Predictor Of Future Events Research Articles

An Efficient Deep Learning Framework for Optimized Event Forecasting

There have been several catastrophic events that have impacted multiple economies and resulted in thousands of fatalities, and violence has generated a severe political and financial crisis. Multiple studies have been centered around the artificial intelligence (AI) and machine learning (ML) approaches that are most widely used in practice to detect or forecast violent activities. However, machine learning algorithms become less accurate in identifying and forecasting violent activity as data volume and complexity increase. For the prediction of future events, we propose a hybrid deep learning (DL)-based model that is composed of a convolutional neural network (CNN), long short-term memory (LSTM), and an attention layer to learn temporal features from the benchmark the Global Terrorism Database (GTD). The GTD is an internationally recognized database that includes around 190,000 violent events and occurrences worldwide from 1970 to 2020. We took into account two factors for this experimental work: the type of event and the type of object used. The LSTM model takes these complex feature extractions from the CNN first to determine the chronological link between data points, whereas the attention model is used for the time series prediction of an event. The results show that the proposed model achieved good accuracies for both cases—type of event and type of object—compared to benchmark studies using the same dataset (98.1% and 97.6%, respectively).

Inflammation and cholesterol associated residual cardiovascular risk and its dependence on chronic kidney disease in statin treated patients with coronary artery disease

Abstract Background/Aim In statin-treated patients with chronic coronary syndrome (CCS) there is uncertainty about the relative contribution of inflammation and dyslipidaemia to residual cardiovascular risk, particularly in the presence of chronic kidney disease. We thus aimed to evaluate high sensitive C-reactive protein (hsCRP) and low density lipoprotein-cholesterol (LDL-C) as predictors of future events in these patients, stratified by estimated glomerular filtration rate (eGFR). Methods We identified patients with CCS on stable statin treatment from a contemporary all-comers cohort recruited at a large tertiary care center in Germany. We excluded patients with active malignancy, infectious disease or in a post-operative state. Cox regression for major adverse cardiovascular events (MACE, composite of cardiovascular death, myocardial infarction, stroke) and all-cause death was performed adjusting for age, sex, arterial hypertension, diabetes mellitus, active smoking, body mass index and change in the intensity of statin therapy at discharge, with log(hsCRP) or LDL-C as the exposure variables. Interaction with eGFR (< vs. ≥ 60 ml/min/1.73m²) was investigated. Results 937 patients with CCS (mean age 71.1 years, 78.7% male, median hsCRP 2.0 mg/L, median LDL-C 73.0 mg/dl, median eGFR 71.3 ml/min/1.73m²) were followed-up over 4 years, with 101 cases of MACE and 171 cases of all-cause death recorded. The hazard ratios (HR) for MACE per standard deviation (SD) increase were as follows: log(hsCRP) 1.08 (95% confidence interval (CI) 0.87, 1.34; p = 0.48), LDL-C 1.10 (95% CI 0.89, 1.35; p = 0.37). For all-cause death the HRs for log(hsCRP) and LDL-C were 1.60 (95% CI 1.36, 1.88; p <0.001) and 0.88 (95% CI 0.74, 1.04; p = 0.14), respectively. Kaplan-Meier curves stratifying the cohort by median values of LDL-C and hsCRP consistently demonstrated a significantly higher risk of death in the subgroups with elevated hsCRP (Figure 1). Contrastingly for LDL-C, levels below (vs. above) the median were associated with a higher mortality rate. This is in part a reflection of a higher prevalence of high-intensity statin therapy in this subgroup (21.0 vs. 14.2%, p = 0.0081), in turn a reflection of greater severity of coronary artery disease (history of myocardial infarction 43.9 vs. 31.9%, p <0.001). Further confounding by unadjusted for diseases with a catabolic metabolism is possible. Interaction analysis with eGFR revealed that log(hsCRP) was predictive of all-cause mortality both in the presence and absence of chronic kidney disease, whilst lower LDL-C values were found to be predictive only at eGFR <60 ml/min/1.73m² (Table 1). Conclusion Residual inflammation as measured by hsCRP predicts all-cause mortality in statin-treated patients with CCS in this contemporary all-comers cohort, both in the presence and absence of chronic kidney disease. The predictive utility of LDL-C is limited, likely due to confounding e.g. by the intensity of lipid lowering medication.Kaplan-Meier curves (all-cause death)Interaction analysis with eGFR

