcome: protocol for a cohort study in Scotland . BMJ Open 10 , 35854 ( 2020 ). OpenUrl 58. ↵ Mirzayi , C. et al. Reporting guidelines for human microbiome research: the STORMS checklist . Nat Med 27 , 1885 – 1892 ( 2021 ). OpenUrl CrossRef PubMed 59. ↵ Sullivan , G. et al. Breast milk exposure is associated with cortical maturation in preterm infants . Ann Neurol 93 , 591 – 603 ( 2022 ). OpenUrl PubMed 60. ↵ Biesbroek , G. et al. Deep sequencing analyses of low density microbial communities: Working at the boundary of accurate microbiota detection . PLoS One 7 , e32942 ( 2012 ). OpenUrl CrossRef PubMed 61. ↵ Beghini , F. et al. Integrating taxonomic, functional, and strain-level profiling of diverse microbial communities with biobakery 3 . Elife 10 , ( 2021 ). 62. ↵ Valles-Colomer , M. et al. The neuroactive potential of the human gut microbiota in quality of life and depression . Nat Microbiol 4 , 623 – 632 ( 2019 ). OpenUrl PubMed 63. ↵ Gartstein , M. A. & Rothbart , M. K . Studying infant temperament via the Revised Infant Behavior Questionnaire . Infant Behav Dev 26 , 64 – 86 ( 2003 ). OpenUrl CrossRef Web of Science 64. ↵ Putnam , S. P. , Gartstein , M. A. & Rothbart , M. K . Measurement of fine-grained aspects of toddler temperament: The Early Childhood Behavior Questionnaire . Infant Behav Dev 29 , 386 – 401 ( 2006 ). OpenUrl CrossRef PubMed Web of Science 65. ↵ Gioia , G. A. , Espy , K. A. & Isquith , P. K. Behavior Rating Inventory of Executive Function–Preschool Version . in Lutz, FL : Psychological Assessment Resources ( 2003 ). 66. ↵ Allison , C. et al. The Q-CHAT (Quantitative CHecklist for Autism in Toddlers): A normally distributed quantitative measure of autistic traits at 18-24 months of age: Preliminary report . J Autism Dev Disord 38 , 1414 – 1425 ( 2008 ). OpenUrl CrossRef PubMed Web of Science 67. ↵ Bayley , N. Bayley Scales of Infant and Toddler Development , Third Edition . in Pearson , San Antonio, TX : Harcourt Assessment Inc ( 2006 ). 68. ↵ Oksanen , J. et al. vegan: Community Ecology Package . R package . ( 2019 ) doi: https://github.com/vegandevs/vegan/ . 69. ↵ Paradis , E. & Schliep , K. ape 5.0: an environment for modern phylogenetics and evolutionary analyses in R . Bioinformatics 35 , 526 – 528 ( 2019 ). OpenUrl CrossRef PubMed 70. ↵ Cailliez , F . The analytical solution of the additive constant problem . Psychometrika 1983 48:2 48 , 305 – 308 ( 1983 ). OpenUrl CrossRef Web of Science 71. ↵ Mallick , H. et al. Multivariable Association Discovery in Population-scale Meta-omics Studies . PLoS Comput Biol 17 , e1009442 ( 2021 ). OpenUrl CrossRef PubMed 72. ↵ Zhou , H. , He , K. , Chen , J. & Zhang , X . LinDA: linear models for differential abundance analysis of microbiome compositional data . Genome Biol 23 , 1 – 23 ( 2022 ). OpenUrl CrossRef PubMed 73. ↵ Bastiaanssen , T. F. S. , Quinn , T. P. & Loughman , A . Bugs as features (part 1): concepts and foundations for the compositional data analysis of the microbiome–gut–brain axis . Nature Mental Health 1 , 930 – 938 ( 2023 ). OpenUrl 74. ↵ Lin , H. & Peddada , S. Das. Multigroup analysis of compositions of microbiomes with covariate adjustments and repeated measures . Nat Methods 21 , 83 – 91 ( 2024 ). OpenUrl CrossRef PubMed 75. ↵ Rohart , F. , Gautier , B. , Singh , A. & Lê Cao , K. A . mixOmics: An R package for ‘omics feature selection and multiple data integration . PLoS Comput Biol 13 , e1005752 ( 2017 ). OpenUrl CrossRef PubMed 76. ↵ Boardman , J. (Owner), Hall , J. (Data manager) & Boardman , J. (Creator). Theirworld Edinburgh Birth Cohort (TEBC) and PREterm birth as a determinant of Neurodevelopment and COGnition in children: mechanisms and causal evidence (PRENCOG) . Edinburgh DataVault ( 2022 ). 