Journal: Frontiers in Neuroscience

Article Title: A Comparison of Functional Networks Derived From Representational Similarity, Functional Connectivity, and Univariate Analyses

doi: 10.3389/fnins.2019.01348

Figure Lengend Snippet: Representational similarity, univariate and functional connectivity analyses. Visualization of the several steps of the methodologies of these analyses to clarify the (dis)similarities. Gray and white squares drawn on the brain depict two random regions of interest (ROIs), consisting of 9 voxels each. Within these squares, colors in the 3 × 3 matrix depict the BOLD signal intensity of the voxels in a condition 1, 2, or 3, modeled through a general linear model (GLM). High intensity is shown in red, medium intensity in yellow, low intensity in blue. (A) In univariate analyses (UNIVAR), the signal is averaged across all voxels within the ROI, and this for every ROI and every condition. This step is visualized by large squares next to each matrix with a color depicting the calculated average. Consequently, these results are averaged across participants. To understand if the ROIs are similarly activated in light of the conditions, the results of every ROI can be correlated to each other. (B) In representational similarity analysis (RSA), the signal is not averaged across voxels within ROIs. Instead, the signal of an array of voxels in a ROI in light of a certain condition is correlated to that of another condition. All possible pairwise comparisons are made. This step is depicted by a square matrix containing the correlation results for each condition pair: the color of each element of the matrix indicates the strength of the correlation (black indicates a correlation of 1 and white indicates 0). These results show if conditions are represented similarly within a ROI. They are then averaged across subjects. Then, these results can be correlated between ROIs to understand if different ROIs represent the conditions in a similar way. (C) In functional connectivity analysis, the signal of the voxels in a ROI at different time-points (i.e., change in BOLD signal intensity over time for each voxel) is averaged across all voxels within that ROI. This step is visualized by large squares with a color depicting the calculated average. After, the resulting average BOLD signal time course can be correlated between ROIs and consequently averaged across subjects. This analysis method depicts if two ROIs communicate.

Article Snippet: Univariate analysis (UNIVAR) and functional connectivity analysis (FCA) are two other popular methods to identify functional networks.

Techniques: Functional Assay

Journal: AMIA Summits on Translational Science Proceedings

Article Title: Factors Affecting Electroconvulsive Therapy Ictal Outcomes: Duration and Postictal Suppression

doi:

Figure Lengend Snippet: Mixed-effects model coefficients and p-values for univariate and multivariate linear regressions of ictal duration and postictal suppression index. Significance: p <0.05*, p <0.01**, p <0.001***. Univar coef = Univariate mixed model coefficients; Selective Multivar coef = multivariate model coefficients including age, sex, and significant univariate variables; Complete Multivar coef = multivariate model coefficients including all univariate model variables.

Article Snippet: Univar coef = Univariate mixed model coefficients; Selective Multivar coef = multivariate model coefficients including age, sex, and significant univariate variables; Complete Multivar coef = multivariate model coefficients including all univariate model variables.

Techniques:

Journal: AMIA Summits on Translational Science Proceedings

Article Title: Factors Affecting Electroconvulsive Therapy Ictal Outcomes: Duration and Postictal Suppression

doi:

Figure Lengend Snippet: Restricted Maximum Likelihood (REML) of univariate and multivariate mixed-effects models for (A) ictal duration, and (B) postictal suppression, rank ordered. REML is the unbiased estimate of variance and covariance parameters for mixed-effects models, with lower REML relating to better fit of the data.

Article Snippet: Univar coef = Univariate mixed model coefficients; Selective Multivar coef = multivariate model coefficients including age, sex, and significant univariate variables; Complete Multivar coef = multivariate model coefficients including all univariate model variables.

Techniques:

Journal: Frontiers in Veterinary Science

Article Title: The Prevalence of Integument Injuries and Associated Risk Factors Among Canadian Turkeys

doi: 10.3389/fvets.2021.757776

Figure Lengend Snippet: Explanatory variables associated ( P ≤ 0.25) with the presence of head/neck (HN) injuries at the univariable level.

Article Snippet: Univariable and multivariable models testing the association of the variables with injuries were performed using RStudio (version 3.5.3).

Techniques:

Journal: Frontiers in Veterinary Science

Article Title: The Prevalence of Integument Injuries and Associated Risk Factors Among Canadian Turkeys

doi: 10.3389/fvets.2021.757776

Figure Lengend Snippet: Explanatory variables associated ( P ≤ 0.25) with the presence of back/tail (BT) injuries at the univariable level.

Article Snippet: Univariable and multivariable models testing the association of the variables with injuries were performed using RStudio (version 3.5.3).

Techniques: