Leveraging artificial intelligence for analysis of the gut microbiome for dementia diagnosis: a scoping review and...

Dementia, a multifactorial disease with progressive cognitive decline, has been linked to imbalances in the gut-brain axis. Emerging artificial intelligence tools have augmented the identification of several gastrointestinal biomarkers for differential dementia detection and severity, but current literature lacks a comprehensive review.

This study aims to better quantify the applications of AI in the exploration of the gut microbiome for diagnosis of specific subtypes of dementia.

Primary articles (n = 896) from any point in time through July 2025 were identified from PubMed, Web of Science, Scopus, and ScienceDirect. Title and abstract screenings filtered articles from 896 to 28 for critical appraisal and review for key bacteria, fungi, and metabolites. Methods adhered to the PRISMA extension for Scoping Reviews (PRISMA-ScR) guidelines.

Several studies utilized predictive models including random forests and neural networks to demonstrate alterations in the gut microbiota of Alzheimer's disease, an increasingly prevalent dementia subtype. These individuals have notably reduced levels of butyrate-producing bacteria, such as Butyrivibrio, Eubacterium, and Faecalibacterium, which contribute to anti-inflammation and gut-barrier maintenance. Increased levels of Odoribacter splanchnicus and Klebsiella pneumoniae, as well as bacteria from generas Bacteroides and Prevotella, which help generate short-chain fatty acids, have been implicated in neuroinflammation; Roseburia inulinivorans negatively correlates with functional ability. Interestingly, superagers also display unique microbiome profiles that seemingly have protective effects linked to superior cognitive resilience.

Distinct gut microbial compositions are associated with dementia. Furthermore, elucidating gut-brain interactions and their neurodegenerative implications can identify targets for earlier, synergistic diagnostics.

https://osf.io/yw2dc/overview.