77. ↵ Gibson , M. K. et al. Developmental dynamics of the preterm infant gut microbiota and antibiotic resistome . Nat Microbiol 1 , 16024 ( 2016 ). OpenUrl PubMed 78. Reyman , M. et al. Effects of early-life antibiotics on the developing infant gut microbiome and resistome: a randomized trial . Nat Commun 13 , 893 ( 2022 ). OpenUrl CrossRef PubMed 79. ↵ Vaher , K. et al. The neonatal gut microbiota: A role in the encephalopathy of prematurity . Cell Rep Med 5 , 101845 ( 2024 ). OpenUrl PubMed 80. ↵ Vaher , K. et al. Neonatal amygdala microstructure and structural connectivity are associated with autistic traits at 2 years of age . medRxiv 2024.11.29.24318196 ( 2024 ) doi: 10.1101/2024.11.29.24318196 . OpenUrl Abstract / FREE Full Text 81. ↵ Adams , J. B. , Johansen , L. J. , Powell , L. D. , Quig , D. & Rubin , R. A . Gastrointestinal flora and gastrointestinal status in children with autism - comparisons to typical children and correlation with autism severity . BMC Gastroenterol 11 , 1 – 13 ( 2011 ). OpenUrl CrossRef PubMed 82. ↵ Kleine , I. et al. Postnatal maternal depressive symptoms and behavioural outcomes in term-born and preterm-born toddlers: a longitudinal UK community cohort study . BMJ Open 12 , e058540 ( 2022 ). OpenUrl PubMed 83. ↵ Ye , F. et al. Comparison of gut microbiota in autism spectrum disorders and neurotypical boys in China: A case-control study . Synth Syst Biotechnol 6 , 120 – 126 ( 2021 ). OpenUrl PubMed 84. ↵ Abuljadayel , D. et al. Gut microbiota of children with autism spectrum disorder and healthy siblings: A comparative study . Exp Ther Med 28 , 430 ( 2024 ). OpenUrl PubMed 85. ↵ Girlich , D. et al. Uncovering the novel Enterobacter cloacae complex species responsible for septic shock deaths in newborns: a cohort study . Lancet Microbe 2 , e536 – e544 ( 2021 ). OpenUrl 86. ↵ Graf , M. et al. Gut microbiome is associated with insula structure in neonates . Physiol Behav 299 , 115001 ( 2025 ). OpenUrl PubMed 87. ↵ Bäckhed , F. et al. Dynamics and stabilization of the human gut microbiome during the first year of life . Cell Host Microbe 17 , 690 – 703 ( 2015 ). OpenUrl CrossRef PubMed 88. ↵ Roswall , J. et al. Developmental trajectory of the healthy human gut microbiota during the first 5 years of life . Cell Host Microbe 29 , 765 – 776 .e3 ( 2021 ). OpenUrl CrossRef PubMed 89. ↵ Carlson , A. L. et al. Infant gut microbiome composition is associated with non-social fear behavior in a pilot study . Nat Commun 12 , Article 3294 ( 2021 ). 90. ↵ Vogel , S. C. et al. Longitudinal associations between the infant gut microbiome and negative affect in toddlerhood . Dev Psychopathol 1 – 13 ( 2025 ) doi: 10.1017/S0954579425100229 . OpenUrl CrossRef 91. ↵ Pausan , M.-R. , et al. Human Milk Oligosaccharides Modulate the Risk for Preterm Birth in a Microbiome-Dependent and -Independent Manner . mSystems 5 , ( 2020 ). 