Applying univariate vs. multivariate statistics to investigate therapeutic efficacy in (pre)clinical trials: A Monte Carlo simulation study on the example of a controlled preclinical neurotrauma trial

Abstract

Background Small sample sizes combined with multiple correlated endpoints pose a major challenge in the statistical analysis of preclinical neurotrauma studies. The standard approach of applying univariate tests on individual response variables has the advantage of simplicity of interpretation, but it fails to account for the covariance/correlation in the data. In contrast, multivariate statistical techniques might more adequately capture the multi-dimensional pathophysiological pattern of neurotrauma and therefore provide increased sensitivity to detect treatment effects. Results We systematically evaluated the performance of univariate ANOVA, Welch's ANOVA and linear mixed effects models versus the multivariate techniques, ANOVA on principal component scores and MANOVA tests by manipulating factors such as sample and effect size, normality and homogeneity of variance in computer simulations. Linear mixed effects models demonstrated the highest power when variance between groups was equal or variance ratio was 1:2. In contrast, Welch's ANOVA outperformed the remaining methods with extreme variance heterogeneity. However, power only reached acceptable levels of 80% in the case of large simulated effect sizes and at least 20 measurements per group or moderate effects with at least 40 replicates per group. In addition, we evaluated the capacity of the ordination techniques, principal component analysis (PCA), redundancy analysis (RDA), linear discriminant analysis (LDA), and partial least squares discriminant analysis (PLS-DA) to capture patterns of treatment effects without formal hypothesis testing. While LDA suffered from a high false positive rate due to multicollinearity, PCA, RDA, and PLS-DA were robust and PLS-DA outperformed PCA and RDA in capturing a true treatment effect pattern.

INTACT vs. FANS for Cell-Type-Specific Nuclei Sorting: A Comprehensive Qualitative and Quantitative Comparison

“…Additionally, 3’UTR distribution bias was observed in the Clontech library preparation as previously reported in poly(A) based strategies [ 3 ]. Using principal component analysis (PCA) [ 36 ], we observed that all samples clustered according to their isolation technique and library preparation strategy ( Supplementary Materials Figure S4C ). To avoid 3’UTR distribution bias and since we were using low-input material isolated from the selected neuronal populations, we decided to use the Nugen Ovation Solo as our library preparation strategy for the following transcriptional analyses.…”

Neuroimaging signatures predicting motor improvement to focused ultrasound subthalamotomy in Parkinson’s disease

“…First, the study’s sample size is relatively small. As considering more than one variable increases sensitivity in multivariate analyses, our approach with multivariate cross-correlation has benefits in overcoming small sample sizes with a high amount of features as well as detecting treatment effects with longitudinal data 55 . Second, although the primary goal was to evaluate the predictive capacity of imaging rather than to determine the factors contributing to clinical outcomes, measures of lesion topography shall be included in future research.…”

Statistical and Artificial Neural Networks Models for Electricity Consumption Forecasting in the Brazilian Industrial Sector

“…The literature concerning time-series forecasting has to a heterogeneous and dynamic degree taken into account an extensive amount of scientific production and competitions for evaluating models applied to observed data in different fields of knowledge [19][20][21]. It is possible to find in the literature prediction comparisons in which univariate time-series models are superior to (or as good as) multivariate models (as in the discussion proposed in [22][23][24][25]). A possible interpretation for this result is a sparse representation, in large-scale models, of the dynamic interactions in a system of variables [26].…”