BACKGROUND: The utilization of non-invasive techniques for fetal cardiac health surveillance is pivotal in evaluating fetal well-being throughout the gestational period. This process requires clean and interpretable fetal Electrocardiogram (fECG) signals.
METHODS: The proposed work is the novel framework for the elicitation of fECG signals from abdominal ECG (aECG) recordings of the pregnant mother. The comprehensive approach encompasses pre-processing of the raw ECG signal, Blind Source Separation techniques (BSS), Decomposition techniques like Empirical Mode Decomposition (EMD), and its variants like Ensemble Empirical Mode Decomposition (EEMD), and Complete Ensemble Empirical Mode Decomposition with Additive Noise (CEEMDAN). The Robust Set Membership Affine Projection (RSMAP) Algorithm is deployed for the enhancement of the obtained fECG signal.
RESULTS: The results show significant improvements in the elicited fECG signal with a maximum Signal Noise Ratio (SNR) of 31.72 dB and correlation coefficient = 0.899, Maximum Heart Rate(MHR) obtained in the range of 108-142 bpm for all the records of abdominal ECG signals. The statistical test gave a p-value of 0.21 accepting the null hypothesis. The Abdominal and Direct Fetal Electrocardiogram Database (ABDFECGDB) from PhysioNet has been used for this analysis.
CONCLUSIONS: The proposed framework demonstrates a robust and effective method for the elicitation and enhancement of fECG signals from the abdominal recordings.

The concentration prediction of PM2.5 plays a vital role in controlling the air and improving the environment. This paper proposes a prediction model (namely EEMD-ALSTM) based on Ensemble Empirical Mode Decomposition (EEMD), Attention Mechanism and Long Short-Term Memory network (LSTM). Through the combination of decomposition and LSTM, attention mechanism is introduced to realize the prediction of PM2.5 concentration. The advantage of EEMD-ALSTM model is that it decomposes and combines the original data using the method of ensemble empirical mode decomposition, reduces the high nonlinearity of the original data, and Specially reintroduction the attention mechanism, which enhances the extraction and retention of data features by the model. Through experimental comparison, it was found that the EEMD-ALSTM model reduced its MAE and RMSE by about 15% while maintaining the same R2 correlation coefficient, and the stability of the model in the prediction process was also improved significantly.

Accurate predictions of stock markets are important for investors and other stakeholders of the equity markets to formulate profitable investment strategies. The improved accuracy of a prediction model even with a slight margin can translate into considerable monetary returns. However, the stock markets\' prediction is regarded as an intricate research problem for the noise, complexity and volatility of the stocks\' data. In recent years, the deep learning models have been successful in providing robust forecasts for sequential data. We propose a novel deep learning-based hybrid classification model by combining peephole LSTM with temporal attention layer (TAL) to accurately predict the direction of stock markets. The daily data of four world indices including those of U.S., U.K., China and India, from 2005 to 2022, are examined. We present a comprehensive evaluation with preliminary data analysis, feature extraction and hyperparameters\' optimization for the problem of stock market prediction. TAL is introduced post peephole LSTM to select the relevant information with respect to time and enhance the performance of the proposed model. The prediction performance of the proposed model is compared with that of the benchmark models CNN, LSTM, SVM and RF using evaluation metrics of accuracy, precision, recall, F1-score, AUC-ROC, PR-AUC and MCC. The experimental results show the superior performance of our proposed model achieving better scores than the benchmark models for most evaluation metrics and for all datasets. The accuracy of the proposed model is 96% and 88% for U.K. and Chinese stock markets respectively and it is 85% for both U.S. and Indian markets. Hence, the stock markets of U.K. and China are found to be more predictable than those of U.S. and India. Significant findings of our work include that the attention layer enables peephole LSTM to better identify the long-term dependencies and temporal patterns in the stock markets\' data. Profitable and timely trading strategies can be formulated based on our proposed prediction model.

