With the booming development of artificial intelligence (AI), machine learning, deep learning and transfer learning have gained growing applications in healthcare, bringing new perspectives and strategies to antidepressant research. Existing antidepressants present slow onset, low efficacy and marked inter-individual variability, with their underlying mechanisms yet to be fully clarified. Traditional research is hampered by small sample sizes, substantial data heterogeneity and poor cross-modal integration capacity. As a pivotal AI technology, transfer learning realizes cross-domain, cross-task and cross-modal knowledge transfer, and mitigates issues such as small-sample modeling, data distribution discrepancy and the curse of dimensionality. This review systematically elaborates the current status and future trends of transfer learning in antidepressant efficacy evaluation and the exploration of their action mechanisms. We analyze the characteristics and performance of transfer learning models construction strategies and application paradigms, and discuss the challenges, technical limitations and clinical translation barriers in practical application. This study provides novel ideas for exploring depression pathogenesis and optimizing clinical antidepressant use, and further promotes the innovative and standardized application of AI in clinical medicine.

Aristolochic acid analogues (AAAs) are a class of nitrophenanthrene derivatives predominantly occurring in Aristolochiaceae plants. Over the past decades, academic research has mostly focused on its strong adverse reactions such as nephrotoxicity and carcinogenicity. In contrast, their potential pharmacological activities—including anti-inflammatory, analgesic, and antitumor effects—have remained relatively underexplored, resulting in an imbalanced understanding of this compound class. In the face of the dual characteristics of toxicity and efficacy of this type of component, how to balance its toxicity and medicinal value is the key. Methods such as processing, compatibility, and structural modification may reduce adverse reactions, which could provide guarantees for the safe use of such components and traditional Chinese medicine. This review integrates current global advances in AAAs research, with a particular focus on their pharmacological activities and detoxification strategies, to provide references for the clinical application of drugs containing AAAs.

Objective To study the protective effect and potential mechanism of soybean isoflavones (SI) on ischemic stroke. Methods SI were isolated from soybeans and components were analyzed. An oxygen glucose deprivation/reperfusion (OGD/R)-induced PC12 cells model was constructed to investigate the effect of SI on cell survival, migration ability, apoptosis, intracellular Ca²⁺, reactive oxygen species (ROS), lactate dehydrogenase (LDH), cysteine aspartate protease-3 (Caspase-3) levels, and expressions of Kelch-like ECH associated protein 1 (Keap1)/nuclear factor erythroid 2-related factor 2 (Nrf2) pathway related proteins. A rat model of global cerebral ischemia-reperfusion injury was established using bilateral common carotid artery ligation. The effect of SI on neurological function score, brain tissue water content, proportion of cerebral infarction volume, pathological changes, superoxide dismutase (SOD), glutathione peroxidase (GSH-Px), catalase (CAT), LDH activities, malondialdehyde (MDA) level, and expressions of Keap1/Nrf2 pathway related proteins in brain tissue were investigated. Results In vitro experiments showed that compared with model group, SI significantly increased cell survival rate and migration rate (P < 0.05, 0.01), decreased cell apoptosis rate and intracellular Ca²⁺, ROS, LDH, Caspase-3 levels (P < 0.05, 0.01), upregulated Nrf2, heme oxygenase-1 (HO-1), NAD(P)H quinone oxidoreductase 1 (NQO1) protein expressions (P < 0.01), and downregulated Keap1 protein expression (P < 0.05, 0.01). In vivo experiments, compared with model group, SI significantly improved neurological deficits, reduced brain edema and infarct volume in rats (P < 0.01), repaired brain tissue pathological damage, reduced MDA level and LDH activity in brain tissue (P < 0.01), increased SOD, CAT and GSH-Px activities in brain tissue (P < 0.01), upregulated Nrf2, HO-1 and NQO1 protein expressions in brain tissue (P < 0.01), and downregulated Keap1 protein expression in brain tissue (P < 0.01). Conclusion SI have a significant neuroprotective effect on ischemic stroke, which may be related to the activation of Keap1/Nrf2 pathway, enhancing the antioxidant capacity of body, inhibiting oxidative stress damage, reducing neuronal apoptosis and alleviating brain tissue damage.

