Construction of a multidimensional quality evaluation system for Scutellariae Radix: Integrated research on chemical components, biological potency, near-infrared spectroscopy technology, and intelligent algorithms.

Scutellariae Radix (SR), a traditional herbal medicine with significant medicinal value, is widely used worldwide. To address the quality differences caused by its various commercial specifications, processing methods, and provenances, a multi-dimensional quality evaluation system was established based on content, biological potency, and spectroscopy, compensating for the inadequacy of using baicalin as sole indicator for assessing SR quality. UHPLC-Q-TOF-MS/MS technology was employed to systematically identify the chemical profiles of SR and confirm differential components among multi-source samples. The varying distribution of flavone glycosides and their aglycones in the samples was identified as the primary reason for differences in commercial specifications and processing methods. Based on this, ten marker compounds with high content and significant bioactivity were selected for quantitative analysis using UHPLC-QqQ-MS/MS. Six components are proposed as the content markers for each sample category, with corresponding limits established. Additionally, the anti-inflammatory activity of the samples was evaluated through biological potency assays, and the potency significantly increased after SR processing. To achieve real-time and efficient detection, an innovative technology combined near-infrared spectroscopy (NIR) with intelligent algorithm was developed. After preprocessing with SG+ SNV + MSC, and selecting characteristic wavelengths by SPA, the GWO-SVM model achieved 100% accuracy in sample classification. Meanwhile, after sequential optimization, the PLSR model reliably predicting the concentration of active constituents and anti-inflammatory potency. This study has accomplished a shift in quality control of SR, moving from single-component quantitative analysis to comprehensive evaluation combining active ingredient detection and bioactivity assessment, while also advancing from precise measurement to rapid predictive methods.