BACKGROUND: Functional brain networks (FBNs) coordinate brain functions and are studied in fMRI using blood-oxygen-level-dependent (BOLD) signal correlations. Previous research links FBN changes to aging and cognitive decline, but various physiological factors influnce BOLD signals. Few studies have investigated the intrinsic components of the BOLD signal in different timescales using signal decomposition. This study aimed to explore differences between intrinsic FBNs and traditional BOLD-FBN, examining their associations with age and cognitive performance in a healthy cohort without dementia.
METHODS: A total of 396 healthy participants without dementia (men = 157; women = 239; age range = 20-85 years) were enrolled in this study. The BOLD signal was decomposed into several intrinsic signals with different timescales using ensemble empirical mode decomposition, and FBNs were constructed based on both the BOLD and intrinsic signals. Subsequently, network features-global efficiency and local efficiency values-were estimated to determine their relationship with age and cognitive performance.
RESULTS: The findings revealed that the global efficiency of traditional BOLD-FBN correlated significantly with age, with specific intrinsic FBNs contributing to these correlations. Moreover, local efficiency analysis demonstrated that intrinsic FBNs were more meaningful than traditional BOLD-FBN in identifying brain regions related to age and cognitive performance.
CONCLUSIONS: These results underscore the importance of exploring timescales of BOLD signals when constructing FBN and highlight the relevance of specific intrinsic FBNs to aging and cognitive performance. Consequently, this decomposition-based FBN-building approach may offer valuable insights for future fMRI studies.

Water availability (WA) is a key factor influencing the carbon cycle of terrestrial ecosystems under climate warming, but its effects on gross primary production (EWA-GPP ) at multiple time scales are poorly understood. We used ensemble empirical mode decomposition (EEMD) and partial correlation analysis to assess the WA-GPP relationship (RWA-GPP ) at different time scales, and geographically weighted regression (GWR) to analyze their temporal dynamics from 1982 to 2018 with multiple GPP datasets, including near-infrared radiance of vegetation GPP, FLUXCOM GPP, and eddy covariance-light-use efficiency GPP. We found that the 3- and 7-year time scales dominated global WA variability (61.18% and 11.95%), followed by the 17- and 40-year time scales (7.28% and 8.23%). The long-term trend also influenced 10.83% of the regions, mainly in humid areas. We found consistent spatiotemporal patterns of the EWA-GPP and RWA-GPP with different source products: In high-latitude regions, RWA-GPP changed from negative to positive as the time scale increased, while the opposite occurred in mid-low latitudes. Forests had weak RWA-GPP at all time scales, shrublands showed negative RWA-GPP at long time scales, and grassland (GL) showed a positive RWA-GPP at short time scales. Globally, the EWA-GPP , whether positive or negative, enhanced significantly at 3-, 7-, and 17-year time scales. For arid and humid zones, the semi-arid and sub-humid zones experienced a faster increase in the positive EWA-GPP , whereas the humid zones experienced a faster increase in the negative EWA-GPP . At the ecosystem types, the positive EWA-GPP at a 3-year time scale increased faster in GL, deciduous broadleaf forest, and savanna (SA), whereas the negative EWA-GPP at other time scales increased faster in evergreen needleleaf forest, woody savannas, and SA. Our study reveals the complex and dynamic EWA-GPP at multiple time scales, which provides a new perspective for understanding the responses of terrestrial ecosystems to climate change.

We propose a method to enhance the accuracy of arrival time picking of noisy microseismic recordings. A series of intrinsic mode functions (IMFs) of the microseismic signal are initially decomposed by employing the ensemble empirical mode decomposition. Subsequently, the sample entropy values of the obtained IMFs are calculated and applied to set an appropriate threshold for selecting IMFs. These are then reconstructed to distinguish between noise and useful signals. Ultimately, the Akaike information criterion picker is used to determine the arrival time of the denoised signal. Test results using synthetic noisy microseismic recordings demonstrate that the proposed approach can significantly reduce picking errors, with errors within the range of 1-3 sample intervals. The proposed method can also give a more stable picking result when applied to different microseismic recordings with different signal-to-noise ratios. Further application in real microseismic recordings confirms that the developed method can estimate an accurate arrival time of noisy microseismic recordings.

