Privacy-Preserving Individual-Level COVID-19 Infection Prediction via Federated Graph Learning (preprint)

Accurately predicting individual-level infection state is of great value since its essential role in reducing the damage of the epidemic. However, there exists an inescapable risk of privacy leakage in the fine-grained user mobility trajectories required by individual-level infection prediction. In this paper, we focus on developing a framework of privacy-preserving individual-level infection prediction based on federated learning (FL) and graph neural networks (GNN). We propose Falcon, a Federated grAph Learning method for privacy-preserving individual-level infeCtion predictiON. It utilizes a novel hypergraph structure with spatio-temporal hyperedges to describe the complex interactions between individuals and locations in the contagion process. By organically combining the FL framework with hypergraph neural networks, the information propagation process of the graph machine learning is able to be divided into two stages distributed on the server and the clients, respectively, so as to effectively protect user privacy while transmitting high-level information. Furthermore, it elaborately designs a differential privacy perturbation mechanism as well as a plausible pseudo location generation approach to preserve user privacy in the graph structure. Besides, it introduces a cooperative coupling mechanism between the individual-level prediction model and an additional region-level model to mitigate the detrimental impacts caused by the injected obfuscation mechanisms. Extensive experimental results show that our methodology outperforms state-of-the-art algorithms and is able to protect user privacy against actual privacy attacks. Our code and datasets are available at the link: https://github.com/wjfu99/FL-epidemic.

Understanding internal migration in the UK before and during the COVID-19 pandemic using twitter data

The COVID-19 pandemic has greatly affected internal migration patterns and may last beyond the pandemic. It raises the need to monitor the migration in an economical, effective and timely way. Benefitting from the advancement of geolocation data collection techniques, we used near real-time and fine-grained Twitter data to monitor migration patterns during the COVID-19 pandemic, dated from January 2019 to December 2021. Based on geocoding and estimating home locations, we proposed five indices depicting migration patterns, which are demonstrated by applying an empirical study at national and local authority scales to the UK. Our findings point to complex social processes unfolding differently over space and time. In particular, the pandemic and lockdown policies significantly reduced the rate of migration. Furthermore, we found a trend of people moving out of large cities to the nearby rural areas, and also conjunctive cities if there is one, before and during the peak of the pandemic. The trend of moving to rural areas became more significant in 2020 and most people who moved out had not returned by the end of 2021, although large cities recovered more quickly than other regions. Our results of monthly migration matrixes are validated to be consistent with official migration flow data released by the Office for National Statistics, but have finer temporal granularity and can be updated more frequently. This study demonstrates that Twitter data is highly valuable for migration trend analysis despite the biases in population representation.

Activation of Nurr1 with Amodiaquine Protected Neuron and Alleviated Neuroinflammation after Subarachnoid Hemorrhage in Rats

Background. Nurr1, a member of the nuclear receptor 4A family (NR4A), played a role in neuron protection, anti-inflammation, and antioxidative stress in multidiseases. We explored the role of Nurr1 on subarachnoid hemorrhage (SAH) progression and investigated the feasibility of its agonist (amodiaquine, AQ) as a treatment for SAH. Methods. SAH rat models were constructed by the endovascular perforation technique. AQ was administered intraperitoneally at 2 hours after SAH induction. SAH grade, mortality, weight loss, neurological performance tests, brain water content, western blot, immunofluorescence, Nissl staining, and qPCR were assessed post-SAH. In vitro, hemin was introduced into HT22 cells to develop a model of SAH. Results. Stimulation of Nurr1 with AQ improved the outcomes and attenuated brain edema. Nurr1 was mainly expressed in neuron, and administration of AQ alleviated neuron injury in vivo and enhanced the neuron viability and inhibited neuron apoptosis and necrosis in vitro. Besides, AQ reduced the amount of IL-1β+Iba-1+ cells and inhibited the mRNA level of proinflammatory cytokines (IL-1β and TNF-α) and the M1-like phenotype markers (CD68 and CD86). AQ inhibited the expression of MMP9 in HT22 cells. Furthermore, AQ reduced the expression of nuclear NF-κB and Nurr1 while increased cytoplasmic Nurr1 in vivo and in vitro. Conclusion. Pharmacological activation of Nurr1 with AQ alleviated the neuron injury and neuroinflammation. The mechanism of antineuroinflammation may be associated with the Nurr1/NF-κB/MMP9 pathway in the neuron. The data supported that AQ might be a promising treatment strategy for SAH.

