Article
Multidisciplinary Sciences
Jakob Runge
Summary: Detecting and quantifying causal relations in ecosystem functioning is challenging and involves reasoning about underlying assumptions. A global study on grasslands highlights the importance of considering confounding, nonlinearity, and determinism in modern causal inference approaches in ecology.
NATURE COMMUNICATIONS
(2023)
Article
Mathematics, Interdisciplinary Applications
Alvaro Fuentes, Oliver Luedtke, Alexander Robitzsch
Summary: Propensity score methods are widely recommended to adjust for confounding and recover treatment effects. This article reviews propensity score weighting estimators for multilevel data and shows that estimates based on calibration weights should be preferred under many scenarios. Large cluster sizes are needed for accurate estimates of treatment effect when covariate effects vary strongly across clusters.
MULTIVARIATE BEHAVIORAL RESEARCH
(2022)
Article
Mathematical & Computational Biology
Eleanor Pullenayegum, Catherine Birken, Jonathon Maguire, TARGet Kids Collaboration
Summary: Data collected in the context of usual care provide a valuable resource for research. An inverse-weighting approach has been proposed to handle irregular assessment times. In this paper, the approach is extended to handle a special case of non-random assessment. Multiple outputation is used to achieve the same purpose, and an alternative joint model that does not require covariates is developed. The methods are examined through simulation and applied to a study on the causal effect of wheezing on time spent playing outdoors among children.
STATISTICS IN MEDICINE
(2023)
Article
Neurosciences
Fuleah A. Razzaq, Maria L. Bringas Vega, Marlis Ontiveiro-Ortega, Usama Riaz, Pedro A. Valdes-Sosa
Summary: This study investigates the causal relationship between anatomical descriptors of the cingulate cortex and cognitive task performance. It found that the posterior cingulate surface area has a positive causal effect on inhibition and cognitive flexibility, while the anterior cingulate surface area only affects inhibition and cognitive flexibility partially. The curvature-corrected mean thickness showed no causal effect on cognitive tasks.
HUMAN BRAIN MAPPING
(2022)
Article
Public, Environmental & Occupational Health
Vivian C. Wong, Kylie Anglin, Peter M. Steiner
Summary: Recent interest in promoting replication efforts assumes well-established methodological guidance, however, no consensus exists in the methodology literature. This article addresses these challenges by describing design-based approaches for planning systematic replication studies, derived from the Causal Replication Framework (CRF). Testing CRF assumptions systematically evaluates replicability of effects and identifies sources of effect variation in replication failure.
PREVENTION SCIENCE
(2022)
Article
Mathematical & Computational Biology
Di Shu, Peisong Han, Sean Hennessy, Todd A. Miano
Summary: There is increasing interest in developing causal inference methods for multi-valued treatments, particularly focusing on average treatment effects when considering drug-drug interactions (DDIs). This paper proposes two empirical likelihood-based weighting approaches for confounding adjustment in studying the effects of DDIs, and evaluates their performance through simulation. The results demonstrate that these new estimators outperform the standard method in terms of robustness and efficiency. Applying the proposed methods to real-world data, the impact of renin-angiotensin system inhibitors (RAS-I) on the comparative nephrotoxicity of nonsteroidal anti-inflammatory drugs (NSAID) and opioids is evaluated.
STATISTICS IN MEDICINE
(2023)
Article
Substance Abuse
Gary C. K. Chan, Carmen Lim, Tianze Sun, Daniel Stjepanovic, Jason Connor, Wayne Hall, Janni Leung
Summary: This paper introduces the potential outcomes framework for causal inference and summarizes well-established causal analysis methods for observational data in addiction research. It provides examples and analysis codes to assist researchers in conducting these analyses.
Article
Economics
Huijuan Ma, Jing Qin, Yong Zhou
Summary: It is well known that conditioning on covariates can improve the estimation of the marginal outcome distribution. This article establishes a connection between marginal quantile and conditional quantile regression and proposes two novel estimation approaches using conditional quantile regression. The consistency and asymptotic normality of the estimators are proven, and the second approach achieves semiparametric efficiency.