It has been proven that marine heatwaves (MHWs) have increased in frequency, duration, and intensity over the past few decades, and this trend will accelerate further under continued global warming. While more intense and frequent MHWs are an expected consequence of rising sea surface temperatures (SSTs) under continued global warming, it remains unclear to what degree per Celsius warming trend of SSTs contributes to the changes in the MHW metrics. Here, we focus on how the MHW metrics evolve with the SST warming trend by using an adaptive data analysis method based on observational datasets covering the past four decades. We find that the globally averaged increasing rates of the annual MHW frequency, duration, and maximum intensity are approximately 3.7 events, 7.5 days, and 2.2° Celsius per degree Celsius of SST rise, respectively. The increasing rates for the annual MHW days and the fraction of the spatial extents to the global ocean affected by MHWs are approximately 58.8 days and 13.9 % per degree Celsius of SST rise, respectively. Based on these observational-based increasing rates and the projected SST warming from the selected Coupled Model Intercomparison Project Phase 6 (CMIP6) models, the spatial distributions of changes in annual MHW days, frequency, and cumulative intensity are projected to exhibit 2-fold, 4-fold, and 6 to 8-fold increases under the three socioeconomic pathways (i.e., SSP126, SSP245, and SSP585), respectively. The globally averaged annual MHW days will increase to approximately 224.2 ± 26.9 days, and the largest changes are projected to occur in the northeast Pacific, the North Atlantic, the south Indian Oceans, and parts of the Southern Ocean, with approximately 14.8 ± 5.7 % of the global ocean reaching a permanent MHW state by the end of the twenty-first century under SSP585.

The flow of information between markets is important to guide investors and policymakers in the effective allocation of assets and proactive market regulation, respectively. This study examines the impact of information flow from global financial market stress on the African stock markets using the daily US financial stress index (USFSI) and other advanced economies\' financial stress index (OAEFSI) to proxy the global financial stress index. To understand the information flow dynamics across various investment horizons, the ensemble empirical mode decomposition (EEMD)-based transfer entropy is employed. Our findings reveal that African equity markets are highly risky for information flow from global financial market stress. However, we identify diversification prospects based on market conditions for Ghana and Egypt in the short term and Tanzania, Cote D\'Ivoire, and Egypt in the medium term. Empirical results also show that the information flow from global financial stress to African stock markets depends on time scales, economic relations, and the state of global financial markets. The findings are important for investors, portfolio managers, practitioners, and policymakers.

Global vegetation photosynthesis and productivity have increased substantially since the 1980s, but this trend is heterogeneous in both time and space. Here, we categorize the secular trend in global vegetation greenness into sustained greening, sustained browning and greening-to-browning. We found that by 2016, increased global vegetation greenness had begun to level off, with the area of browning increasing in the last decade, reaching 39.0 million km2 (35.9% of the world\'s vegetated area). This area is larger than the area with sustained increasing growth (27.8 million km2, 26.4%); thus, 12.0% ± 3.1% (0.019 ± 0.004 NDVI a-1) of the previous earlier increase has been offset since 2010 (2010-2016, P < 0.05). Global gross primary production also leveled off, following the trend in vegetation greenness in time and space. This leveling off was caused by increasing soil water limitations due to the spatial expansion of drought, whose impact dominated over the impacts of temperature and solar radiation. This response of global gross primary production to soil water limitation was not identified by land submodels within Earth system models. Our results provide empirical evidence that global vegetation greenness and primary production are offset by water stress and suggest that as global warming continues, land submodels may overestimate the world\'s capacity to take up carbon with global vegetation greening.

The novel coronavirus pneumonia that began to spread in 2019 is still raging and has placed a burden on medical systems and governments in various countries. For policymaking and medical resource decisions, a good prediction model is necessary to monitor and evaluate the trends of the epidemic. We used a long short-term memory (LSTM) model and the improved hybrid model based on ensemble empirical mode decomposition (EEMD) to predict COVID-19 trends; Methods: The data were collected from the Harvard Dataverse. Epidemic data from 21 January 2020 to 25 April 2021 for California, the most severely affected state in the United States, were used to develop an LSTM model and an EEMD-LSTM hybrid model, which is an LSTM model combined with ensemble empirical mode decomposition. In this study, ninety percent of the data were adopted to fit the models as a training set, while the subsequent 10% were used to test the prediction effect of the models. The mean absolute percentage error, mean absolute error, and root mean square error were used to evaluate the prediction performances of the models; Results: The results indicated that the number of confirmed cases in California was increasing as of 25 April 2021, with no obvious evidence of a sharp decline. On 25 April 2021, the LSTM model predicted 3666418 confirmed cases, whereas the EEMD-LSTM predicted 3681150. The mean absolute percentage errors for the LSTM and EEMD-LSTM models were 0.0151 and 0.0051, respectively. The mean absolute and root mean square errors were 5.58 × 104 and 5.63 × 104 for the LSTM model and 1.9 × 104 and 2.43 × 104 for the EEMD-LSTM model, respectively; Conclusions: The results showed the advantage of an EEMD-LSTM model over a single LSTM model, and the established EEMD-LSTM model may be suitable for monitoring and evaluating the epidemic situation and providing quantitative analysis evidence for epidemic prevention and control.