Alzheimer's disease (AD) is one of the most common diseases in elderly people with a high incidence of dementia at approximately 60-80%. The pathogenesis of AD was quite complicated and currently there is no unified conclusion in the academic community, so no efficiently clinical treatment is available. In recent years, with the development of traditional Chinese medicine (TCM), researchers have proposed the idea of relying on TCM to prevent and treat AD based on the characteristic of multiple targets of TCM. This study reviewed the pathological hypothesis of AD and the potential biomarkers found in the current researches. And the potential targets of berberine and evodiamine from Evodia rutaecarpa in AD were summarized and further analyzed. A compound-targets-pathway network was carried out to clarify the mechanism of action of berberine and evodiamine for AD. Furthermore, the limitations of current researches on the TCM and AD were discussed. It is hoped that this review will provide some references for development of TCM in the prevention and treatment of AD.

The chemical fingerprinting and metabolite profile in a rat plasma sample after intragastric administration of Yangyin qingfei decoction (YYQFD, 14 g/kg) were investigated. First, YYQFD was analyzed by UPLC/Q‐TOF MS to establish the chemical composition database by comparing their retention behavior, accurate molecular mass and MS2 data with those of references or known compounds in the literature. In this database, 100 chemical constituents with information on retention time, molecular mass, molecular formula, MS2 data and compound name were identified, which can provide compound information for further metabolite profiling studies. Furthermore, 64 compounds including 37 prototypes and 27 metabolites were detected in the dosed rat plasma sample, and the metabolic pathways of YYQFD were hydrolyzation, hydroxylation, dehydrogenation, glucuronidation, glucosylation, sulfation and mixed modes. Among the five component herbs in the YYQFD, Glycyrrhizae Radix et Rhizome and Fritillariae Thunbergii bulbs were actively metabolized, contributing 16 and 7 metabolites, respectively. It is suggested that chemical characterization and metabolite profiling studies are valuable to elucidate the material basis of herbal preparations.

Alzheimer’s disease (AD) is a common and serious neurodegenerative disease in the elderly; however, the treatment of AD is still lacking of rational drugs. In this paper, the active constituents and targets of the self-developed Chinese medicine Formula 9002A in the treatment of AD were investigated from three aspects: pharmacodynamics based on cell and animal experiments, network pharmacology analysis, and pharmacokinetic analysis. A total of 124 compounds were screened in Formula 9002A, and four constituents including salidroside, gastrodin, niacinamide, and umbelliferone were screened as potential active components for the treatment of AD by network pharmacology. Among them, salidroside and gastrodin showed higher relevance with AD targets, such as ESR1 and AR. The pharmacokinetic study showed that they could be absorbed and identified in plasma; the half-life and mean residence times of salidroside and gastrodin in plasma were nearly increased 2-fold by the administration of Formula 9002A compared with those by the administration of a monomer, indicating the extended action time of active compounds in vivo. Formula 9002A exerted the efficacy in the treatment of AD mainly by regulating APP, GSK3β, ESR1, and AR targets based on the anti-β-amyloid protein deposition, anti-oxidation and anti-apoptosis pathways. Two genes enriched in Alzheimer’s disease pathway, APP and GSK3β, were further validated. The experiments also demonstrated that Formula 9002A could downregulate APP and GSK3β protein expression in the model mice brain and improved their cognitive ability. In summary, Formula 9002A has the characteristics of multiple targets and multiple pathways in the treatment of AD, and salidroside and gastrodin might be the main active constituents, which could provide a foundation for further research and application.

3D human pose estimation is an important and challenging task in computer vision. In this paper, we propose a method to estimate 3D human pose from RGB-D images. We adopt a 2D pose estimator to extract color features from the RGB image. The color features are integrated with the depth image in the form of point cloud. To fully exploit geometric information, we design a 3D learning module to extract point-wise features. To take advantage of local information as well as facilitate the convergence of the model, we design a dense prediction module. It estimates the offset vectors and closeness scores from points to target keypoints. The point-wise estimations are weighted and summed up to a final 3D pose. Experimental results show that our method achieves state-ofthe-art performance on MHAD and SURREAL datasets.