Dear Bing Yu,
To analyse your data using the "partitioned error" approach, you would
specify two second level models, one for each effect of interest.
https://en.wikibooks.org/wiki/SPM/Group_Analysis
For the main effect of group, at the first level, compute the average
effect over conditions (with t contrast [1 1 1]) for each patient and
control, and enter these images in a two-sample t-test at the second
level (then testing using a contrast [1 -1]).
For the group x condition interaction, at the first level, compute two
differential effects for your three conditions (with t-contrasts [1 -1
0] and [0 1 -1]) and enter the two contrast images in a two-way (2x2)
ANOVA (using a Full Factorial design) at the second level. The contrast
to test for an interaction would then be [1 0 -1 0; 0 1 0 -1].
Best regards,
Guillaume.
On 05/09/17 09:03, 于兵 wrote:
>
> Dear Chih-Hao:
>
> The article is of great help to me. Thanks for your direction.
>
> Bing Yu
>
>
>
>
> At 2017-09-05 10:15:22, "連 志浩" <[log in to unmask]> wrote:
>
> Dear Bing Yu,
>
>
> I think the information you need can be found in this paper.
>
>
> Gläscher, J., & Gitelman, D. (2008). Contrast weights in flexible
> factorial design with multiple groups of subjects. /SPM@ JISCMAIL.
> AC. UK) Sml, Editor/.
>
> Best regards,
> Chih-Hao
>
> ------------------------------------------------------------------------
> *寄件者:* SPM (Statistical Parametric Mapping) <[log in to unmask]
> <mailto:[log in to unmask]>> 代表 Bing Yu <[log in to unmask]
> <mailto:[log in to unmask]>>
> *寄件日期:* 2017年9月5日 上午 12:28:54
> *收件者:* [log in to unmask] <mailto:[log in to unmask]>
> *主旨:* [SPM] Analysis of repeat measured fMRI Data
>
> Dear SPM experts:
>
> Now I am performing a fMRI research, to detect the
> abnormalities in some pediatric patients. The subject were divided
> into 2 groups (patients, controls) ,they all underwent fMRI scan in
> 3 conditions. Now I will perform repeat measured analysis using SPM
> "Flexible factorial" design and we will focus on the brain area
> exhibit interactions (group* conditions) and main effect of groups.
>
> I will be very appreciate if I got instructions from how to
> define the "effects" in specify 2nd level step and how to define the
> contrast after model estimation.
>
> Best regards,
> Bing Yu
>
>
>
>
>
>
--
Guillaume Flandin, PhD
Wellcome Trust Centre for Neuroimaging
University College London
12 Queen Square
London WC1N 3BG
|