Advancing Subseasonal Surface Air Temperature and Heat Wave Prediction Skill in China by Incorporating Scale Interaction in a Deep Learning Model

AbstractAccurate subseasonal predictions of high surface air temperature (SAT) and heat wave events 10–30 days in advance are crucial for mitigating the risks of extreme weather; however, they pose a challenge for current operational models. In this study, we trained a convolutional neural network (CNN)‐based deep learning model to exploit the modulations in China's SAT by precursor signals across different timescales to improve predictions of future SAT and heat wave events. This CNN model demonstrated superior capability in capturing the evolution of SAT anomalies and the occurrence of heat wave events with forecast lead times beyond 20 days, compared with that of the operational models of the China Meteorological Administration and European Centre for Medium‐Range Weather Forecasts. Explainability analysis highlighted that subseasonal SAT predictability in China is driven primarily by large‐scale intraseasonal perturbations from both lower‐ and higher‐latitude regions of Eurasia and interannual variability.

Flammable futures—storylines of climatic impacts on wildfire events and palm oil plantations in Indonesia

Abstract Wildfire events are driven by complex interactions of climate and anthropogenic interventions. Predictions of future wildfire events, their extremity, and their impact on the environment and economy must consider the interactions between these drivers. Economic policy and land use decisions influence the susceptibility of an area to climate extremes, the probability of burning, and future decision making. To better understand how climate-driven drought events and adaptation efforts affect burned area, agricultural production losses, and land use decisions, we developed a storyline approach centered on Indonesia’s 2015 fire events, which saw significant (>5%) production losses of palm oil. We explored analogous events under three warming conditions and two storylines (multi-model ensemble mean climate change and high impact). We employed a model chain consisting of CMIP6 climate modeling to quantify climate change impacts, a wildfire climate impacts and adaptation model (FLAM) to predict burned areas, and the Global Biosphere Management Model (GLOBIOM) to predict resultant production losses and socio-economic consequences in the oil palm sector in Indonesia and, by extension, the EU. FLAM is a mechanistic fire model used to reproduce and project wildfires based on various criteria and input variables; GLOBIOM is a global economic land use model, which assesses land use competition and provides economic impacts based on scenario data. We found that total burned area and production loss can increase by up to 25% and lead to local price increases up to 70%, with only minor differences beyond 2.5 degrees of warming. Our results highlight the importance of considering the interactions of future warming, drought conditions, and extreme weather events when predicting their impacts on oil palm losses and burned area. This study sets the stage for further exploration on the impacts of land management policies on local and international environments and economies in the context of global warming.

A Longitudinal Study of the Impact of COVID-19 on Optimism Prediction

The purpose of this longitudinal study was to investigate how optimistic predictions, hopelessness, and depressive symptoms changed as a result of the COVID-19 pandemic, and the causal relationships between these variables. To achieve this purpose, we used data from online surveys conducted in 2019 and 2021 among men and women aged 20–79. Based on item response theory, we developed a future prediction task for the assessment of optimistic predictions. Our comparison of online survey responses found a decline in optimistic predictions before and after the pandemic. More specifically, there were no change in predictions of negative future events, but there was the decrease in predictions of positive future events. Furthermore, we found that those who were more stressed by COVID-19 were less likely to have an optimistic view of the near future. We also found a relationship between optimistic predictions and hopelessness and depressive symptoms with lower optimism predicting more hopelessness and more depression predicting lower optimism. To prevent feelings of hopelessness, it is important to help people develop positive expectations about the future.