Objective To establish an accurate and efficient molecular identification method for specific DNA Signature Sequence Tags (DSS) of Tianshanxuelian (Saussurea involucrate) and its adulterants Shuimuxuetuzi (S. medusa) and Miantouxuetuzi (S. laniceps). Methods The Chloroplast Genome Information Resource (CGIR) were used to obtain the complete chloroplast genome sequences of Saussurea DC, the IdenDSS software was used to screen for specific DNA signature sequence tags (DSS) of the S. involucrata. Five sets of DSS markers were selected, and the corresponding specific primer pairs were designed using IdenDSS. Through DNA extraction from the sample, PCR amplification, Sanger sequencing and sequence alignment verification, the presence of DSS or its reverse complementary sequence of the target species S. involucrata in the test sample was detected for original identification. Results By comparing the selected five pairs of DSS sequences with the authentic source and its adulterants sequences of S. involucrata, all five sets of DNA Signature Sequence Tags could be used for identification of S. involucrata were finally determined. The above sequences were uploaded to the standard database of the Traditional Chinese Medicine Molecular Identification Platform (www.herbsdna.com), and combined with the DSS identification function of the database, accurate and efficient identification of S. involucrata origin can be achieved. Conclusion The five selected DNA Signature Sequence Tags had strong specificity and high accuracy, which can effectively solve the problem of original identification of S. involucrata.

The incidence and mortality of pulmonary diseases remain persistently high. Conventional therapeutic modalities encounter restrictions, including drug resistance and notable side effects, underscoring the urgent need for novel treatment strategies. Nucleic acid drugs, with their strong target specificity and long-lasting effects, provide a promising new direction for treating pulmonary diseases. Meanwhile, traditional Chinese medicine (TCM), with its extensive theoretical and practical foundation the prevention and treatment of pulmonary diseases, offers innovative intervention strategies through its derived nucleic acid components, which can function via trans-kingdom regulatory mechanisms. This paper explores the potential of TCM-derived nucleic acid drugs as an emerging trans-kingdom therapeutic strategy. Guided by TCM theory, it systematically elaborates drug, type, mechanisms of action, and advances in areas such as drug stability and delivery efficiency, as well as the unique advantages of these agents. Furthermore, the article reviews the application status of nucleic acid drugs in various pulmonary diseases, including chronic obstructive pulmonary disease, asthma, pneumonia, lung cancer, acute lung injury, pulmonary fibrosis, and acute respiratory distress syndrome. Finally, the discussion then addresses the current challenges and future prospects to inform the theoretical foundation and research ideas for the systematic development of this drug class.

This paper reviews the traditional application basis and modern research progress of Kushen (Sophorae Flavescentis Radix), and performs a comprehensive analysis focusing on its chemical constituents, shared common pharmacological mechanisms, and safety profiles, thereby providing a scientific reference for its research and standardized application. Using keywords such as Sophorae Flavescentis Radix, alkaloids, total flavonoids, immune-inflammatory regulation, and safety evaluation, this review collected relevant domestic and foreign relevant literature published in the past decade was retrieved from databases including Wanfang, China National Knowledge Infrastructure, PubMed, and Web of Science. The retrieved studies were systematically analyzed to summarize the traditional medical understanding and modern pharmacological research, with special emphasis on the pharmacodynamic material basis and clinical application mechanisms of Sophorae Flavescentis Radix. Sophorae Flavescentis Radix contains multiple classes of chemical components, predominantly alkaloids, while flavonoids, polysaccharides, and other constituents also contribute to its holistic efficacy. Modern pharmacological studies have revealed that Sophorae Flavescentis Radix exerts its overall effects through immune-inflammatory regulation as the core mechanism, specifically by targeting desmoglein-2 for barrier repair and bidirectionally modulating immune polarization (M1/M2). Its material basis exhibits clear structure-synergy patterns, non-alkaloid components can regulate the toxicity threshold of alkaloids by inhibiting the reactive oxygen species/NOD-like receptor family pyrin domain containing 3 pathway, demonstrating a significant “efficacy-toxicity” balance feature. Meanwhile, adverse reactions related to Sophorae Flavescentis Radix are closely associated with its alkaloid components, dosage, and administration routes, suggesting that the safety risks are manageable to a certain extent. Modern research on Sophorae Flavescentis Radix has, to some extent, validated its traditional efficacy basis. However, critical issues remain, such as unclear relationships among components, insufficient integration of mechanisms, and limited systematic safety evaluations. Future research should strengthen studies on multi-component synergistic mechanisms and safety risk control to promote the rational development and standardized clinical application of Sophorae Flavescentis Radix.

Objective To evaluate the quality of Fagopyri Dibotryis Rhizoma Tablets (FDRT) produced by different companies. Methods A total of 18 batches of FDRT produced by company A and 20 batches of FDRT produced by company B were collected, respectively. The fingerprints of 38 batches FDRT were determined using HPLC, the common peaks were assigned and the constituents of these peaks were identified by quadrupole time-of-flight mass spectrometry (Q-TOF-MS/MS). Using quantitative common peak areas as variables, pattern recognition analysis of the fingerprint was conducted using hierarchical cluster analysis (HCA), principal component analysis (PCA), and orthogonal partial least squares discriminant analysis (OPLS-DA). The content of the nine constituents confirmed by comparison with reference substances and have good separation in samples were determined. Results There were 33 common peaks in the fingerprint of 38 batches of FDRT, and a total of 53 constituents were identified by Q-TOF-MS/MS, including 21 tannins, 12 phenolics, seven flavonoids, seven phenylpropanoid glycosides, two amino acids, two organic acids, one alkaloid, and one terpenoid, among them, three phenolics (gallic acid, protocatechuic acid, and protocatechualdehyde), five tannins (procyanidin B1, B2, B3, C1, C2), four flavonoids (catechin, epicatechin, epicatechin gallate, and rutin) were confirmed by reference substance. The separation degree of nine constituents except for procyanidin C2, B2, and rutin was relatively good. The average content of these nine constituents in FDRT produced by company B was significantly higher than that in FDRT produced by company A (P < 0.01). The results of HCA, PCA, and OPLS-DA were consistent, clearly distinguished the FDRT produced by company A and company B into two categories. The constituents corresponding to common peaks 30, 29, 16, 17, 2, 21, 10, 11, 18, 20, and 27, such as procyanidin C2, B2, C1, B1, B3, gallic acid, epicatechin, etc., are the marker constituents responsible for the quality differences in FDRT produced by two enterprises. Conclusion There is a significant quality difference in FDRT produced by two companies.

Objective To address the challenges of insufficient standardization in the processing of Orpiment (arsenic disulfide, As₂S₂) and the difficulty in controlling its toxicity, this study elucidated the interactions among “particle size-morphology-porosity-crystal defects-arsenic (As) release”. It analyzed the effects of different processing methods on elemental composition and microstructure, and established a rapid identification model based on multimodal spectral fusion. Methods A multi-dimensional “element-morphology-spectrum” data matrix was constructed. Inductively coupled plasma optical emission spectrometry (ICP-OES) was used to determine elemental content. Scanning electron microscopy (SEM) was employed to observe morphology and analyze surface porosity, combined with dissolution experiments to quantify As release. Furthermore, Fourier transform infrared spectroscopy (FTIR) and Raman spectroscopy (RS) data were integrated. Partial least squares-discriminant analysis (PLS-DA) and support vector machine (SVM) discrimination models were built using both low-level and mid-level data fusion strategies. Results The zirconia-ball water-grinding method reduced As content and increased sulfur (S) content in Orpiment. SEM analysis revealed that particles processed by this method exhibited a more concentrated size distribution, their morphology became more rounded and blunt, and surface porosity significantly increased, forming a hierarchical pore structure. In simulated gastric and intestinal fluids, As dissolution decreased by 72.8% and 81.4%, respectively, compared to dry crushing (P < 0.001). The SVM model based on mid-level spectral fusion achieved 100% classification accuracy. Conclusion The water-grinding method achieves detoxification by removing As³⁺ from Orpiment, regulating the particle morphology and pore structure of Orpiment particles, and enhancing the stability of Orpiment crystals. It establishes a multidimensional data matrix integrating elemental composition, microscopic morphology, pore structure, and spectral characteristics, providing efficient and reliable technical support for quality control of the mineral drug Orpiment.

Objective This study aims to systematically analyze the intellectual property landscape of traditional Chinese medicine (TCM) globally, with a focus on comparing the patent layout characteristics and competitive dynamics in five key countries: China, the United States, Japan, South Korea, and Germany. The goal is to provide strategic references for promoting the high-quality development and internationalization of China's TCM industry. Methods Leveraging the Patyee database, this research retrieved and obtained TCM-related patent data globally and from the aforementioned five countries. Employing a combination of quantitative and qualitative analysis methods, an in-depth examination was conducted from multiple dimensions, including patent application trends, geographical distribution, applicant types, core technological topics, and international layout strategies. Results The research findings indicate that despite a vast number of domestic TCM patent applications (approximately 360 000), the proportion of international applications filed via the Patent Cooperation Treaty (PCT) is relatively low (only 1.37%), revealing a significant deficiency in overseas layout. Regarding the distribution of technological topics, patents related to TCM quality control and standardization technologies are relatively scarce globally (including in China), leaving a gap in meeting the high standards of quality controllability required for modern industrialization and internationalization. In terms of applicant types, 44% of TCM patents in China are held by individuals, while enterprises account for only 34%, reflecting a pattern of individual-led inheritance with enterprise-driven R&D yet to be enhanced. In contrast, TCM-related patent activities in the United States, Japan, and Germany are predominantly driven by multinational pharmaceutical companies or consumer goods giants with complete industrial chains. Conclusion To achieve high-quality internationalization of TCM intellectual property, China urgently needs to establish a systematic overseas patent layout strategy, forming a patent protection system covering the entire industrial chain of R&D, quality control, production, and marketing. The core task is to transform domestic technological advantages into internationally recognized technical standards and to promote the transformation of TCM from a traditional empirical model to a modern, evidence-based innovation paradigm by strengthening collaborative innovation among industry, academia, research, and users.

Objective To establish a multi-index weighted optimization strategy for wine-processed Xiangfu (Cyperi Rhizoma) by integrating CRITIC weighting, Box-Behnken design-response surface methodology (BBD-RSM), and an artificial neural network (ANN), and define a robust operating space based on model prediction, thereby providing a reference for process parameter fixation and quality consistency control. Methods The contents of cyperotundone, α-cyperone, total flavonoids, and total volatile oil were used as comprehensive evaluation indices. Objective weights were assigned using the CRITIC method, and an overall desirability (OD) value was calculated. Based on single-factor experiments, a four-factor, three-level Box-Behnken design (BBD) was conducted with processing temperature, processing time, moistening time, and dosage as independent variables, and a regression model was established. An ANN prediction model was then developed, and the Garson algorithm was applied to interpret the relative contributions of factors. In addition, the Pareto non-dominated solution set was used to screen the process-parameter space to obtain a robust operating interval (model-predicted). Results The CRITIC-derived weights for cyperotundone, α-cyperone, total volatile oil, and total flavonoids were 0.274 9, 0.255 4, 0.253 4, and 0.216 3, respectively. The optimal processing conditions were determined as follows: processing temperature 140 ℃, processing time 19 min, moistening time 6.9 h, and dosage 32 g/L. The OD obtained from validation experiments was 0.655 1, RSD was 2.97%, which was close to the predicted value (0.653 3). The robust parameter interval defined based on ANN prediction and the Pareto non-dominated solution set was as follows: processing temperature 140—150 ℃, processing time 18—22 min, moistening time 6.5—7.5 h, and dosage 30—35 g/L. The model suggested that OD could remain relatively stable under ± 5% parameter fluctuations within this interval. Conclusion The constructed BBD-ANN-Pareto coupling model demonstrated good predictive ability and stability, which may provide a reference for optimizing the processing technology and ensuring the quality consistency of wine-processed Cyperi Rhizoma.