sdm-help-list > Widespread activations despite strict thresholding - results checking?
Showing 1-1 of 1 posts
Sep 22, 2021 09:09 AM | Molly Rowlands
Widespread activations despite strict thresholding - results checking?
Hello,
Thanks again for developing this great software and providing easy-to-follow tutorials alongside it. I'm just re-posting this as there has been no response for a few weeks and I'm still very stuck:
I'm using SDM-Psi v6.21, and I'm concerned about the extent of the spread of the activations the meta-analysis produces - especially as the network involved in what I am meta-analysing is very well outlined thus far and is usually to a much less extent as SDM suggests, e.g. some extra regions being shown by SDM that usually haven't survived single studies nor other meta-analyses (albeit using different meta-analytic software, e.g. GingerALE). I've attached some photos (on this post and replied posts) of the extensive activations I'm talking about, with a strict threshold of tfce-corrected p .0001. From your experience, would you say the extent of these activations is unusual, or could this be a valid result?
It's my first time using this software, so I'm concerned I've made an error that may contribute to the extensive activations - is it possible to have these results verified somehow? I'm happy to send on the input data and other needed details if possible.
(In response to a previous forum post re large activation clusters, I've reduced the FWHM at pre-processing from 20 to 10 - but this did not limit the size/spread of the activations by much)
Best wishes,
Molly
Thanks again for developing this great software and providing easy-to-follow tutorials alongside it. I'm just re-posting this as there has been no response for a few weeks and I'm still very stuck:
I'm using SDM-Psi v6.21, and I'm concerned about the extent of the spread of the activations the meta-analysis produces - especially as the network involved in what I am meta-analysing is very well outlined thus far and is usually to a much less extent as SDM suggests, e.g. some extra regions being shown by SDM that usually haven't survived single studies nor other meta-analyses (albeit using different meta-analytic software, e.g. GingerALE). I've attached some photos (on this post and replied posts) of the extensive activations I'm talking about, with a strict threshold of tfce-corrected p .0001. From your experience, would you say the extent of these activations is unusual, or could this be a valid result?
It's my first time using this software, so I'm concerned I've made an error that may contribute to the extensive activations - is it possible to have these results verified somehow? I'm happy to send on the input data and other needed details if possible.
(In response to a previous forum post re large activation clusters, I've reduced the FWHM at pre-processing from 20 to 10 - but this did not limit the size/spread of the activations by much)
Best wishes,
Molly