Historical Trends and Normalizing Flow for One-shot Temporal Knowledge Graph Reasoning

Temporal Knowledge Graphs (TKGs) have garnered significant scholarly interest for their ability to dynamically represent the evolution of real-world events. However, the presence of the long tail effect in knowledge graphs necessitates the exploration of one-shot temporal knowledge graph reasoning. Concurrently, the issues of out-of-distribution and overfitting heavily impact the performance of models. Empirical evidence from real-world cases has demonstrated the significance of historical trend information in accurately predicting future events, and normalizing flows effectively address out-of-distribution and overfitting challenges. Thus, this paper proposes a novel approach named Historical Trends and Normalizing Flow for One-shot Temporal Knowledge Graph Reasoning (HNOT). Specifically, HNOT combine normalizing flows with a historical information aggregator to obtain more complex distributions, thereby resolving out-of-distribution and overfitting problems and capturing global representations of TKGs. In contrast to existing models that focus solely on entity and relationship modeling, we have introduced a novel historical trend information aggregator. This aggregator is designed to extract trend information from temporal knowledge subgraphs. By employing trend information as the primary modeling object, it facilitates the exploration of trend variations at different timestamps. This approach serves as an auxiliary source of global information. Subsequently, we integrate these two distinct types of information to achieve predictions of future events. Experimental results demonstrate the superiority of the proposed method over state-of-the-art baselines on three benchmark datasets.

TEE4EHR: Transformer event encoder for better representation learning in electronic health records

Irregular sampling of time series in electronic health records (EHRs) is one of the main challenges for developing machine learning models. Additionally, the pattern of missing values in certain clinical variables is not at random but depends on the decisions of clinicians and the state of the patient. Point process is a mathematical framework for analyzing event sequence data consistent with irregular sampling patterns. Our model, TEE4EHR, is a transformer event encoder (TEE) with point process loss that encodes the pattern of laboratory tests in EHRs. The utility of our TEE has been investigated in various benchmark event sequence datasets. Additionally, we conduct experiments on two real-world EHR databases to provide a more comprehensive evaluation of our model. Firstly, in a self-supervised learning approach, the TEE is jointly learned with an existing attention-based deep neural network, which gives superior performance in negative log-likelihood and future event prediction. Besides, we propose an algorithm for aggregating attention weights to reveal the events’ interactions. Secondly, we transfer and freeze the learned TEE to the downstream task for the outcome prediction, where it outperforms state-of-the-art models for handling irregularly sampled time series. Furthermore, our results demonstrate that our approach can improve representation learning in EHRs and be useful for clinical prediction tasks.

Detection and Classification of Lanes Using Reannotated Lane Dataset to Aid Autonomous Driving

The core functionality of advanced driver assistance systems and self-driving cars depends on the ability to recognize drivable road areas. As there are many different categories of useful markings incorporated within the road area, lane detection and classification form a major step for taking appropriate actions leading to truly autonomous driving. Existing datasets do not provide ample classification of lane types and adequate granularity for precise localization of lane markings. A new dedicated dataset for semantic segmentation of 11 varied lane types obtained by reannotating the BDD100K dataset is presented. The reannotation process involves pixel-level lane markings of 76,000 image instances of the BDD100K dataset, which were originally represented as a sequence of coordinate points. This opens up the possibility of high resolution both spatially and semantically in the context of lane understanding for autonomous driving. Baseline results on the proposed dataset based on the Bilateral Segmentation Network (BiSeNetV2), which considers the spatial data and the categorical semantics distinctly, are presented. The performance is pinned at 85% accuracy during testing using BiSeNetV2 architecture. The dataset is expected to open up new directions of research to address problems such as severe class imbalance, segmentation of multiple classes of texture-less and eccentric features (lanes), and so forth. As a result, applications such as lane-centric activity interpretation, future event prediction, and continuous learning are expected.

Machine Learning Techniques for Spatio-Temporal Air Pollution Prediction to Drive Sustainable Urban Development in the Era of Energy and Data Transformation

Sustainable urban development in the era of energy and digital transformation is crucial from a societal perspective. Utilizing modern techniques for analyzing large datasets, including machine learning and artificial intelligence, enables a deeper understanding of historical data and the efficient prediction of future events based on data from IoT sensors. This study conducted a multidimensional historical analysis of air pollution to investigate the impacts of energy transformation and environmental policy and to determine the long-term environmental implications of certain actions. Additionally, machine learning (ML) techniques were employed for air pollution prediction, taking spatial factors into account. By utilizing multiple low-cost air sensors categorized as IoT devices, this study incorporated data from various locations and assessed the influence of neighboring sensors on predictions. Different ML approaches were analyzed, including regression models, deep neural networks, and ensemble learning. The possibility of implementing such predictions in publicly accessible IT mobile systems was explored. The research was conducted in Krakow, Poland, a UNESCO-listed city that has had long struggle with air pollution. Krakow is also at the forefront of implementing policies to prohibit the use of solid fuels for heating and establishing clean transport zones. The research showed that population growth within the city does not have a negative impact on PMx concentrations, and transitioning from coal-based to sustainable energy sources emerges as the primary factor in improving air quality, especially for PMx, while the impact of transportation remains less relevant. The best results for predicting rare smog events can be achieved using linear ML models. Implementing actions based on this research can significantly contribute to building a smart city that takes into account the impact of air pollution on quality of life.

