When selection of trials was modeled by trial size and intensity of treatment, the most powerful test was the arcsine test.
When selection of trials was modeled by trial size and treatment effect magnitude, the power adjusted for type I error rate increased when the true average odds ratios moved away from 1.
We addressed this issue by generating data under two previously used selection models: by trial size and intensity of treatment effect [ 9, 31, 32] or by p-value for the treatment effect associated with the trial [ 5, 7] and we found similar results.
To model trial-by-trial adaptation, I used a state-space approach [11], in which the initial direction of the left and the right hand (y) was expressed as a function of the force applied to the hand (a) and the internal state of the system (z), (4) [ y L y R ] n = [ D L 0 0 D R ] [ a L a R ] n − [ z L z R ] n, in which D L and D R relate to the stiffness of the left and right arm.
Briefly, all data from individual participants were combined in a single meta-analysis based on a simple Cox proportional regression model stratified by trial, containing random effects (frailty models) to assess the impact of stent type for instance, cobalt-chromium everolimus eluting stent versus bare metal stent on outcome measures.
Furthermore, these scenarios presented computational problems when fitting a linear mixed model, while the by-trial model is computationally straightforward.
Manual FE model tuning by trial and error found that flexible supports in the longitudinal direction should be introduced at the girder ends to improve correlation between the measured and FE-calculated modes.
In the analysis adjusting for patient characteristics, individual patient time-to-event data were synthesised across trials using a Cox proportional hazard model stratified by trial.
Recent theoretical work on perceptual uncertainty advocates the utility of modelling trial-by-trial updating across individuals in a Bayesian framework [ 53], which is argued to be of significant benefit in conceptualising the current findings from the signal detection task.
The results highlight the importance of modeling trial-by-trial RT in fMRI analyses and raise the possibility that RT variability may provide a powerful probe for investigating the previously elusive white matter BOLD signal.
