目的 比较河南、河北、山东产地金银花挥发性成分、气味和味道的差异，为鉴别不同产地金银花提供参考。方法 基于顶空-气相色谱-质谱法（head space-gas chromatography-mass spectrometry，HS-GC-MS）及电子鼻和电子舌2种智能感官技术对不同产地金银花药材进行测定。以变量投影重要性（variable importance projection，VIP）＞1、P＜0.05为标准筛选出每个产地间的差异挥发性成分。对电子鼻与HS-GC-MS结果进行pearson相关分析，使用SPSS25软件建立不同产地金银花味道判别模型。结果 经HS-GC-MS分析从金银花中共鉴定出73个挥发性成分。筛选出河南与河北、河南与山东、河北与山东样品间差异挥发性成分分别为15、11、13个，其中己醛和丁香醛为3个产地的共有差异性挥发成分。电子鼻和电子舌均能较好区分不同产地的金银花样品。通过pearson相关分析确定了11种引起金银花气味差异的挥发性成分。另外，建立的味道判别函数模型可对不同产地金银花进行准确判别。结论 HS-GC-MS、电子鼻和电子舌技术结合多元统计分析可较好的区分不同产地的金银花，并阐明其挥发性成分、气味与味道的差异，对金银花的产地鉴别和质量评价具有一定指导意义。

Objective To compare and analyze the volatile components, odor and taste of Jinyinhua (Lonicerae Japonicae Flos) from Henan, Hebei and Shandong, so as to provide a reference for identifying Lonicera Japonica Flos from different producing areas. Methods The volatile components, odors and taste of Lonicerae Japonicae Flos of producing areas were determined based on headspace-gas chromatography-mass spectrometry (HS-GC-MS), E-nose and E-tongue. VIP > 1, P < 0.05 were used to screen out the differential volatile components between each producing area. A pearson correlation analysis was performed differences of Lonicerae Japonicae Flos samples from different producing areas, SPSS25 software was used to establish the taste discrimination model of Lonicerae Japonicae Flos from different producing areas. Results By HS-GC-MS analysis, a total of 73 volatile components were identified from Lonicerae Japonicae Flos. Fifteen, 11 and 13 volatile components were screened for differences between Henan and Hebei, Henan and Shandong, and Hebei and Shandong samples, respectively. Hexanal and Lilac aldehyde were the co-differential volatile components of the three producing areas. The E-nose and E-tongue could better distinguish the Lonicerae Japonicae Flos samples from different producing areas. Through pearson correlation analysis, 11 volatile components were screened as the main material basis for the difference in odor production. In addition, the established taste discriminant function model can accurately identify Lonicerae Japonicae Flos between different producing areas. Conclusion HS-GC-MS, E-nose and E-tongue techniques combined with multivariate statistical analysis can better distinguish Lonicerae Japonicae Flos from different origins and elucidate the differences in volatile components, odor and taste, which have certain guiding significance for its origin identification and quality evaluation.

R286.2

河北省省级科技计划项目资助（21372503D，23372502D）；河北省自然科学基金项目资助（H2022423341，H2022423335）