Time Evolution of the Stroke Symptom-Herb Networks Based on TCM Prescriptions

Traditional Chinese Medicine (TCM) has its origins in distant antiquity and has piled up over a long time with much knowledge about diseases, especially stroke. Different combinations of symptom variables yield different combinations of herbs to form a myriad of prescriptions, and they have undergone repeated confirmation and are worthy objects of excavation and analysis. Herbal studies on stroke have developed from genomics to transcriptomics, proteomics and metabolomics, yet more thought is needed on putting time in a wider perspective of the symptoms and herbs for stroke in Chinese medicine. Due to this, we studied the dynamic structure of TCM prescriptions on stroke, using over 270 TCM prescription books containing 2231 prescriptions related to stroke recorded from 341 to 2000 CE. We labeled the functions of the prescriptions with the symptoms based on subject terms in MESH Neurologic Manifestations, then standardized the herbs in the prescriptions, and finally connected the co-occurring symptoms and herbs in the prescriptions to build an undirected complex network. The Stroke Symptom-Herb Networks (SSHNs) can be seen from its network characteristics that it is not a random network and has small-world characteristics. It has experienced two peaks in its nearly 1700-year history, during the Song dynasty, the Ming and Qing dynasties. From 600 years onwards, the core herb cluster has been initially formed. The comparison of sub-network similarities allowed us to identify several symptoms with similar herb clusters. We divided the community based on modularity, and by analyzing the community evolution, we found a more fixed historical evolutionary trend with Hemiplegia and Sialorrhea nodes and their associated symptom and herb nodes. In the time series analysis, we found many symptom-herb combinations that were consistently closely related to historic time depends on assessing the similarity between the symptoms and the herbs. The complex network provides a distinctive perspective for understanding the symptom-herb relationships embedded in TCM prescriptions in remote antiquity.