Neuro-fuzzy model for predicting insulin delivery from crosslinked agar-carbomer hydrogels

Abstract

This study focuses on the innovation of an inhaled sustained release form of insulin and the development of a neuro-fuzzy model specifically tailored to predict insulin release kinetics from polycondensed agar-carbomer hydrogels. These were synthesized by blending agar and carbomer, incorporating propylene glycol and glycerol, and then cross-linking by polycondensation. The structure and morphology of the hydrogel were analyzed via Fourier Transform Infrared Spectroscopy, Scanning Electron Microscopy and Proton Nuclear Magnetic Resonance Spectroscopy. The neuro-fuzzy model, a combination of artificial neural networks and fuzzy logic, employs inputs such as concentrations of crosslinking agents, polycondensation time, and release time, with the output being the rate of insulin release. The model demonstrated a strong correlation with experimental data, highlighting its effectiveness and precision in predicting insulin delivery from hydrogel compositions and temporal parameters. This emphasizes the importance of intelligent modelling for forecasting the kinetic release of therapeutic agents from novel drug delivery systems.