As Isabelle said, including an explicit mask in your model should
control for stray clusters, and you can also lower the threshold
(maybe not -inf, but perhaps to 0.5).
Best,
Leyla
On Fri, Feb 3, 2023 at 2:50 PM SUBSCRIBE SPM Terry Zhang
<[log in to unmask]> wrote:
>
> Dear SPM experts,
>
> I'm doing my first task-fMRI study using SPM 12 and am quite confused about masking. I know there are many posts about this issue in the forum, but I still can't come to an accurate conclusion after reading them, so I hope to get some help here.
>
> By default, SPM uses an implicit mask with a threshold = 0.8 in the first-level analysis. I have two questions about this:
> 1. The implicit mask often dropout some important brain areas, such as OFC. Some people strongly suggest using an explicit mask with threshold = -inf, instead of the implicit mask. I’m confused about which method is more accepted in the community.
> 2. When setting the threshold to 0.8, the implicit mask usually includes a lot of white matter and CSF (in my case, the whole brain is included). So, in the results of first and second-level analyses, there will also be clusters composed of WM, or clusters containing both GM and WM. Should I report these results as significantly activated brain regions? Or should I apply the subject's gray matter (such as the wc1 image from segmentation) as a mask somewhere?
>
> Any answer or discussion will be highly appreciated. Thanks in advance.
>
> With best wishes,
> Terry
--
Leyla Loued-Khenissi, PhD
Telephone: +41 (0) 78 7344217
Email: [log in to unmask]
@NeuroLeyla
https://github.com/LLouedKhen/
www.zeitheist.org
*** If this is a work-related email received outside of your normal
working hours, please feel free to respond at a time convenient to
you***
|