Spatiotemporal spread pattern of the COVID-19 cases in China (preprint)

COVID-19 is currently spreading widely around the world, causing huge threats on public safety and global society. This study analyzes the spatiotemporal spread pattern of the COVID-19 in China, reveals China’s epicenters of the epidemic through spatial clustering, and delineates the substantial effect of distance to Wuhan on the COVID-19 spread. The results show that the daily new COVID-19 cases mostly occurred in and around Wuhan before March 6, and then moved to the Grand Bay Area (Shenzhen, Hong Kong and Macau). The total COVID-19 cases in China were mainly distributed in the east of the Huhuanyong Line, where the epicenters account for more than 60% of the country’s total on 24 January and 7 February, half on 31 January, and more than 70% from 14 February. The total cases finally stabilized around 84,000, and the inflection point for Wuhan was on 14 February, one week later than those of Hubei (outside Wuhan) and China (outside Hubei). The generalized additive model-based analysis shows that population density and distance to provincial cities significantly associated with the total number of the cases, while distances to prefecture cities and inter-city traffic stations, and population inflow from Wuhan after 24 January, had no strong relationships with the total number of cases. The results and findings should provide valuable insights for understanding the changes in the COVID-19 transmission and controlling the global COVID-19 spread.

Stimulator of IFN genes mediates neuroinflammatory injury by suppressing AMPK signal in experimental subarachnoid hemorrhage.

BACKGROUND: Neuroinflammation is closely associated with the poor prognosis in subarachnoid hemorrhage (SAH) patients. This study was aimed to determine the role of stimulator of IFN genes (STING), an essential regulator to innate immunity, in the context of SAH. METHODS: A total of 344 male C57BL/6 J mice were subjected to endovascular perforation to develop a model of SAH. Selective STING antagonist C-176 and STING agonist CMA were administered at 30 min or 1 h post-modeling separately. To investigate the underlying mechanism, the AMPK inhibitor compound C was administered intracerebroventricularly at 30 min before surgery. Post-SAH assessments included SAH grade, neurological test, brain water content, western blotting, RT-PCR, and immunofluorescence. Oxygenated hemoglobin was introduced into BV2 cells to establish a SAH model in vitro. RESULTS: STING was mainly distributed in microglia, and microglial STING expression was significantly increased after SAH. Administration of C-176 substantially attenuated SAH-induced brain edema and neuronal injury. More importantly, C-176 significantly alleviated both short-term and persistent neurological dysfunction after SAH. Meanwhile, STING agonist CMA remarkably exacerbated neuronal injury and deteriorated neurological impairments. Mechanically, STING activation aggravated neuroinflammation via promoting microglial activation and polarizing into M1 phenotype, evidenced by microglial morphological changes, as well as the increased level of microglial M1 markers including IL-1ß, iNOS, IL-6, TNF-α, MCP-1, and NLRP3 inflammasome, while C-176 conferred a robust anti-inflammatory effect. However, all the mentioned beneficial effects of C-176 including alleviated neuroinflammation, attenuated neuronal injury and the improved neurological function were reversed by AMPK inhibitor compound C. Meanwhile, the critical role of AMPK signal in C-176 mediated anti-inflammatory effect was also confirmed in vitro. CONCLUSION: Microglial STING yielded neuroinflammation after SAH, while pharmacologic inhibition of STING could attenuate SAH-induced inflammatory injury at least partly by activating AMPK signal. These data supported the notion that STING might be a potential therapeutic target for SAH.