Search

Horizon scanning for innovative technologies that might be applied to medical products and requires new assessment approaches to prepare regulators, allowing earlier access to the product for patients and an improved benefit/risk ratio. The purpose of this study is to confirm that citation network analysis and text mining for bibliographic information analysis can be used for horizon scanning of the rapidly developing field of AI-based medical technologies and extract the latest research trend information from the field. We classified 119,553 publications obtained from SCI constructed with the keywords "conventional," "machine-learning," or "deep-learning" and grouped them into 36 clusters, which demonstrated the academic landscape of AI applications. We also confirmed that one or two close clusters included the key articles on AI-based medical image analysis, suggesting that clusters specific to the technology were appropriately formed. Significant research progress could be detected as a quick increase in constituent papers and the number of citations of hub papers in the cluster. Then we tracked recent research trends by re-analyzing "young" clusters based on the average publication year of the constituent papers of each cluster. The latest topics in AI-based medical technologies include electrocardiograms and electroencephalograms (ECG/EEG), human activity recognition, natural language processing of clinical records, and drug discovery. We could detect rapid increase in research activity of AI-based ECG/EEG a few years prior to the issuance of the draft guidance by US-FDA. Our study showed that a citation network analysis and text mining of scientific papers can be a useful objective tool for horizon scanning of rapidly developing AI-based medical technologies. [ABSTRACT FROM AUTHOR]

Certain innovative technologies applied to medical product development require novel evaluation approaches and/or regulations. Horizon scanning for such technologies will help regulators prepare, allowing earlier access to the product for patients and an improved benefit/risk ratio. This study investigates whether citation network analysis and text mining of scientific papers could be a tool for horizon scanning in the field of immunology, which has developed over a long period, and attempts to grasp the latest research trends. As the result of the analysis, the academic landscape of the immunology field was identified by classifying 90,450 papers (obtained from PubMED) containing the keyword "immune* and t lymph*" into 38 clusters. The clustering was indicative of the research landscape of the immunology field. To confirm this, immune checkpoint inhibitors were used as a retrospective test topic of therapeutics with new mechanisms of action. Retrospective clustering around immune checkpoint inhibitors was found, supporting this approach. The analysis of the research trends over the last 3 to 5 years in this field revealed several candidate topics, including ARID1A gene mutation, CD300e, and tissue resident memory T cells, which shows notable progress and should be monitored for future possible product development. Our results have demonstrated the possibility that citation network analysis and text mining of scientific papers can be a useful objective tool for horizon scanning of life science fields such as immunology. [ABSTRACT FROM AUTHOR]

Despite the intensive study of the viral spread of fake news in political echo chambers (ECs) on social networking services (SNSs), little is known regarding the underlying structure of the daily information spread in these ECs. Moreover, the effect of SNSs on opinion polarisation is still unclear in terms of pluralistic information access or selective exposure to opinions in an SNS. In this study, we confirmed the steady, highly independent nature of left- and right-leaning ECs, both of which are composed of approximately 250,000 users, from a year-long reply/retweet network of 42 million Japanese Twitter users. We found that both communities have similarly efficient information spreading networks with densely connected and core-periphery structures. Core nodes resonate in the early stages of information cascades, and unilaterally transmit information to peripheral nodes. Each EC has resonant core users who amplify and steadily spread information to a quarter of a million users. In addition, we confirmed the existence of extremely aggressive users of ECs who co-reply/retweet each other. The connection between these users and top influencers suggests that the extreme opinions of the former group affect the entire community through the top influencers. [ABSTRACT FROM AUTHOR]

Conventionally, the importance of nodes in a network has been debated from the viewpoint of the amount of information received by the nodes and its neighbors. While node evaluation based on the adjacency relationship mainly uses local proximity information, the community structure that characterizes the network has hardly been considered. In this study, we propose a new node index that contributes to the understanding of the inter-community structure of a network by combining the measures of link distribution and community relevance. The visualization of node rankings and rank correlations with respect to the attack tolerance of networks demonstrated that the proposed index shows the highest performance in comparison with five previously proposed indexes, suggesting a new way to detect latent mediators in heterogeneous networks. [ABSTRACT FROM AUTHOR]