Modeling of Continuous Ultrasonication to Improve Total Phenolic Content and Antioxidant Activity in Sorghum Flour: A Comparison between Response Surface Methodology and Artificial Neural Network
Abstract Fermentation followed by continuous ultrasonication was applied to release the bound phenolics in sorghum flour (SF). Total phenolic content (TPC) and antioxidant activity (AA) increased with decrease in fermentation time (FT), flour to water ratio (FWR), flow rate (FR) and ultrasonication intensity (UI). The influence of process variables was investigated by Box–Behnken design and multi-layer perceptron neural network. The optimum conditions for maximum TPC and AA were obtained as 12 h FT, 10 % (w/v) FWR, 20 W/cm2 UI, 4 ml/s FR and 120 s UT. The values observed for TPC and AA at optimum conditions were 90.1 mg GAE/100 g dm and 190.1 µmol TE/100 g dm, respectively, while these values for control SF were observed as 63.9 mg GAE/100 g dm and 133.5 µmol TE/100 g dm. Ultrasonication improved the free phenolic acid content by releasing bound phenolics in SF. The ANN model prediction was more precise compared to the RSM model.