Understanding drivers of wildfires in California

This research study investigates the intricate relationship between El Niño events, precipitation patterns, and wildfires in California over the past two decades. California, known for its susceptibility to wildfires, has experienced significant variations in precipitation influenced by the El Niño-Southern Oscillation (ENSO) phenomenon. The objective of this research is to examine the impact of El Niño events on California's precipitation patterns and subsequently assess their association with wildfire occurrences. Predicting wildfire and precipitation wildfires are important to California’s significant agriculture production and better preparing for possible droughts or wildfires. To achieve this objective, a comprehensive analysis of climate data spanning the last 20 years was conducted. Historical records of El Niño episodes, obtained from reputable sources, were measured by correlation with precipitation data acquired from meteorological databases. Additionally, wildfire occurrence and severity data were collected from relevant agencies and organizations to be measured for correlation. As it turns out, while there was some positive correlation with El Nino and precipitation, there was little direct correlation between El Nino and wildfires. This means that there are many more complex factors to be assessed when figuring out causes of wildfires in addition to El Nino. The findings of this research will contribute to a better understanding of the main causes of wildfires in California. The results can inform policymakers, land managers, and disaster response agencies in developing more effective strategies to mitigate the risks associated with wildfires. Additionally, the study will provide valuable insights into the influence of climate variability on fire-prone ecosystems, aiding in the prediction and management of future wildfire events.

Continuous Sargassum monitoring across the Caribbean Sea and Central Atlantic using multi-sensor satellite observations

Recurrent transnational Sargassum blooms across the Caribbean Sea and Atlantic Ocean have received growing attention. Different multispectral sensors, including Moderate Resolution Imaging Spectroradiometer (MODIS), Visible Infrared Imager Radiometer Suite (VIIRS), and Ocean and Land Color Instrument (OLCI), have been used to map their distributions. However, the synergistic use of multi-sensor observations for high temporal resolution Sargassum monitoring is lacking. Here, by combining MODIS (on Aqua and Terra), VIIRS (on JPSS1 and SNPP), and OLCI (on Sentinel-3A and -3B) observations, 3-day mean Sargassum distributions were mapped across the Caribbean Sea and Central Atlantic. The Sargassum biomass densities were derived using the sensor-specific Alternative Floating Algae Index (AFAI)-biomass model, and the consistency between the six sensors was examined using MODIS Aqua as the reference sensor. Comparison of the Sargassum biomass derived from different sensors shows that they have strong linear correlations (R2 ≥ 0.95), demonstrating high consistency and continuity between the six-sensor observations. On average, the combined six-sensor datasets provide ∼1.6 times more valid observations compared to the MODIS-only dataset in 2021, enabling the generation of the 0.5° 3-day mean products over ∼90% of the study region. Such 0.5° 3-day mean products detected ∼10-20% more biomass in the bloom peak month (June 2021) compared to the monthly mean counterpart. Increasing the spatial resolution to 0.1°, the 3-day mean products can continuously monitor Sargassum dynamics with eddies and tropical cyclones, which cannot be well captured by single sensors. This study highlights that combining multiple polar-orbiting satellite observations can achieve 3-day gap-free monitoring of floating macroalgae dynamics in the Caribbean Sea and tropical Atlantic, thus facilitating the analyses of the bloom response to different environmental conditions and the prediction of future bloom events.