92. ↵ Urushiyama , D. et al. Vaginal microbiome as a tool for prediction of chorioamnionitis in preterm labor: a pilot study . Sci Rep 11 , 18971 ( 2021 ). OpenUrl CrossRef PubMed 93. ↵ Zhou , L. et al. Association of maternal postpartum depression symptoms with infant neurodevelopment and gut microbiota . Front Psychiatry 15 , 1385229 ( 2024 ). OpenUrl PubMed 94. ↵ Alduncin , N. , Huffman , L. C. , Feldman , H. M. & Loe , I. M . Executive function is associated with social competence in preschool-aged children born preterm or full term . Early Hum Dev 90 , 299 – 306 ( 2014 ). OpenUrl PubMed 95. van Houdt , C. A. , Oosterlaan , J. , van Wassenaer-Leemhuis , A. G. , van Kaam , A. H. & Aarnoudse-Moens , C. S. H . Executive function deficits in children born preterm or at low birthweight: a meta-analysis . Dev Med Child Neurol 61 , 1015 – 1024 ( 2019 ). OpenUrl CrossRef PubMed 96. ↵ Sandoval , C. C. , Gaspardo , C. M. & Linhares , M. B. M . The impact of preterm birth on the executive functioning of preschool children: A systematic review . Appl Neuropsychol Child 11 , 873 – 890 ( 2022 ). OpenUrl PubMed 97. ↵ Aarnoudse-Moens , C. S. H. , Smidts , D. P. , Oosterlaan , J. , Duivenvoorden , H. J. & Weisglas-Kuperus , N . Executive function in very preterm children at early school age . J Abnorm Child Psychol 37 , 981 – 993 ( 2009 ). OpenUrl CrossRef PubMed Web of Science 98. ↵ Spiegel , J. A. , Goodrich , J. M. , Morris , B. M. , Osborne , C. M. & Lonigan , C. J . Relations between executive functions and academic outcomes in elementary school children: A meta-analysis . Psychol Bull 147 , 329 – 351 ( 2021 ). OpenUrl PubMed 99. Pascual , A. C. , Moyano , N. & Robres , A. Q . The Relationship Between Executive Functions and Academic Performance in Primary Education: Review and Meta-Analysis . Front Psychol 10 , 1582 ( 2019 ). OpenUrl CrossRef PubMed 100. Best , J. R. , Miller , P. H. & Naglieri , J. A . Relations between Executive Function and Academic Achievement from Ages 5 to 17 in a Large, Representative National Sample . Learn Individ Differ 21 , 327 ( 2011 ). 101. ↵ Samuels , W. E. , Tournaki , N. , Blackman , S. & Zilinski , C . Executive functioning predicts academic achievement in middle school: A four-year longitudinal study . J Educ Res 109 , 478 – 490 ( 2016 ). OpenUrl 102. ↵ Halse , M. , Steinsbekk , S. , Hammar , Å. & Wichstrøm , L . Longitudinal relations between impaired executive function and symptoms of psychiatric disorders in childhood . Journal of Child Psychology and Psychiatry 63 , 1574 – 1582 ( 2022 ). OpenUrl CrossRef PubMed 103. Hawkey , E. J. , Tillman , R. , Luby , J. L. & Barch , D. M . Preschool Executive Function Predicts Childhood Resting-State Functional Connectivity and Attention-Deficit/Hyperactivity Disorder and Depression . Biol Psychiatry Cogn Neurosci Neuroimaging 3 , 927 – 936 ( 2018 ). OpenUrl PubMed 104. ↵ Willoughby , M. T. , Wylie , A. C. & Blair , C. B . Using Repeated-Measures Data to Make Stronger Tests of the Association between Executive Function Skills and Attention Deficit/Hyperactivity Disorder Symptomatology in Early Childhood . J Abnorm Child Psychol 47 , 1759 – 1770 ( 2019 ). OpenUrl PubMed 105. ↵ Eckermann , H. A. , Ou , Y. , Lahti , L. & de Weerth , C . Can gut microbiota throughout the first 10 years of life predict executive functioning in childhood? Dev Psychobiol 64 , e22226 ( 2022 ). OpenUrl PubMed 106. ↵ Han , S. , Márquez-Gómez , R. , Woodman , M. & Ellender , T . Histaminergic Control of Corticostriatal Synaptic Plasticity during Early Postnatal Development . Journal of Neuroscience 40 , 6557 – 6571 ( 2020 ). OpenUrl Abstract / FREE Full Text 107. ↵ Carthy , E. & Ellender , T . Histamine, Neuroinflammation and Neurodevelopment: A Review . Front Neurosci 15 , 680214 ( 2021 ). OpenUrl CrossRef PubMed 108. ↵ Passani , M. B. , Panula , P. & Lin , J. S . Histamine in the brain . Front Syst Neurosci 8 , 82804 ( 2014 ). OpenUrl 109. ↵ Chikahisa , S. et al. Histamine from Brain Resident MAST Cells Promotes Wakefulness and Modulates Behavioral States . PLoS One 8 , e78434 ( 2013 ). OpenUrl CrossRef PubMed 110. ↵ Molina-Hernández , A. , Díaz , N. F. & Arias-Montaño , J. A . Histamine in brain development . J Neurochem 122 , 872 – 882 ( 2012 ). OpenUrl CrossRef PubMed 111. ↵ Khedkar , T. , Koushik , S. & Gadhikar , Y . Expression pattern of histaminergic neurons in the human fetal hypothalamus at second and third trimester . Ann Neurosci 19 , 116 ( 2012 ). OpenUrl PubMed 112. ↵ Engevik , K. A. et al. Phylogenetically diverse bacterial species produce histamine . Syst Appl Microbiol 47 , ( 2024 ). 113. ↵ Sánchez-Pérez , S. et al. Intestinal Dysbiosis in Patients with Histamine Intolerance . 14 , 1774 ( 2022 ). OpenUrl 114. ↵ Rocha , S. M. , Pires , J. , Esteves , M. , Baltazar , G. & Bernardino , L . Histamine: a new immunomodulatory player in the neuron-glia crosstalk . Front Cell Neurosci 8 , 120 ( 2014 ). OpenUrl CrossRef PubMed 115. ↵ Stone , T. W. & Darlington , L. G . The kynurenine pathway as a therapeutic target in cognitive and neurodegenerative disorders . Br J Pharmacol 169 , 1211 – 1227 ( 2013 ). OpenUrl CrossRef PubMed 116. ↵ Guillemin , G. J. et al. Quinolinic acid in the pathogenesis of alzheimer’s disease . Adv Exp Med Biol 527 , 167 – 176 ( 2003 ). OpenUrl CrossRef PubMed Web of Science 117. ↵ Ting , K. K. , Brew , B. J. & Guillemin , G. J . Effect of quinolinic acid on human astrocytes morphology and functions: Implications in Alzheimer’s disease . J Neuroinflammation 6 , 1 – 13 ( 2009 ). OpenUrl CrossRef PubMed 118. ↵ Lim , C. K. et al. Altered kynurenine pathway metabolism in autism: Implication for immune-induced glutamatergic activity . Autism Research 9 , 621 – 631 ( 2016 ). OpenUrl PubMed 119. ↵ Cathomas , F. , Guetter , K. , Seifritz , E. , Klaus , F. & Kaiser , S . Quinolinic acid is associated with cognitive deficits in schizophrenia but not major depressive disorder . Sci Rep 11 , 9992 ( 2021 ). OpenUrl PubMed 120. ↵ Marx , W. et al. The kynurenine pathway in major depressive disorder, bipolar disorder, and schizophrenia: a meta-analysis of 101 studies . Mol Psychiatry 26 , 4158 – 4178 ( 2021 ). OpenUrl PubMed 121. ↵ Spichak , S. et al. Mining microbes for mental health: Determining the role of microbial metabolic pathways in human brain health and disease . Neurosci Biobehav Rev 125 , 698 – 761 ( 2021 ). OpenUrl CrossRef PubMed 122. ↵ Low , L. Y. et al. Concurrent host-pathogen transcriptional responses in a clostridium perfringens murine myonecrosis infection . mBio 9 , ( 2018 ). 