I'm afraid there's no right or wrong answer on that. As pointed out by Klaas Stefan in his "Ten Simple Rules" paper, DCM is used to provide a connectivity-based explanation for the effects that you see. So if there's no effect to explain, DCM won't give useful results.
Note that if many of your subjects don't show activation in an ROI, it could mean they are using alternative strategies / brain regions to perform your task, which you could explore using SPM and DCM.
Best,
P.
-----Original Message-----
From: Fuyu [mailto:[log in to unmask]]
Sent: 17 December 2014 14:05
To: Zeidman, Peter
Cc: [log in to unmask]
Subject: Re: SPM12
Dear Peter,
I tried various thresholds and it seems like for some subjects, the ROI only appears at 0.1. Is that too liberal a threshold at the subject level, or the threshold at subject level doesn't matter?
Many thanks.
Kind Regards,
Fuyu
On 17 Dec, 2014, at 8:44 pm, Zeidman, Peter <[log in to unmask]> wrote:
> Dear Fuyu,
> You're very welcome. I think you are asking, if you do not get activation at the single subject level within your region of interest, should I include this in my DCM analysis?
>
> One option is to exclude the subject, however that may be overly conservative - you may only find a significant effect at the group level rather than the single subject level. In this case, you could relax the statistical threshold at the single subject level for positioning your ROIs. Just make sure you stay within the area of the group level significant activation.
>
> Best,
> Peter.
>
> -----Original Message-----
> From: Fuyu [mailto:[log in to unmask]]
> Sent: 17 December 2014 12:42
> To: Zeidman, Peter
> Cc: [log in to unmask]
> Subject: Re: SPM12
>
> Dear Peter,
>
> Thanks a lot for your prompt reply, you've been most helpful.
>
> Also, just to clarify, if the VOI of the subj is not within the ROI, do we exclude the subject from the DCM analyses?
>
> Many thanks.
>
> Kind Regards,
> Fuyu
> On 17 Dec, 2014, at 8:27 pm, Zeidman, Peter <[log in to unmask]> wrote:
>
>> Dear Fuyu (please CC queries to the SPM mailing list), If you click
>> "current cluster" in the results window, you should get the results table appear, and it should also jump to a peak. If not, you can click the coordinates in the table to make it jump.
>>
>> Best,
>> Peter.
>>
>> -----Original Message-----
>> From: Fuyu [mailto:[log in to unmask]]
>> Sent: 17 December 2014 11:55
>> To: Zeidman, Peter
>> Subject: SPM12
>>
>> Dear Peter,
>>
>> I'm encountering some issues with SPM 12, which I hope you could kindly advice me on.
>>
>> I'm currently creating VOI at a subject level for my DCM analyses. In the past, when I enter a coordinate for an ROI (Eg. -15,-9,6 for thalamus), if there is no activation found at the coordinate, the red arrow will point to the nearest coordinate that has activation. Unfortunately, I've been trying a few times but the red arrow doesn't seem to bring me to the nearest point of activation.
>>
>> It would be great to get your advice on this matter.
>>
>> Many thanks.
>>
>>
>> Kind Regards,
>> Fuyu
>>
>
> Kind Regards,
> Fuyu
>
Kind Regards,
Fuyu
|