this is for holding javascript data
Anisha Keshavan edited We_needed_to_emphasize_that__.tex
over 8 years ago
Commit id: 36d865aed78af5815e99e57f9685aee7b4a391a4
deletions | additions
diff --git a/We_needed_to_emphasize_that__.tex b/We_needed_to_emphasize_that__.tex
index 9ef8874..d3ac4b8 100644
--- a/We_needed_to_emphasize_that__.tex
+++ b/We_needed_to_emphasize_that__.tex
...
We needed to emphasize that the The number of phantom subjects do not contribute to the power equation in Figure 1. The power equation
we that was derived does not account for any sort of calibration or scaling. Instead, it requires an estimate for $CV_a$, which is the variability of scaling biases between sites.
Because we are not asking our Our consortium
was not asked to change protocols,
we thought that our so the true
bias variability
would certainly might be higher than that of a harmonized study, and sample sizes we would calculate would be smaller
than we what is actually
need needed if
we used harmonized data
was used to calculate
it. bias. For other researchers planning multisite studies,
they could use our measurements of $CV_a$
could be used to get a better power
estimates if the estimate for protocols
in their consortiums that are close to the 3D-MPRAGE protocols
that were described here. This is especially useful
if they want to include for the inclusion of retrospective
data in their analyses. data.
To power an estimate of $CV_a$,
we'd need enough subjects to get accurate estimates of scaling factors within sites, two sample sizes are required - the subject level, and
enough sites to get a better variance estimates. the site level. The sample size of 12 is enough to estimate the slope of the calibration line (which is around 1), because the scan-rescan reliability of the ROIs is very
high. high (around 0.9). Ideally, we would have more than 12 subjects to get better a better estimate of the
intercept (This is intercept, and this was listed as a study limitation in the
discussion). We sampled from discussion. 20 sites
were sampled because these same sites will be used in a future study on MS genetics and MRI phenotypes, although the inclusion of more sites would also improve the $CV_a$ estimate.
In the text,
we've it was emphasized that the derived power equation does not require phantom subjects or calibration to be used:
"\textit{We emphasize that the use of phantom subjects do not directly contribute to the power equation in Figure 1, as it does not account for any sort of calibration or scaling. However, it requires an estimate for $CV_a$, which is the variability of scaling biases between sites. The goal of this study is to provide researchers with estimates of $CV_a$ from our set of calibration phantoms and our set of non-standardized MRI acquisitions. For a standardized set of scanners, these values may be considered an upper bound. }"