Optimization of the Microwave-Assisted Extraction of Total Phenolic Compounds (TPCs) From Almond Skins Through Artificial Neural Networks (ANNs) and Assessment of the Antioxidant and Antihyperglycemic Activity of the Extracts

Abstract

Optimization by response surface methodology (RSM) and artificial neural networks (ANNs) was efficaciously applied to study the operating parameters of microwave-assisted extraction (MAE) in the recovery of total phenolic compounds (TPCs) from almond skins. These models were used to evaluate the effects of process variables and their interaction towards the attainment of their optimum conditions. A comparison of statistical parameters showed that ANN was more consistent (R2 = 0 99) than RSM (R2 = 0 97) to predict a TPC by MAE. Therefore, the following conditions were proposed: microwave power of 562W, extraction time of 30 s, and ethanol concentration of 53%, corresponding to an optimal TPC yield of 560.79 mg gallic acid equivalents (GAEs)/ 100 g of dry weight (DW). The almond skin extract exhibited a high antioxidant activity tested by 1,1-diphenyl-picrylhydrazyl (DPPH) radical scavenging activity (IC50 = 5 39 ± 0 35 μg/mL), phosphomolybdate ammonium essay, hydroxyl radical scavenging activity (IC50042002mg/mL), and ability of chelating ferrous ions. The in vitro antihyperglycemic activity test revealed that the almond skin extract inhibits strongly α-amylase activity with IC50 = 27 87 μg/mL which was close to IC50 of the therapeutic drug acarbose (IC50 = 14 24 μg/mL).