Hybrid modeling of ultrasound-assisted extraction of phytochemicals from Himalayan Ganoderma lucidum using artificial neural network-genetic algorithm

Résumé fourni par la source

Objective Ganoderma lucidum is a bioactive-rich traditional Chinese medicine widely recognized for its nutraceutical and pharmaceutical applications. This study optimized the ultrasound-assisted extraction (UAE) from Himalayan G. lucidum , applying artificial-neural network (ANN) modelling coupled with genetic algorithm (GA) optimization to enhance efficiency. Methods The UAE extraction was performed under various process parameters, including treatment duration ( i 1 , 5–15 min), ultrasonic amplitude ( i 2 , 30%−80%), and solvent concentration ( i 3 , 40%−80%), to optimize the yield ( w / w %, O 1 ), total triterpene content [mg/g, dry weight (dw), O 2 ], 2,2-diphenyl-1-picrylhydrazyl scavenging activity (%, O 3 ) and total phenolic content (mg/g dw, O 4 ). Results The ANN-GA approach accurately modeled the complex non-linear relationships between input and output parameters, identifying optimal conditions as 5 min ( i 1 ), 45.63% ( i 2 ) and 80% ( i 3 ). Experimental validations under these conditions resulted in (7.63 ± 0.2) %, (34.11 ± 0.01) mg/g dw, (41.55 ± 0.01) % and (10.98 ± 0.01) mg/g dw for O 1 , O 2 , O 3 , and O 4 respectively. Compared to response surface methodology (RSM) approach, which served only as a benchmark, the ANN-GA approach demonstrated superior predictive accuracy and robustness in optimization. Conclusion The ANN-GA approach provides an effective and robust strategy for optimizing herbal extraction processes. This study demonstrates the applicability of ANN-GA as an advanced optimization tool for herbal extraction processes and highlights G. lucidum as a valuable source of natural antioxidants and triterpenes with potential therapeutic relevance.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.