Thank you for your clarification, Steve.
Yes, my second level analysis pools over runs for each subject with the
contrasts setup to compare the two levels of factor A. And the resulting
cope images from all subjects are fed up into my third level analysis to
pool across subjects.
In order to use RT as covariates in this third level analysis, I made
EV2 and put in RT differences between level 2 and level 1 and made it
orthogonal to EV1 (which is a column of 1s).
Does this sound OK?
Soohyun.
-----Original Message-----
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On
Behalf Of Steve Smith
Sent: Monday, May 15, 2006 11:15 PM
To: [log in to unmask]
Subject: Re: [FSL] RT as covariate within third level analysis
contrasting conditions
Hi, it's not quite clear to me what model you used at second-level
and what is the proposed third-level model. If at second level you
have achieve the level2 vs level 1 contrast, and are feeding this
differential response as a single datapoint for each subject up to
the third level, this all sounds fine - then at third level you can
have a constant height (all 1s) mean differential effect EV1, and EV2
should be the demeaned RT (or RT differential) values.
Does this sound right? Cheers, Steve.
On 14 May 2006, at 03:42, Soohyun Cho wrote:
> Hi,
>
>
>
> I have a question about using RT as covariates in the third-level
> analysis (pooling across subjects).
>
> I have a within subject design, in which the subject data is
> collected under 2 levels of one factor.
>
> For simplicity, let's call it Factor A having level1 & level2.
>
>
>
> I am interested in finding brain regions that show greater
> activations for level 2 vs. level 1 of Factor A.
>
> My third-level analysis is loading up cope images from the second
> level that show greater activations for level 2 compared to level1.
>
> If I want to see whether the degree of differential brain
> activation correlate with difference in RT between levels,
>
> should I calculate the RT difference between the two levels and
> demean those scores and put them into EV2 orthogonalized with EV1?
>
>
>
> Thank you so much for your help,
>
> Soohyun.
>
>
>
>
------------------------------------------------------------------------
---
Stephen M. Smith, Professor of Biomedical Engineering
Associate Director, Oxford University FMRIB Centre
FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
[log in to unmask] http://www.fmrib.ox.ac.uk/~steve
------------------------------------------------------------------------
---
|