Control and prediction of degradation of biopolymer based hydrogels with poly(ɛ-caprolactone) subunits

Abstract

Complex hydrogels consisting of natural and synthetic polymers, stabilized by combining different physical and chemical cross-linking methods, were assessed by in vitro degradation tests performed at 37°C in phosphate buffer solution. Four biopolymer-based hydrogel series were comparatively evaluated as sponge or dense films as regards mass loss, morphology changes and thermal behavior over the fixed incubation period, considering as main factors of influence the composition, the adopted stabilization mode and the microstructure of the 3D construct. To facilitate the selection of the appropriate material for envisaged applications, matching the specific needs, the obtained data were used to generate an artificial neural network (ANN) model, able to establish correlations of the examined formulations and preparation parameters with mass loss. The comparison of experimental and calculated data showed that the developed ANN shows reasonable predictive performance (rtraining/validation(2)=0.99).

Keywords: Artificial neural network; Degradation; Hydrogel.