Prognostic impact of MRI-derived feature tracking myocardial strain in patients with non-ischaemic dilated cardiomyopathy: a systematic review and meta-analysis

To evaluate the clinical utility of feature tracking (FT)-derived myocardial strain in patients with non-ischaemic dilated cardiomyopathy (NIDCM). Electronic database searches of PubMed, Web of Science Core Collection, Cochrane advanced search, and EMBASE were performed. Studies on NIDCM were divided into categories according to left ventricular ejection fraction (LVEF; <30%, 30-40%, >40%), and correlations between strains and prevalence of late gadolinium enhancement (LGE) were evaluated by weighted correlation coefficients. Global longitudinal strain (GLS) hazard ratios were also integrated for prediction of future adverse events. The present meta-analysis analysed data from 5,767 patients with NIDCM from 30 eligible studies. GLS and global circumferential strain significantly differed across the three LVEF categories (all p<0.05); however, global radial strain did not. Only GLS showed a strong correlation with the prevalence of LGE (Spearman's correlation coefficient=0.61). The pooled HR of GLS for predicting adverse events was 1.15 (95% confidence interval [CI]: 1.07-1.23, p<0.001). In this meta-analysis, FT-derived GLS was strongly correlated with myocardial fibrosis and was an important predictor of future adverse events. These results suggest that FT-derived GLS may be useful in the pathological evaluation and risk stratification of NIDCM.

Recovering sedimentary ancient DNA of harmful dinoflagellates accumulated over the last 9000years off Eastern Tasmania, Australia.

Harmful algal blooms (HABs) have had significant adverse impacts on the seafood industry along the Tasmanian east coast over the past 4 decades. To investigate the history of regional HABs, we performed analyses of sedimentary ancient DNA (sedaDNA) in coastal sediments up to ~9000years old collected inshore and offshore of Maria Island, Tasmania. We used metagenomic shotgun sequencing and a hybridisation capture array ("HABbaits1") to target three harmful dinoflagellate genera, Alexandrium, Gymnodinium, and Noctiluca. Bioinformatic and DNA damage analyses verified the authenticity of the sedaDNA sequences. Our results show that dinoflagellates of Alexandrium genera have been present off eastern Tasmania during the last ~8300years, and we sporadically detected and unambiguously verified sequences of Gymnodinium catenatum that were present offshore up to ~7600years ago. We also recovered sedaDNA of the fragile, soft-bodied Noctiluca scintillans with increased relative abundance since 2010, consistent with plankton surveys. This study enabled us to identify challenges of sedaDNA sequence validation (in particular for G. catenatum, a microreticulate gymnodinoid species) and provided guidance for the development of tools to monitor past and present HAB species and improvement of future HAB event predictions.

Who Should I Engage With at What Time? A Missing Event-Aware Temporal Graph Neural Network.

Temporal graph neural network (GNN) has recently received significant attention due to its wide application scenarios, such as bioinformatics, knowledge graphs, and social networks. There are some temporal GNNs that achieve remarkable results. However, these works focus on future event prediction and are performed under the assumption that all historical events are observable. In real-world applications, events are not always observable, and estimating event time is as important as predicting future events. In this article, we propose, a missing event-aware temporal GNN, which uniformly models evolving graph structure and timing of events to support predicting what will happen in the future and when it will happen. models the dynamic of both observed and missing events as two coupled temporal point processes (TPPs), thereby incorporating the effects of missing events into the network. Experimental results on several real-world temporal graphs demonstrate that significantly outperforms the existing methods with up to 89% and 112% more accurate time and link prediction. Code can be found on https://github.com/HIT-ICES/TNNLS-MTGN.

Rapid calibration to dynamic temporal contexts.

The prediction of future events and the preparation of appropriate behavioural reactions rely on an accurate perception of temporal regularities. In dynamic environments, temporal regularities are subject to slow and sudden changes, and adaptation to these changes is an important requirement for efficient behaviour. Bayesian models have proven a useful tool to understand the processing of temporal regularities in humans; yet an open question pertains to the degree of flexibility of the prior that is required for optimal modelling of behaviour. Here we directly compare dynamic models (with continuously changing prior expectations) and static models (a stable prior for each experimental session) with their ability to describe regression effects in interval timing. Our results show that dynamic Bayesian models are superior when describing the responses to slow, continuous environmental changes, whereas static models are more suitable to describe responses to sudden changes. In time perception research, these results will be informative for the choice of adequate computational models and enhance our understanding of the neuronal computations underlying human timing behaviour.

