Biomimic models for in vitro glycemic index: Scope of sensor integration and artificial intelligence

Abstract

The accurate quantification of glycemic index (GI) remains crucial for diabetes management, yet current methodologies are constrained by resource intensiveness and methodological limitations. In vitro digestion models face challenges in replicating the dynamic conditions of the human gastrointestinal tract, such as enzyme variability and multi-time point analysis, leading to suboptimal predictive accuracy. This review proposes an integrated technological framework combining non-enzymatic electrochemical sensing with artificial intelligence to revolutionize GI assessment. Non-enzymatic sensors offer superior stability and repeatability in complex matrices, enabling real-time glucose quantification across multiple timepoints without enzyme degradation constraints. Machine learning algorithms, both supervised and unsupervised, enhance predictive accuracy by elucidating complex relationships within digestion data. This technological convergence represents a paradigm shift in food science analytics, promising improved throughput and precision in GI assessment. Future developments should focus on system scalability and broader applications across nutritional science, advancing diabetic management and personalized nutrition strategies.

Keywords

• Glycemic index
• In vitro models
• Starch hydrolysis
• Electrochemical sensor
• Artificial intelligence