New direction of digital twin application: analyzing infant development

Research led by the University of Chicago shows that AI technology-driven "digital twins" can model the baby's microbiome to predict possible neurodevelopmental problems that may occur later in the baby's growth. .

Using very early gut microbiome-related data from fecal samples of premature infants, the digital twin can very accurately predict their later microbiome composition and corresponding neurodevelopmental defects.

The paper was published in the journal Science Exhibition. "We simply looked at a snapshot of the microbiome and analyzed various types of bacteria," lead author Ishanu Chattopadhyay from the University of Chicago said in a statement. This is because the microbiome continues to change and mature during early childhood.”

Therefore, we developed a method to use generative AI to build a microbiome. A new approach to digital twins of systems capable of simulating interactions as bacterial populations change.

The research is still in its early stages, but if validated, the team believes it could help predict which babies may need early microbiome transplants to help improve their neurodevelopment.

In the discussion, the authors pointed out, “As more evidence shows that microbial imbalance can lead to the occurrence and development of a variety of diseases, including affecting basic digestive processes and even accessing the microbiota-gut-brain line TRAVEL CENTRAL NERVOUS SYSTEM ”

Natural Science has observed the role of the microbiome in human brain development, including in preterm infants, and the relationship between microbial dysbiosis and neuroinflammation and neurodevelopmental disorders, but The specific mechanism of its glial brain axis operation is still a mystery that has not been completely revealed.

To advance exploration in this area, Chattopadhyay and colleagues used 16S ribosomal RNA profiles from 398 samples taken from 88 premature infants to guide and train a digital model. The data provided by the babies, some with neurodevelopmental problems and others with no symptoms, allowed the AI to learn how to predict potential developmental problems in newborns.

The research team found that Digital Kids was able to predict the risk of developmental delays and poor head circumference growth, with a rate of correct observation of subject characteristics as high as 76%. The positive prediction accuracy rate at 30 weeks was 95%, and the specificity prediction accuracy rate was 98%.

The researchers calculated that early microbiome transplantation can help about 45% of infants avoid developmental problems, but the specific situation must be further verified in future work, especially the possibility of incorrect supplementation of microbiota. negative effects.

Chatopadhyay explained, "We cannot expect to reduce developmental risks by simply giving probiotics. The baby's microbiome is very important, and supplementation needs to be precisely controlled from multiple angles."

The researchers also mentioned that the digital twin model may focus research on specific conditions and treatment targets in the gut microbiome in the future. Compared with existing research methods, it is expected to significantly shorten the development cycle of diagnosis and treatment plans.