Dear Takanori,
Thanks for the explanations.
I had believed that "adjusted" is to remove the effect of
injected dose's diffences among scans.
But now I can understand the meaning of "adjusted".
"Adjuted" is to remove "effects of no interest".
In your explanation, mean response for each subject is described $B!H (BB*b2 $B!I (B.
The mean response is an average of all voxel values in a brain, isn $B!G (Bt it ?
If my understanding is correct, what is the advantage to calculate adj_Y ?
To reduce noise ?
>Some explanations are below.
>Suppose you extract the data from only one voxel
>First, you set up general linear model :
>Y = X*b+e
>Desgin matrix X (and parameters b) is partitioned into
>effects of interest: H (b1) and effects of no interest: B (b2).
>X = [H B]
>b =[b1 b2]
>In simple PET design, H is a matrix including dummy
>variables for several conditions. B is subject's block.
>You can rewrite
>Y = H*b1+B*b2+e.
>If you choose "Adjusted data for effects of interest"
>in V.O.I. data extraction, SPM calculate
>adj_Y = Y - B*b2,
>that is, mean response for each subjects is subtracted from
>(globally normalized) raw data.
>If you choose "don't adjust" ,the extracted data is
>(globally normalized) raw data.
>Using this data, you may calculate
>"Percent signal change" as follows.
>%Signal change for task_1 = (mean(Y_task_1)-mean(Y_rest))/mean(Y_rest)
>This value is also useful for VOI analysis.
>
>Takanori Kochiyama
>-----------------------------------------------
>Graduate School of Human and Environmental Studies
>Kyoto University
>E-mail [log in to unmask]
>Tel +81 75 753 7862
>
_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/
Hideki Mochiduki. E-mail: [log in to unmask]
Department of Pharmacology
Tohoku University School of Medicine
ADRESS: 2-1 Seiryo-machi, Aoba-ku, Sendai, 980-8575,Japan
TEL:+81-(0)22-717-8057
FAX:+81-(0)22-717-8059
_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/
|