Logo image
Back
Predicting Dietary Impact on Multiple Sclerosis-Related Symptoms With the Gut Microbiome: A Pilot Study Using Unsupervised Machine Learning
Journal article   Open access   Peer reviewed

Predicting Dietary Impact on Multiple Sclerosis-Related Symptoms With the Gut Microbiome: A Pilot Study Using Unsupervised Machine Learning

Leeann Aguilar Meza, Rachel L Fitzjerrells, Farnoosh Shemirani, Tyler J Titcomb, Linda M Rubenstein, Patrick Ten Eyck, Linda G Snetselaar, Shailesh K Shahi, Terry L Wahls and Ashutosh K Mangalam
Brain and behavior, Vol.16(4), e71394
04/2026
DOI: 10.1002/brb3.71394
PMCID: PMC13093897
PMID: 42007545
Appears in  UI Libraries Support Open Access
url
https://doi.org/10.1002/brb3.71394View
Published (Version of record) Open Access

Abstract

Multiple sclerosis (MS) is a neurodegenerative disease where dietary intervention has emerged as a potential adjunct treatment. Recently, the modified Paleolithic elimination (MPE) diet, also known as the Wahls diet, and the low-saturated fat (LSF) diet, also known as the Swank diet, were linked to reduced fatigue and improved quality of life (QoL) in the WAVES study (NCT02914964). However, how diet impacts these outcomes remains unclear. As diet impacts gut microbiota, we investigated whether the baseline gut microbiota can predict response to diet in people with MS (pwMS). We performed fecal 16s rRNA sequencing to profile the microbiome changes associated with pwMS receiving the MPE (n = 11) and LSF diet (n = 12). Next, we utilized topic modeling, a machine learning technique, to determine whether baseline microbiome features predicted diet response in the combined MPE + LSF dietary cohort (n = 23). Specific genera significantly differed over time on both diets. On the MPE diet, Hungateiclostridiaceae, Ruminiclostridium, and Shuttleworthia decreased, while Coriobacteriaceae Collinsella decreased on LSF. Predictive machine-learning analysis associated a baseline microbiome enriched with Akkermansia, Bacteroides, and Barnesiella with fatigue response in the combined MPE + LSF cohort. For a non-response in Mental QoL improvement in the combined MPE + LSF cohort, our analysis associated an enrichment of Faecalibacterium and Alistipes at the start of the diet. Utilizing topic modeling, this pilot study identified baseline microbiota communities linked to improvements in fatigue and Mental QoL in pwMS on dietary intervention. These findings highlight the microbiota's role in dietary response and the potential for personalized nutrition. Given the small cohort and exploratory design, the results are hypothesis-generating and require validation in larger mechanistic studies.
 Quality of Life Adult Diet Diet, Paleolithic Feces - microbiology Female Gastrointestinal Microbiome - physiology Humans Male Middle Aged Multiple Sclerosis - diet therapy Multiple Sclerosis - microbiology Pilot Projects RNA, Ribosomal, 16S Unsupervised Machine Learning UIOWA OA Agreement

Details

Metrics

5 Record Views
Logo image