目的 针对传统“口尝微有麻舌感”评价草乌Aconiti Kusnezoffii Radix炮制终点的主观性和安全性问题，探究利用电子舌技术量化草乌炮制品麻舌感等级的可行性，建立基于味觉特征值的双酯型生物碱含量预测方法。方法 选取生草乌，按不同时间段炮制，制作不同麻舌感草乌样品，通过人工感官评价、电子舌味觉检测及HPLC法测定6种生物碱（双酯型生物碱及其单酯型衍生物）含量，结合多元统计分析方法，分析感官强度评分、生物碱含量与电子舌特征响应值的动态关联规律，建立基于电子舌特征响应值的草乌麻舌感等级评价和双酯型生物碱总量预测曲线方程。结果 随着炮制时间延长，麻舌感评分显著降低（r＝−0.85，P＜0.01），电子舌苦味值显著升高（r＝0.96，P＜0.01），新乌头碱、次乌头碱和乌头碱等双酯型生物碱含量逐步减少，苯甲酰新乌头原碱、苯甲酰乌头原碱和苯甲酰次乌头原碱等单酯型生物碱含量逐步增加；偏最小二乘法-判别分析（partial least squares-discriminant analysis，PLS-DA）结合单因素方差分析表明，苦味值（VIP值＞1，P＜0.01）是电子舌区分不同麻舌感草乌的重要味觉指标。电子舌苦味值与麻舌感评分（r＝−0.93，P＜0.01）、双酯型生物碱总量（r＝−0.78、P＜0.01）均呈现显著相关性。电子舌苦味值（X）与麻舌感评分（Y）进行曲线拟合，得到了方程式为Y＝−3.123 0 lnX＋5.915 4，R²＝0.937 3的对数回归方程；电子舌苦味值（X）与双酯型生物碱总量（Y）进行曲线拟合，得到了方程式为Y＝3.814 5 e^−0.760 X，R²＝0.875 5的指数回归方程，该方程预测结果良好。结论 电子舌苦味值可作为客观量化草乌炮制品麻舌感等级和预测双酯型生物碱含量的指标，为草乌炮制终点的判断提供了客观量化依据。

Objective In view of the subjectivity and safety of the traditional “tasting slightly hemp tongue” to evaluate the end point of processing of Caowu (Aconiti Kusnezoffii Radix, AKR), the feasibility of using electronic tongue technology to quantify the grade of hemp tongue of processed products of AKR was explored, and a prediction method of diester alkaloid content based on taste characteristic value was established. Methods AKR was selected and processed according to different time periods to make different samples of AKR. The contents of six alkaloids (diester alkaloids and their monoester derivatives) were determined by artificial sensory evaluation, electronic tongue taste detection and HPLC. Combined with multivariate statistical analysis method, the dynamic correlation between sensory intensity score, alkaloid content and electronic tongue characteristic response value was analyzed, and the evaluation of AKR’s tongue feeling grade and the prediction curve equation of total diester alkaloids based on electronic tongue characteristic response value were established. Results With the prolongation of processing time, the score of Ma-tongue sensation decreased significantly (r = −0.85, P < 0.01), the bitterness value of electronic tongue increased significantly (r = 0.96, P < 0.01), the content of diester alkaloids such as mesaconitine, hypaconitine and aconitine decreased gradually, and the content of monoester alkaloids such as benzoylmesaconine, benzoylaconitine and benzoylhypaconine increased gradually. Partial least squares-discriminant analysis (PLS-DA) combined with one-way analysis of variance showed that the bitterness value (VIP value > 1, P < 0.01) was an important taste index for the electronic tongue to distinguish different numb-tongue sensation of AKR. The bitterness value of electronic tongue was significantly correlated with the score of numbness (r = −0.93, P < 0.01) and the total amount of diester alkaloids (r = −0.78, P < 0.01). The logarithmic regression equation of Y = −3.123 0 lnX + 5.915 4, R² = 0.937 3 was obtained by fitting the bitterness value (X) of the electronic tongue with the numbness score (Y) curve. The exponential regression equation of Y = 3.814 5 e^−0.760 X, R² = 0.875 5 was obtained by fitting the electronic tongue bitterness value (X) with the total amount of diester alkaloids (Y), and the prediction result of the equation was good. Conclusion The bitterness value of electronic tongue can be used as an index to objectively quantify the level of numb-tongue sensation and predict the content of diester-type alkaloids in processed products of AKR, which provides an objective quantitative basis for judging the end point of processing of AKR.

R283.6

国家自然科学基金项目（82160790）；内蒙古医科大学蒙药学“一流学科”研究生科研能力提升计划项目（MYX2022-R13）；蒙药药味理论现代研究科研创新团队（2024MYYLXK005）