123. ↵ Smajdor , J. et al. The impact of gut bacteria producing long chain homologs of vitamin K2 on colorectal carcinogenesis . Cancer Cell Int 23 , 1 – 14 ( 2023 ). OpenUrl CrossRef PubMed 124. ↵ Bonaldo , F. & Leroy , F . Review: Bacterially produced vitamin K2 and its potential to generate health benefits in humans . Trends Food Sci Technol 147 , 104461 ( 2024 ). OpenUrl 125. ↵ Popescu , A. & German , M . Vitamin K2 Holds Promise for Alzheimer’s Prevention and Treatment . Nutrients 2021, Vol. 13, Page 2206 13 , 2206 ( 2021 ). OpenUrl PubMed 126. ↵ Roumeliotis , S. et al. The role of vitamin K2 in cognitive impairment: linking vascular health to brain health . Front Aging Neurosci 16 , 1527535 ( 2024 ). OpenUrl PubMed 127. ↵ Hou , Y. et al. Vitamin K2 mitigates cognitive, and motor impairments induced by multiple surgery with anesthesia and analgesia in neonatal stage . Biochem Biophys Res Commun 765 , 151784 ( 2025 ). OpenUrl PubMed 128. ↵ Tang , C. et al. RGS2 promotes estradiol biosynthesis by trophoblasts during human pregnancy . Exp Mol Med 55 , 240 ( 2023 ). OpenUrl PubMed 129. Tulchinsky , D. & Hobel , C. J . Plasma human chorionic gonadotropin, estrone, estradiol, estriol, progesterone, and 17α-hydroxyprogesterone in human pregnancy . III. Early normal pregnancy. Am J Obstet Gynecol 117 , 884 – 893 ( 1973 ). OpenUrl 130. ↵ Barrera , D. et al. Estradiol and progesterone synthesis in human placenta is stimulated by calcitriol . J Steroid Biochem Mol Biol 103 , 529 – 532 ( 2007 ). OpenUrl CrossRef PubMed 131. ↵ Arevalo , M. A. , Santos-Galindo , M. , Bellini , M. J. , Azcoitia , I. & Garcia-Segura , L. M . Actions of estrogens on glial cells: Implications for neuroprotection . Biochimica et Biophysica Acta (BBA) - General Subjects 1800 , 1106 – 1112 ( 2010 ). OpenUrl PubMed 132. McCarthy , M. M. Estradiol and the Developing Brain . Physiol Rev 88 , 91 ( 2008 ). OpenUrl CrossRef PubMed Web of Science 133. ↵ Pansiot , J. , Mairesse , J. & Baud , O . Protecting the developing brain by 17β-estradiol . Oncotarget 8 , 9011 ( 2017 ). OpenUrl PubMed 134. ↵ Sharma , D. R. et al. Oestrogen treatment restores dentate gyrus development in premature newborns by IGF1 regulation . J Cell Mol Med 27 , 2467 ( 2023 ). OpenUrl PubMed 135. Panda , S. et al. Estrogen Treatment Reverses Prematurity-Induced Disruption in Cortical Interneuron Population . The Journal of Neuroscience 38 , 7378 ( 2018 ). OpenUrl Abstract / FREE Full Text 136. ↵ Tibrewal , M. et al. Disruption of Interneuron Neurogenesis in Premature Newborns and Reversal with Estrogen Treatment . Journal of Neuroscience 38 , 1100 – 1113 ( 2018 ). OpenUrl Abstract / FREE Full Text 137. ↵ Trotter , A. et al. Follow-Up Examination at the Age of 15 Months of Extremely Preterm Infants after Postnatal Estradiol and Progesterone Replacement . J Clin Endocrinol Metab 86 , 601 – 603 ( 2001 ). OpenUrl CrossRef PubMed Web of Science 138. ↵ Li , D. et al. Gut-microbiome-expressed 3β-hydroxysteroid dehydrogenase degrades estradiol and is linked to depression in premenopausal females . Cell Metab 35 , 685 – 694 .e5 ( 2023 ). OpenUrl CrossRef PubMed 139. ↵ Christensen , R. et al. Preterm Sex Differences in Neurodevelopment and Brain Development from Early Life to 8 Years of Age . Journal of Pediatrics 276 , ( 2025 ). 