Making sense of music: Insights from neurophysiology and connectivity analyses in naturalistic listening conditions

According to the latest frameworks, auditory perception and memory involve the constant prediction of future sound events by the brain, based on the continuous extraction of feature regularities from the environment. The neural hierarchical mechanisms for predictive processes in perception and memory for sounds are typically studied in relation to simple acoustic features in isolated sounds or sound patterns inserted in highly certain contexts. Such studies have identified reliable prediction formation and error signals, e.g., the N100 or the mismatch negativity (MMN) evoked responses. In real life, though, individuals often face situations in which uncertainty prevails and where making sense of sounds becomes a hard challenge. In music, not only deviations from predictions are masterly set up by composers to induce emotions but sometimes the sheer uncertainty of sound scenes is exploited for aesthetic purposes, especially in compositional styles such as Western atonal classical music. In very recent magnetoencephalography (MEG) and electroencephalography (EEG) studies, experimental and technical advances in stimulation paradigms and analysis approaches have permitted the identification of prediction-error responses from highly uncertain, atonal contexts and the extraction of prediction-related responses from real, continuous music. Moreover, functional connectivity analyses revealed the emergence of cortico-hippocampal interactions during the formation of auditory memories for more predictable vs. less predictable patterns. These findings contribute to understanding the general brain mechanisms that enable us to predict even highly uncertain sound environments and to possibly make sense of and appreciate even atonal music.

Predictive Analysis for Inventory Management of Coconut Warehouse (Case Study: Banio Lahewa)

Inventory management plays a pivotal role in the coconut farming business, directly influencing sales and income. An essential component of this management is warehousing, which not only affect revenue but also involves suppliers in the coconut storage process. Warehousing management and technology are two elements that can help companies operate more effectively and efficiently. This research focuses on efforts to improve warehouse management efficiency in the agricultural sector, particularly at Banio Lahewa, a company that operates as a coconut supplier in a small village with limited resources. Currently, the company still records data manually and lacks a real-time system to monitor demand patterns, stock rotation, and restocking frequency in the warehouse. This situation is caused by uncertainty about the products entering the warehouse, leading to the company’s focus being more limited to daily operational issues rather than future planning. To address this challenge, this research uses future event prediction methods, specifically forecasting by applying two neural network models: the Feed Forward Neural Network and the Long Short Term Memory. The implementation of this system is expected to provide new insights to the company, enabling them to be more adaptive in efficiently managing warehouse systems. With an understanding of patterns and predictions of future events, it is expected that the company can be more prepared and responsive to changes in customer demand and able to expand products more quickly. The results of this research are expected to make a positive contribution to the company, helping them optimize warehouse management and become more adaptive to market dynamics.

Baseline high sensitivity C-Reactive Protein predicts incident cardiovascular events and all-cause mortality in the ASCOT Study at 20-years

