Statistical Methods and Bayesian Inference

EPHierStats: a statistical tool to model the hierarchical relationships in electrophysiological data

EPHierStats models hierarchical relationships in electrophysiological data to prevent pseudoreplication and improve statistical accuracy in neuroscience studies.

Evidence for the null hypothesis in functional magnetic resonance imaging using group-level Bayesian inference

This paper compares Bayesian parameter inference with classical null hypothesis testing in fMRI group analysis, arguing that Bayesian methods provide more intuitive evidence for both null and alternative hypotheses.

Bayesian inference of population prevalence

This paper introduces a Bayesian method to estimate population prevalence from individual participant data, offering a quantitative estimate with uncertainty instead of binary inferences.

Multi-Level Multi-Growth Models: New opportunities for addressing developmental theory using longitudinal designs

This paper introduces multi-level multi-growth models to address developmental theory using longitudinal designs, demonstrating their ability to model multiple growth processes affecting a single outcome.

OCRbayes: A Bayesian hierarchical modeling framework for Seahorse extracellular flux oxygen consumption rate data analysis

OCRbayes is a Bayesian framework for analyzing Seahorse XF OCR data, accounting for complex data structures to improve statistical inference.

Joint Modeling of Zero-inflated Longitudinal Proportions and Time-to-event Data with Application to a Gut Microbiome Study

The paper introduces JointMM, a joint modeling framework for analyzing longitudinal microbiome proportions and time-to-event data, showing its effectiveness in identifying microbial associations with disease progression.

Using joint models to adjust for informative drop-out when modelling a longitudinal biomarker: an application to type 2 diabetes disease progression

The paper applies joint longitudinal-survival models to adjust for censoring bias in HbA1c trajectory analysis in type 2 diabetes patients, finding a higher gradient with the joint model compared to linear mixed effects models.

partR2: Partitioning R2 in generalized linear mixed models

partR2 quantifies part R2 and structure coefficients for fixed effects in generalized linear mixed models, providing variance decomposition and confidence intervals.

Bayesian Multivariate Growth Mixture Modeling of Longitudinal Data: An Application to Alzheimer's Disease Study

The paper applies a Bayesian multivariate growth mixture model to identify cognitive decline patterns in Alzheimer's Disease Neuroimaging Initiative data.

Trimmed Constrained Mixed Effects Models: Formulations and Algorithms.

This paper introduces Trimmed Constrained Mixed Effects Models with efficient algorithms and an open-source Python package, showing improved accuracy and computational efficiency in handling outliers and nonlinear data.

Application of the hierarchical bootstrap to multi-level data in neuroscience

Treating neurons as independent gave about 46–96% false positives in simulated nested data; a hierarchical bootstrap stayed near 5% with many subjects and kept more power than per-subject averaging.

A Bayesian predictive approach for dealing with pseudoreplication

The paper introduces a Bayesian predictive approach to address pseudoreplication in biological research, demonstrating its effectiveness with two in vivo datasets.

A framework for testing different imputation methods for tabular datasets

The paper develops a framework to test imputation methods on a clinical stroke dataset and finds that listwise deletion performs poorly, while the best method depends on the missing data mechanism.

Hierarchical Bayesian inference for concurrent model fitting and comparison for group studies

This paper introduces a hierarchical Bayesian inference framework for simultaneous model comparison and parameter estimation in group studies, showing improved accuracy and robustness.

Highly Efficient Hypothesis Testing Methods for Regression-type Tests with Correlated Observations and Heterogeneous Variance Structure

The paper introduces efficient hypothesis testing methods for regression-type tests with correlated data and heterogeneous variance, outperforming existing methods in speed and accuracy.

Bayesian Model Selection Maps for group studies using M/EEG data

This paper introduces a Bayesian Model Selection method for group-level M/EEG data analysis using SPM, enabling hypothesis comparison across brain regions and participants.

Exact Tests of Zero Variance Component in Presence of Multiple VarianceComponents with Application to Longitudinal Microbiome Study

The paper develops exact tests for variance components in longitudinal microbiome studies and applies them to HIV-infected patients, identifying associations between lung microbiome genera and lung function.

Characterizing and Managing Missing Structured Data in Electronic Health Records

The paper evaluates imputation methods for missing structured data in EHRs using clinical lab measures from 602,366 patients, showing some methods perform well while others are unsuitable for multiple imputation.

Multiple Linear Regression: Bayesian Inference for Distributed and Big Data in the Medical Informatics Platform of the Human Brain Project

Variational Bayesian parameter estimation techniques for the general linear model

The paper provides a theoretical and practical analysis of variational Bayes, VML, ReML, and ML methods applied to the general linear model in fMRI data analysis.

When null hypothesis significance testing is unsuitable for research: a reassessment

The paper critiques null hypothesis significance testing (NHST) and advocates for alternative statistical methods in biomedical and psychological research.

A Bayesian approach to multivariate and multilevel modelling with non-random missingness for hierarchical clinical proteomics data

The paper introduces a Bayesian multivariate multilevel model to analyze hierarchical clinical proteomics data with non-random missingness, showing improved performance over existing methods.