140. ↵ Boghossian , N. S. et al. Trends in sex differences in neurodevelopmental outcomes among extremely preterm infants . Arch Dis Child Fetal Neonatal Ed 110 , 269 – 278 ( 2025 ). OpenUrl Abstract / FREE Full Text 141. ↵ Nearing , J. T. et al. Microbiome differential abundance methods produce different results across 38 datasets . Nat Commun 13 , 1 – 16 ( 2022 ). OpenUrl CrossRef PubMed 142. ↵ Sullivan , G. et al. Brain 3T magnetic resonance imaging in neonates: features and incidental findings from a research cohort enriched for preterm birth . Arch Dis Child Fetal Neonatal Ed 110 , 85 – 90 ( 2025 ). OpenUrl Abstract / FREE Full Text View the discussion thread. Back to top Previous Next Posted September 05, 2025. Download PDF Supplementary Material Email Thank you for your interest in spreading the word about medRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following From microbes to milestones: Gut bacterial abundances and functional pathways associate with neurodevelopment following preterm birth Message Subject (Your Name) has forwarded a page to you from medRxiv Message Body (Your Name) thought you would like to see this page from the medRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share From microbes to milestones: Gut bacterial abundances and functional pathways associate with neurodevelopment following preterm birth Kadi Vaher , Aisling Kenny , Paula Lusarreta Parga , Lorena Jiménez-Sánchez , Helen Turner , Rebekah Smikle , Amy Corrigan , Hilary Cruickshank , Magda Rudnicka , Sue Fletcher-Watson , Debby Bogaert , James P Boardman medRxiv 2025.09.03.25334925; doi: https://doi.org/10.1101/2025.09.03.25334925 Share This Article: Copy Citation Tools From microbes to milestones: Gut bacterial abundances and functional pathways associate with neurodevelopment following preterm birth Kadi Vaher , Aisling Kenny , Paula Lusarreta Parga , Lorena Jiménez-Sánchez , Helen Turner , Rebekah Smikle , Amy Corrigan , Hilary Cruickshank , Magda Rudnicka , Sue Fletcher-Watson , Debby Bogaert , James P Boardman medRxiv 2025.09.03.25334925; doi: https://doi.org/10.1101/2025.09.03.25334925 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Pediatrics Subject Areas All Articles Addiction Medicine (568) Allergy and Immunology (863) Anesthesia (300) Cardiovascular Medicine (4433) Dentistry and Oral Medicine (444) Dermatology (382) Emergency Medicine (608) Endocrinology (including Diabetes Mellitus and Metabolic Disease) (1508) Epidemiology (15227) Forensic Medicine (30) Gastroenterology (1124) Genetic and Genomic Medicine (6594) Geriatric Medicine (667) Health Economics (997) Health Informatics (4533) Health Policy (1368) Health Systems and Quality Improvement (1613) Hematology (540) HIV/AIDS (1264) Infectious Diseases (except HIV/AIDS) (15915) Intensive Care and Critical Care Medicine (1103) Medical Education (623) Medical Ethics (145) Nephrology (667) Neurology (6596) Nursing (346) Nutrition (998) Obstetrics and Gynecology (1144) Occupational and Environmental Health (957) Oncology (3332) Ophthalmology (972) Orthopedics (369) Otolaryngology (420) Pain Medicine (436) Palliative Medicine (130) Pathology (663) Pediatrics (1692) Pharmacology and Therapeutics (691) Primary Care Research (711) Psychiatry and Clinical Psychology (5447) Public and Global Health (9227) Radiology and Imaging (2198) Rehabilitation Medicine and Physical Therapy (1370) Respiratory Medicine (1196) Rheumatology (593) Sexual and Reproductive Health (712) Sports Medicine (530) Surgery (712) Toxicology (99) Transplantation (289) Urology (265) (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'a00019093afe0db4',t:'MTc3OTQ5Nzc1NQ=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();