Abstract
Gut microbial alteration is closely associated with brain disorders including cognitive impairment (CI). Gut microbes have the potential to predicate the development of diseases. However, the gut microbial markers for CI remain to be elucidated. In this study, the gut microbial alterations were assessed using16S rRNA sequencing, and identified the gut microbial markers using a random forest model. The results showed that there were significant gut microbial differences between the control and CI groups based on beta diversity (p < 0.002). Patients with CI had higher abundances of Actinobacteria and Proteobacteria but lower proportions of Bcateroidetes and Firmicutes vs. that in the control group. Patients had 39 special genera and the control subjects had 11 special genera. Furthermore, 11 genera such as Blautia, Roseburia, and Lactococcus and 18 genera such as Lactobacillus, Ruminococcus 2, and Akkermansia were the differential taxa in the control and CI groups, respectively. Gene functions related to nutrient metabolisms were upregulated in patients with CI. This suggested that the huge differences in gut microbes between the two groups and gut microbiota had the potential to predicate the development of CI. Based on machine learning results, 15 genera such as Lactobacillus, Bifidobacterium, and Akkermansia were selected as the optimal marker set to predicate CI with an area under curve (AUC) value of 78.4%. The results revealed the gut microbial markers for CI and provided a potential diagnosis tool to prevent the development of CI in the elderly.