Is R^2 statistiques in 3ddeconvolve the percent of variance explained ?

AFNI Message Board
,
1

AFNI version info (afni -ver): (version 7.12, April 14, 2020)

Hello folks,

I need to compute some statistics on 3dDeconvolve in AFNi.
I want to compute the percent of variance explained for different regressions on same data.

How do we calculate this ?
I found on internet that this might correspond to the R^2 from the output of 3dDeconvolve.
But in 3dDeconvolve's help, it is mentioned that these statistics are marginal and that for more clarification, I should see with the AFNI message board guru entities (when they can be lured from the peak of Mt Taniquetil or Kailash).

So I am summoning you for a council.

Please :D.
Any help or explanation would be greatly appreciated.

2

What is your concern regarding marginal effects?

You probably don't to embark on an epic quest to the peaks of Mt. Taniquetil or Kailash to uncover the secrets of the coefficient of determination R^2. Add the option -rout to your 3dDeconvolve script, and the coefficient of determination R^2 will be provided together with each individual statistic such as t or F.

Gang

3

Hello Gang,

Thanks for your reply. To give some precision, I'm trying to reproduce some analysis.
I ran the same protocol, and in their analysis, they did "the amount of variance explained
by the regressor was computed" cf. Improving the Assessment of Breath-Holding Induced Cerebral Vascular Reactivity Using a Multiband Multi-echo ASL/BOLD Sequence | Scientific Reports

But I don't know if this refers to r2 results from regressor or not.

When we talk about amount of variance explained, does it refer to r2 ?

Thanks

4

See the discussion on this Wikipedia page.

Gang

1 Like
5

Perfect, thanks