To the best of our knowledge, this is the first study to indicate a correlation between the gut microbiome and brain structure and function in SZs. The main findings are as follows: 1) Consistent with previous studies, significant between-group differences in gut microbiota and MRI indexes were found; 2) Both the alpha diversity evenness and Shannon indexes showed a positive correlation with the GMV and ReHo indexes in several brain regions; and 3) A lower ReHo indexes in the right STC, left cuneus and right MTC were negatively correlated with a lower relative abundance of the genus Roseburia.
The results from 16S rRNA sequencing demonstrated significant changes in microbial composition from the Bray-Curtis distances between the two groups, suggesting greater intragroup differences in the gut microbiome of SZs. Furthermore, shifts in taxonomic abundance in SZs were consistent with previous studies. For instance, Roseburia (order Clostridiales) and Ruminococcus (order Lactobacillales) showed depletion in SZs, which was observed in multiple previous studies on psychiatric diseases 4,27,28. Roseburia and Ruminococcus are representative bacteria that produce short-chain fatty acids (SCFAs) 29. Microbial-derived SCFAs can cross the blood-brain barrier (BBB) and activate specific receptors in relevant brain regions pertinent to depression and anxiety-related behaviors 30. In fact, the decrease in the relative abundance of Roseburia may be detrimental to insulin sensitivity and thus affect the concentrations of branched-chain amino acids (BCAAs) 31. Due to the brain transporters shared between BCAAs and tryptophan, the presence of excessive BCAAs would cause decreased efficiency in the transportation of tryptophan. It has been reported that excessive consumption of BCAAs could lead to decreased cerebral concentrations of tryptophan, the precursor of serotonin, and thus result in decreased cerebral 5-hydroxytryptamine concentrations 32. Li et al. 33 speculated that depletion of the Clostridiales taxa, which degrade BCAAs, leads to an elevated concentration of BCAAs in the circulatory system and therefore indirectly decreases cerebral serotonin concentrations that regulate mood.
In this study we found the presence of structural abnormalities in SZs. Our results showed that the GMV of some brain regions, including the bilateral insula, frontal and temporal (see Supplementary Table S2 for the detailed regions) in SZs were decreased compared with NCs. These results are in line with previous studies which found that SZs showed decreased GMV in the insula 34, superior temporal pole 35, amygdala, anterior cingulate, and frontal cortices (superior, middle, opercular inferior, and orbital frontal gyrus) 34,35. Van Rheenen et al. reported the entire cortex volume reductions in SZs with cognitive impairments 36. Actually, a larger cortical volume or greater grey matter density in most brain regions is often associated with better computational efficacy 37. Thus, we inferred that the cognitive impairments may related with the reduction of GMV in SZs.
We found that most brain regions showed a lower functional brain activity in SZs than those in NCs (Supplementary Table S2). This result is consistent with previous studies that reported altered ReHo in SZs in the bilateral STC 38, the MTC 22,38, the bilateral superior medial prefrontal cortex (mPFC) 20, the right superior frontal gyrus (SFG), and the fusiform gyrus 20,22 compared with those in NCs. Most of the brain regions that showed altered ReHo, such as the STG, MTC and SFG, are related to visual and auditory perception and are mainly located in the frontal and temporal areas (Supplementary Table S2). Several studies have reported that temporal lobe abnormalities may be related to the emergence of auditory hallucinations 39, abnormal language processing 39, thought disorder 40, and other psychotic symptoms in SZs. Additionally, the frontal lobe mediates a number of important processes that may impact executive function, working memory, abstract reasoning, social behavior, empathy, self-monitoring, and impulse control in SZ 39,41. We also found that SZs showed increased ALFF in the right caudate than NCs (Supplementary Table S2). Previous studies reported that SZs showed significantly increased ALFF in the right caudate nucleus 42,43, middle temporal gyrus, inferior parietal lobule 42, bilateral prefrontal and parietal cortex, and left superior temporal cortex 43 compare with NCs. These findings suggested that the local synchronization of spontaneous activity and amplitude of fluctuation in SZ brains was widely disrupted, which could explain the psychopathology of SZ.
We observed significant positive associations between the gut microbial diversity measures and the reduction GMV in SZs (Fig. 3). Previous studies found that variations of genes were associated with GMV reduction in SZs, especially in the prefrontal cortex and anterior cingulate cortex 44-46. The transcription-neuroimaging association analysis found that expression levels of 98 genes were significantly correlated with GMV changes in SZs 47. Although no prior study reported the relationship between GMV and gut microbiome in SZs, decreased structural integrity of both white and gray matter regions including hippocampus in mice that were colonized with attention-deficit/hyperactivity disorder (ADHD) microbiota 48. A study reported that the relative abundance of Bacteroides showed greater prominence in the cerebellum, frontal regions, and the hippocampus in women, which further explained that microbial modulation may affect mood and behavior 49.
Some microbiota have been reported to be associated with brain function. Prevotella has been shown to be associated with the development of brain abscesses and other neurological syndromes via production of IgA proteases that promote virulence and initiate an immune response 50,51. Lin et al. reported that the genus Neisseria is negatively associated with functional network connectivity (FNC) loading, especially the FNC between the left angular gyrus and right inferior occipital gyrus, which is related to visual processing function 52. In addition, members of the genus Neisseria, including the species Neisseria meningitidis, stimulate the immune system through a variety of mechanisms and invade the neurological nervous system during infection 53. In this study, we found that the depletion of the genus Roseburia was significantly associated with the local synchronization of spontaneous activity of the right STC and the right MTC, which are related to auditory verbal hallucinations 54,55 and thought disturbances 56 in SZs. Dhiman et al. 57 reported that Roseburia was associated with good cognitive performance, which may further support our findings. A preliminary study on the gut microbiome and brain functional connectivity in infants revealed that alpha diversity was significantly associated with functional connectivity between the amygdala and thalamus, and between the anterior cingulate cortex and anterior insula 58, suggesting a potential pathway linking gut microbial diversity and cognitive outcomes. We found that both evenness and Shannon of microbial diversity in SZs were positively associated with ReHo indexes of the bilateral calcarine cortex, bilateral lingual gyrus, left superior occipital cortex and right superior parietal cortex. Additionally, evenness showed positive correlations with ReHo indexes of the right cuneus lobe, bilateral fusiform gyrus, left postcentral gyrus and left superior parietal cortex (Table S3). We inferred that the sensory and cognitive impairments may be related to the alteration of ReHo indexes and microbial alpha diversity in SZs. For example, gut microbial alpha diversity in SZs may be associated with visual hallucinations 54,59, and this relationship could be mediated by functional brain activity in the calcarine cortex.
The present study has several limitations. First, the SZs varied considerably in their medications. Specifically, the durations and types of medication for the SZs were distinct. It is important to note that medication can affect microbiota composition and brain activity, which may further affect the results. Second, although we detected correlations between the MRI indexes and microbiota composition as well as diversity, we cannot determine the causal relationship between them. A future longitudinal study may contribute to solving this problem. Finally, the sample size was moderate. A larger independent sample is needed to examine the reproducibility of our findings.