Optimization of sequential smashing tissue-microwave assisted extraction and resin purification of salvianolic acids from Salviae miltiorrhizae Radix et Rhizoma using a hybrid RSM-ANN-GA model

This study developed a sequential smashing tissue-microwave assisted extraction (ST-MAE) method to extract salvianolic acids (SAs) from Salviae miltiorrhizae Radix et Rhizoma (SMR). Macroporous resin purification further enhanced the extract’s biological activity. Response surface methodology (RSM) and an artificial neural network embedded in a genetic algorithm (ANN-GA) optimized the extraction and purification parameters. Extraction kinetic studies revealed that the Weibull model accurately described the time-dependent behavior of 10 major SAs, reflecting efficient diffusion and compound stability under optimized conditions. Compared with RSM, the hybrid ANN-GA model significantly improved the antioxidant activity of obtained products (p < 0.05). The ST-MAE method optimized by ANN-GA increased SAs yield by 2.03, 1.95, and 1.90 fold compared to smashing tissue extraction, microwave extraction, and heating reflux extraction, respectively. ST-MAE extract showed superior antioxidant activity in DPPH, ABTS, FRAP, and ORAC assays. After purification, the SAs content increased from 18.49 % to 65.87 %, and the antioxidant activity was significantly enhanced. In conclusion, this study provides an efficient approach to produce antioxidant-rich SAs from SMR, with the ST-MAE method and the hybrid optimization strategy demonstrating significant advantages.