Rapid and accurate detection of Dendrobium officinale adulterated with lower-price species using NMR characteristic markers integrated with artificial neural network

Dendrobium officinale (D. officinale) as a well-known herbal and functional food material is often adulterated with lower-price Dendrobium species. In this study, we aimed to develop an integrated method of nuclear magnetic resonance spectroscopy and artificial neural network (NMR-ANN) to identify and quantify the adulteration of D. officinale powder with other cheaper species. Microwave-assisted water extraction was selected as a time-saving and green method for sample preparation. The results indicate that the NMR-ANN method effectively quantified the adulteration rate of D. officinale powder with the root mean squared error (RMSE) of 4.92, mean absolute error (MAE) of 3.56 and coefficient of determination (R2) of 0.98 in the model test phase and with the RMSE of 7.65, MAE of 6.30 and R2 of 0.98 in the double-blinded test phase. The whole evaluation process can be done in 6 min and 5 s. Therefore, the NMR-ANN method can be used as a rapid, green and accurate tool for evaluating the quality of D. officinale or even other food materials.