Yes. This looks correct.
Good luck
darren
-----Original Message-----
From: Shoichi Ugai [mailto:[log in to unmask]]
Sent: Friday, November 17, 2006 12:07 AM
To: [log in to unmask]; Darren G
Subject: Re: DCM model
Dear Darren
You helped me a great deal.
For your advice, I gradually understand how to use DCM in SPM2.
I revised my model as follows:
DCM.a = [1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1]
DCM.b = [0 0 0 0 1
0 0 0 0 1
0 0 0 0 1
0 0 0 0 1
0 0 0 0 0]
DCM.c = [1 1 0 0 0]'
ROIs were
R.FFA
L.FFA
R.AMY
L.AMY
PFC
direct input:R.FFA and L.FFA
modulatory effect:PFC->R.FFA and R.AMY; PFC->R.AMY and L.AMY
Because, I do not know the intrinsic connections among the ROIs, we assumed
that all the intrinsic connection among the ROIs are bidirectional.
Those are my procedure of analysis.
I would like to confirm if this procedure is correct.
All the regressors of GLM analyses are common to all the subjects.
(1) We determined ROIs by conventional GLM analyses.
(2) Five ROIs were selected, and were conducted DCM analyses
with above parameters. The coordinates for each region were common
to all the subjects.
(3)To test existence of significant modulation activities, DCM.B(1,5),
DCM.B(2,5), DCM.B(3,5), DCM.B(4,5) are tested with one-sample t-test.
If I want to compare the strength of connection, we compare the two
parameters (for example DCM.B(1,5) and DCM.B(2,5)) using paired-t test.
(4)Finally we assesed the significance at the threshold P<0.05
(corrected if possible).
Thank you in advance,
Shoichi
>Dear Shoichi:
>I am still confused about your model, but here is my guess.
>Input effects are on regions (DCM.c matrix). Intrinsic connections are
>between regions (DCM.a), and modulatory effects are on the intrinsic
>connections (DCM.b)
>So from what you have said, it appears that you have have 1 input
condition,
>that of fearful faces and 0 modulatory effects.
>I assume that in the matrices o = 1 and x = 0; by the way it would be
easier
>to read if you send the matrices as 1's and 0's.
>DCM.c = [1 1 1 1 0]'; (inputs to all regions EXCEPT the PFC)
>In general one might consider just 1 or perhaps 2 input areas. If inputs go
>to all, or nearly all areas, it turns the DCM analysis into somewhat of a
>conventional SPM analysis in the sense that activity in SPM's are analogous
>to stimulus input to all regions (see Friston et al., Neuroimage, vol 19,
>page 1275 (bottom first - top second column), 2006.
>DCM.a = [1 1 1 1 0
> 1 1 1 1 0
> 1 1 1 1 1
> 1 1 1 1 1
> 0 0 0 0 1]
>DCM.b = [0 0 0 0 0
> 0 0 0 0 0
> 0 0 0 0 0
> 0 0 0 0 0
> 0 0 0 0 0]
>If I have read the matrices correctly the other problem is that you have
>inputs into all areas except PFC, none of the areas connect to the PFC, but
>the PFC connects to R and L amygdala. In that case, PFC is receiving no
>inputs either directly or indirectly. Thus it can produce no output.
>
>Darren.
|