JOURNAL OF BUSINESS & ECONOMIC STATISTICS
(2023)
Article
Computer Science, Artificial Intelligence
Shuzhou Sun, Shuaifeng Zhi, Qing Liao, Janne Heikkila, Li Liu
Summary: This paper presents a debiasing procedure for Scene Graph Generation (SGG) task using causal inference. The proposed Two-stage Causal Modeling (TsCM) can achieve state-of-the-art performance in terms of mean recall rate by addressing both long-tailed distribution and semantic confusion as confounders. TsCM can also maintain a higher recall rate compared to other debiasing methods, indicating a better tradeoff between head and tail relationships.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Microbiology
Jie-Hai Chen, Li-Ying Zeng, Yun-Feng Zhao, Hao-Xuan Tang, Hang Lei, Yu-Fei Wan, Yong-Qiang Deng, Ke-Xuan Liu
Summary: By performing Mendelian randomization (MR) analysis using publicly accessible genome-wide association study (GWAS) summary-level data, this study found suggestive evidence of causal associations between gut microbiota and sepsis risk. Specific gut microbiota were found to be negatively or positively correlated with the risk of sepsis. Multiple statistical methods were used to validate the robustness of the findings.
FRONTIERS IN MICROBIOLOGY
(2023)
Article
Physics, Multidisciplinary
Sergei Kalinin, Ayana Ghosh, Rama Vasudevan, Maxim Ziatdinov
Summary: The development of high-resolution imaging methods has provided rich information on the atomic structure and functionalities of solids, necessitating the adaptation of classical macroscopic definitions to understand local imaging data. This data can be used to construct statistical and physical models that generate observed structures. The availability of observational data opens pathways for exploring causal mechanisms underlying solid structure and functionality.
Article
Environmental Sciences
Kathrin Wolf, Sophia Rodopoulou, Jie Chen, Zorana J. Andersen, Richard W. Atkinson, Mariska Bauwelinck, Nicole A. H. Janssen, Doris Tove Kristoffersen, Youn-Hee Lim, Bente Oftedal, Maciek Strak, Danielle Vienneau, Jiawei Zhang, Bert Brunekreef, Gerard Hoek, Massimo Stafoggia, Evangelia Samoli
Summary: Most studies on the health effects of long-term exposure to air pollution have used traditional regression models, but few have applied causal inference approaches. In this study, we compared the associations between exposure to PM2.5 and NO2 and natural-cause mortality using both traditional Cox and causal models in a large cohort setting. The results showed consistent associations between air pollution exposure and natural-cause mortality using both approaches, although there were some differences in estimates among individual cohorts. Multiple modeling methods may help improve causal inference.
ENVIRONMENTAL POLLUTION
(2023)
Article
Computer Science, Artificial Intelligence
Tao Zhang, Hao-Ran Shan, Max A. Little
Summary: This paper introduces a graph neural network model called Causal GraphSAGE (C-GraphSAGE) that incorporates causal inference into the sampling stage of GraphSAGE to improve classifier robustness. Experimental results demonstrate that C-GraphSAGE outperforms other graph neural network models in terms of classification performance when perturbation is present.
PATTERN RECOGNITION
(2022)
Article
Computer Science, Artificial Intelligence
Ciaran M. Gilligan-Lee, Christopher Hart, Jonathan Richens, Saurabh Johri
Summary: The article introduces a new method for detecting hidden common causes in causal relationships. The method modifies existing algorithms to be able to distinguish both purely directed causal relationships and latent common causes. Through testing on synthetic and real data, the effectiveness and performance of this method are demonstrated.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)
Article
Health Care Sciences & Services
D. Zugna, M. Popovic, F. Fasanelli, B. Heude, G. Scelo, L. Richiardi
Summary: This study compared four statistical methods to analyze the mediating role of adverse reproductive outcomes and infant respiratory infections in the effect of maternal pregnancy mental health on infant wheezing. The results revealed a small mediating effect of these factors on the outcome. The choice of method depends on the main effect of interest, the type of variables involved, and confidence in specifying the models for the exposure, mediators, and outcome.
BMC MEDICAL RESEARCH METHODOLOGY
(2022)