Optimization of the extraction methods of Costus pictus rhizomes for a phytochemical-enriched antioxidant extract utilizing response surface methodology and artificial neural network

Plants were tapped for various applications due to their plethora of bioactive compounds. The bioextracts market has grown significantly in recent decades, facilitating the shift toward a bio-centric economy. It was the first study on optimizing extraction techniques for Costus pictus rhizomes and their modeling by “response surface methodology (RSM)” and “artificial neural network (ANN).” Dried C. pictus rhizomes were extracted using water by “microwave-assisted extraction (MAE),” “solvent extraction,” “ultrasound-assisted extraction (UAE)” and “Soxhlet extraction,” and UAE extracts demonstrated better phenolics and flavonoid contents with significant antioxidant properties. “Fourier transform infrared (FTIR) spectroscopy” results justified the preponderance of UAE extracts. “Central composite design (CCD)” in RSM was utilized to ameliorate the extraction efficacy of UAE extracts. Extraction was also modeled using an ANN, and optimal parameters were determined at a 1:20 (w/v) sample-to-solvent ratio, 3 min of extraction and 10 s:10 s pulsing. This extract exhibited optimum antioxidant properties: DPPH radical scavenging assay (10.112 ± 0.09 µg/mL), ABTS cation decolorization assay (127.368 ± 0.025 µg/g), ferric reducing antioxidant power assay (686.139 ± 0.025 mM/g) and phosphomolybdate assay (16.627 ± 0.01 mg/g). The extracts contained 184.321 ± 0.07 mg/g and 276.103 ± 0.01 mg/g of phenolic and flavonoid contents, respectively. The efficacy of RSM and ANN was compared based on statistical indicators, and ANN had better predictability. Liquid chromatographic analysis results revealed the occurrence of phenolic acids and flavonoids. Therefore, green extraction of C. pictus rhizome with water through UAE promises sustainable utilization of this medicinal plant in pharmaceutical and food formulations.