STG-Net: A COVID-19 prediction network based on multivariate spatio-temporal information.

The modern urban population features a high population density and a fast population flow, and COVID-19 has strong transmission ability, long incubation period, and other characteristics. Considering only the time sequence of COVID-19 transmission cannot effectively respond to the current epidemic transmission situation. The distance between cities and population density information also have a significant impact on the transmission of the virus. Currently, cross-domain transmission prediction models do not fully exploit the time-space information and fluctuation trend of data, and cannot reasonably predict the trend of infectious diseases by integrating time-space multi-source information. To solve this problem, this paper proposes the COVID-19 prediction network (STG-Net) based on multivariate spatio-temporal information, which introduces the Spatial Information Mining module (SIM) and the Temporal Information Mining module (TIM) to mine the spatio-temporal information of the data in a deeper level, and uses the slope feature method to further mine the fluctuation trend of the data. Also, we introduce the Gramian Angular Field module (GAF), which converts one-dimensional data into two-dimensional images, further enhancing the network's feature mining capability in the time and feature dimension, ultimately combining spatiotemporal information to predict daily newly confirmed cases. We tested the network on datasets from China, Australia, the United Kingdom, France, and Netherlands. The experimental results show that STG-Net has better prediction performance than existing prediction models, with an average decision coefficient R2 of 98.23% on the datasets from five countries, as well as good long- and short-term prediction ability and overall good robustness.

STG-Net: A COVID-19 prediction network based on multivariate spatio-temporal information

Graphical abstract The modern urban population features a high population density and a fast population flow, and COVID-19 has strong transmission ability, long incubation period, and other characteristics. Considering only the time sequence of COVID-19 transmission cannot effectively respond to the current epidemic transmission situation. The distance between cities and population density information also have a significant impact on the transmission of the virus. Currently, cross-domain transmission prediction models do not fully exploit the time–space information and fluctuation trend of data, and cannot reasonably predict the trend of infectious diseases by integrating time–space multi-source information. To solve this problem, this paper proposes the COVID-19 prediction network (STG-Net) based on multivariate spatio-temporal information, which introduces the Spatial Information Mining module (SIM) and the Temporal Information Mining module (TIM) to mine the spatio-temporal information of the data in a deeper level, and uses the slope feature method to further mine the fluctuation trend of the data. Also, we introduce the Gramian Angular Field module (GAF), which converts one-dimensional data into two-dimensional images, further enhancing the network's feature mining capability in the time and feature dimension, ultimately combining spatiotemporal information to predict daily newly confirmed cases. We tested the network on datasets from China, Australia, the United Kingdom, France, and Netherlands. The experimental results show that STG-Net has better prediction performance than existing prediction models, with an average decision coefficient R2 of 98.23% on the datasets from five countries, as well as good long- and short-term prediction ability and overall good robustness.

Effect of inactivated COVID-19 vaccines on seizure frequency in patients with epilepsy: A multicenter, prospective study.

Objectives: Several COVID-19 vaccines list "uncontrolled epilepsy" as a contraindication for vaccination. This consequently restricts vaccination against COVID-19 in patients with epilepsy (PWE). However, there is no strong evidence that COVID-19 vaccination can exacerbate conditions in PWE. This study aims to determine the impact of COVID-19 vaccination on PWE. Methods: PWE were prospectively recruited from 25 epilepsy centers. We recorded the seizure frequency at three time periods (one month before the first vaccination and one month after the first and second vaccinations). A generalized linear mixed-effects model (GLMM) was used for analysis, and the adjusted incidence rate ratio (AIRR) with 95% CI was presented and interpreted accordingly. Results: Overall, 859 PWE were included in the analysis. Thirty-one (3.6%) and 35 (4.1%) patients were found to have increased seizure frequency after the two doses, respectively. Age had an interaction with time. The seizure frequency in adults decreased by 81% after the first dose (AIRR=0.19, 95% CI:0.11-0.34) and 85% after the second dose (AIRR=0.16, 95% CI:0.08-0.30). In juveniles (<18), it was 25% (AIRR=0.75, 95% CI:0.42-1.34) and 51% (AIRR=0.49, 95% CI:0.25-0.95), respectively. Interval between the last seizure before vaccination and the first dose of vaccination (ILSFV) had a significant effect on seizure frequency after vaccination. Seizure frequency in PWE with hereditary epilepsy after vaccination was significantly higher than that in PWE with unknown etiology (AIRR=1.95, 95% CI: 1.17-3.24). Two hundred and seventeen (25.3%) patients experienced non-epileptic but not serious adverse reactions. Discussion: The inactivated COVID-19 vaccine does not significantly increase seizure frequency in PWE. The limitations of vaccination in PWE should focus on aspects other than control status. Juvenile PWE should be of greater concern after vaccination because they have lower safety. Finally, PWE should not reduce the dosage of anti-seizure medication during the peri-vaccination period.

VOC-DL: Deep learning prediction model for COVID-19 based on VOC virus variants.

BACKGROUND AND OBJECTIVE: The ever-mutating COVID-19 has infected billions of people worldwide and seriously affected the stability of human society and the world economic development. Therefore, it is essential to make long-term and short-term forecasts for COVID-19. However, the pandemic situation in different countries and regions may be dominated by different virus variants, and the transmission capacity of different virus variants diversifies. Therefore, there is a need to develop a predictive model that can incorporate mutational information to make reasonable predictions about the current pandemic situation. METHODS: This paper proposes a deep learning prediction framework, VOC-DL, based on Variants Of Concern (VOC). The framework uses slope feature method to process the time series dataset containing VOC variant information, and uses VOC-LSTM, VOC-GRU and VOC-BILSTM prediction models included in the framework to predict the daily newly confirmed cases. RESULTS: We analyzed daily newly confirmed cases in Italy, South Korea, Russia, Japan and India from April 14th, 2021 to July 3rd, 2021. The experimental results show that all VOC-DL models proposed in this paper can accurately predict the pandemic trend in the medium and long term, and VOC-LSTM model has the best prediction performance, with the highest average determination coefficient R2 of 96.83% in five nations' datasets. The overall prediction has robustness. CONCLUSIONS: The experimental results show that VOC-LSTM is the best predictor for such a series of data and has higher prediction accuracy in the long run. At the same time, our VOC-DL framework combining VOC variants has reference significance for predicting other variants in the future.