Abstract Introduction The prediction of future cardiovascular events in those at greatest risk is essential to optimise preventative therapies appropriately and effectively. The value of high sensitivity C-reactive Protein (hsCRP) has been questioned in this regard, and recent studies have suggested the utility of this widely assessed biomarker for predicting future mortality in those presenting with suspected myocardial infarction (MI), however the relationship in stable patients is still debated. Purpose To investigate the role of baseline hsCRP for predicting long-term incident cardiovascular events in hypertensive patients in the Anglo-Scandinavian Cardiac Outcomes Trial (ASCOT), with follow-up extending to 20-years. Methods This ASCOT legacy study reports events after 20-years of the UK participants in the original ASCOT trial. ASCOT was a multicentre randomised trial, randomising patients with hypertension into amlodipine-based or atenolol-based blood pressure-lowering (BPL) treatment. In addition, those with total cholesterol &amp;lt;6.5 mmol/L and no previous lipid-lowering treatment underwent further randomisation to either atorvastatin or placebo as part of the lipid-lowering (LL) arm of ASCOT. We examined outcomes related to hsCRP levels in the LL arm, dichotomously (&amp;lt; or &amp;gt; 2mg/L), in tertiles or continuously, adjusting in Model 1 (age, sex, socio-economic status [years of education] and ethnicity) and Model 2 (Model 1 plus current smoker, body mass index, baseline systolic blood pressure, creatinine, diabetes, history of vascular diseases, history of antihypertensive medication and allocation to BPL and LL). All-cause mortality, non-fatal and fatal MI, total coronary events and procedures and total cardiovascular events were assessed. Results 5,294 participants were included in the final cohort, after exclusion of 3,286 participants in the LL arm (n=8,580) without hsCRP data. There were no substantial differences in baseline characteristics between treatment allocation arms (non-randomised, placebo or atorvastatin). The highest tertile of hsCRP (median [IQR], 6.41 [4.81-10.44]) strongly related to all-cause mortality, withstanding adjustment in both Model 1 (HR 95% CI, 1.38 [1.27-1.53]; p&amp;lt;0.001; p&amp;lt;0.001 for interaction) and Model 2 (HR 1.25 [1.10-1.42]; p&amp;lt;0.001; p&amp;lt;0.001 for interaction). Moreover, the highest hsCRP tertile also related to fatal and non-fatal MI (Model 2 – HR 1.32 [1.05-1.67]; p=0.020; p=0.019 for interaction); total coronary events and procedures (Model 2 – HR 1.27 [1.09-1.47]; p=0.002; p=0.003 for interaction); and total cardiovascular events (Model 2 – HR 1.22 [1.08-1.37]; p=0.001; p=0.001 for interaction). Conclusions This study demonstrates that one hsCRP reading at baseline independently predicts cardiovascular events and all-cause mortality at very long-term follow-up in patients at high-risk for these events, and may help stratify patients into higher risk categories for intensive preventative strategies.

Induced optimism to lessen hopelessness-related cognitions among young adults at risk for suicidal thoughts and behavior

BackgroundHopelessness is one of the best-studied cognitive predictors of depression and suicide. Previous research suggests that hopelessness may develop through repetitive thinking about the occurrence of positive and negative future outcomes. The present study sought to investigate whether mental rehearsal in making optimistic future-event predictions, or induced optimism, would lead to reductions in hopelessness, particularly among individuals with a history of suicide ideation or suicide attempts. MethodsParticipants with (n = 58) and without (n = 76) a history of suicide ideation or attempts were randomly assigned to either practice making optimistic future-event predictions or to a control condition in which they practiced making a lexical decision (using the same stimuli) over three study sessions, each separated by one week. ResultsFindings offered modest support for the hypothesis that induced optimism would decrease hopelessness but not improve mood; this was regardless of history of suicide ideation or attempts. LimitationsThe sample was predominantly female, and assessment of suicide ideation and attempt history was not confirmed by clinical interview, which may limit generalizability. ConclusionPractice in making optimistic future-event predictions over time may be one way to reduce the hopelessness-related cognitions that confer vulnerability to suicide ideation and behavior.

The Beginning of the AI-Enabled Preventative PAP Therapy Era: A First-in-Human Proof of Concept Interventional Study.

Positive Airway Pressure (PAP) therapy is the most common and efficacious treatment for Obstructive Sleep Apnea (OSA). However, it suffers from poor patient adherence due to discomfort and may not fully alleviate all adverse consequences of OSA. Identifying abnormal respiratory events before they have occurred may allow for improved management of PAP levels, leading to improved adherence and better patient outcomes. Our previous work has resulted in the successful development of a Machine-Learning (ML) algorithm for the prediction of future apneic events using existing airflow and air pressure sensors available internally to PAP devices. Although researchers have studied the use of ML for the prediction of apneas, research to date has focused primarily on using external polysomnography sensors that add to patient discomfort and has not investigated the use of internal-to-PAP sensors such as air pressure and airflow to predict and prevent respiratory events. We hypothesized that by using our predictive software, OSA events could be proactively prevented while maintaining patients' sleep quality. An intervention protocol was developed and applied to all patients to prevent OSA events. Although the protocol's cool-down period limited the number of prevention attempts, analysis of 11 participants revealed that our system improved many sleep parameters, which included a statistically significant 31.6% reduction in Apnea-Hypopnea Index, while maintaining sleep quality. Most importantly, our findings indicate the feasibility of unobtrusive identification and unique prevention of each respiratory event as well as paving the path to future truly personalized PAP therapy by further training of ML models on individual patients.