2018 Fiscal Year Annual Research Report
Social effects on epidemics: A temporal network approach
| Project/Area Number |
18H01655
|
|---|
| Research Institution | Tokyo Institute of Technology |
|---|
Principal Investigator |
Holme Petter 東京工業大学, 科学技術創成研究院, 特任教授 (50802352)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
村田 剛志 東京工業大学, 情報理工学院, 准教授 (90242289)
|
|---|
| Project Period (FY) |
2018-04-01 – 2022-03-31
|
| Keywords | epidemic outbreak / behavior change / temporal network / network epidemiology / theoretical epidemiology |
|---|
| Outline of Annual Research Achievements |
During FY 2018 I published 8 publications in international journals, four of these are directly related to the present KAKENHI project. One paper with RC Naoki Masuda (published in Scientific Reports) develops methods to analyze temporal networks. By our method, we can see what the state of a system is. For example, in a network of school children, we can say if they are attending a class, in recess, etc. This will be combined with the behavioral models we outline in the proposal to increase precision in disease outbreak prediction. Another paper with Korean colleagues (recently published in Rhys. Rev. E) discusses the modeling of behavioral changes through peer pressure and social influence (as we describe in the proposal (stage 2). Another paper in collaboration with Luis Rocha (Greenwich University) published in Network Science discusses models of temporal networks. In particular how robust predictions from temporal network disease spreading models are in presence of noise (something that is inevitable in practice). A fourth paper (published in New Journal of Physics) discussed models for extinction in small graphs. This is less related to the present KAKENHI project but contains results that can be used to validate and fine tune models developed in the project.
|
| Current Status of Research Progress |
Current Status of Research Progress
2: Research has progressed on the whole more than it was originally planned.
Reason
The current state of the project is almost as planned. At the moment, we are working together with co-PI’s Murata and RC Masuda on microscopic models of temporal network. This project is in the write-up phase and close to submission. I am also starting up a project of measure behavioral impact on health decisions with my student. This part of the project is ahead of the original plan whereas the modeling mentioned above is 6 months behind.
Some work is currently under review. One manuscript investigating the role of commuting on the spread of e.g. influenza (in collaboration with researchers from University of Texas) is currently with PLOS Computational Biology, another paper about how to use social mobility data for modeling of disease spreading-in particular about how to correct biases based on the methods of data collection (if it comes from transportation, mobile phones, etc.) with the same authors is also under review with the same journal. With this group, I am also investigating the impact of circadian patterns on disease spreading. Many people are, for example, more sensitive in the morning, which is exactly when they commute to work.
|
| Strategy for Future Research Activity |
We are currently finishing Phases 1 and 2 (that were due in mid FY2019. When this is done we will move on to study how the coevolution of disease and behavior changes results from temporal network epidemiology. By combining the finished models with standard compartmental models-like the susceptible-infectious-recovered model of infectious giving immunity upon recovery-we will be able to better understand how social processes affect disease spreading and models of temporal network epidemiology. We will explore different scenarios of social influence to evaluate methods to exploit temporal network structure in targeted vaccination and sentinel surveillance. We will also analyze how social processes, like awareness propagation and behavioral change, affects estimates of epidemic thresholds. In addition to simulations, we will gather online data to validate our models.
|
