NITRC NiiStat Forum: help

RE: atlas-based VLSM using custom ROI

sen0n — Thu, 18 Apr 2024 18:10:36 GMT

Hi Roger,
I was assuming that the AAL atlas provided with NiiStat is in standard format, an in MNI space.
I used the AAL precentral gyrus as custom ROI. 
Thank you

RE: atlas-based VLSM using custom ROI

Roger Newman-Norlund — Tue, 16 Apr 2024 12:01:11 GMT

Hi senOn,
Is your atlas in standard format, and in MNI space?
Can you share your custom atlas?
Roger
<blockquote>
Hi,
I performed an atlas-based lesion-symptom mapping analysis by considering a custom ROI with NiiStat that revealed a significant region.
When overlaying the thresholded map on an MNI template and the atlas however, only a fraction of the map cover the custom ROI, but also additional other ROIs. What is the interpretation of this finding?
Furthermore, the results were repeated for a subsample, revealing exactly the same thresholded map just with a slightly lower Z-score. 
Does this custom ROI approach for one ROI test only for this ROI?
If so, why does the threholded map include other ROIs?
Could this be due to an offset during reslicing?
Any support is very much appreciated.  
</blockquote>

atlas-based VLSM using custom ROI

sen0n — Tue, 16 Apr 2024 1:49:33 GMT

Hi,
I performed an atlas-based lesion-symptom mapping analysis by considering a custom ROI with NiiStat that revealed a significant region.
When overlaying the thresholded map on an MNI template and the atlas however, only a fraction of the map cover the custom ROI, but also additional other ROIs. What is the interpretation of this finding?
Furthermore, the results were repeated for a subsample, revealing exactly the same thresholded map just with a slightly lower Z-score. 
Does this custom ROI approach for one ROI test only for this ROI?
If so, why does the threholded map include other ROIs?
Could this be due to an offset during reslicing?
Any support is very much appreciated.

nii_merge

danilo mattia — Mon, 11 Dec 2023 15:01:10 GMT

Hi!
<pre id="tw-target-text" class="tw-data-text tw-text-large tw-ta" dir="ltr" data-placeholder="Traduzione" data-ved="2ahUKEwib5OHxzoeDAxUmnf0HHYmmAU0Q3ewLegQIBhAQ">I'm Danilo Mattia, technician of the BCI laboratory at the Santa Lucia foundation, Research and Rehabilitation Hospital, Italy.</pre>
<pre class="tw-data-text tw-text-large tw-ta" dir="ltr" data-placeholder="Traduzione" data-ved="2ahUKEwib5OHxzoeDAxUmnf0HHYmmAU0Q3ewLegQIBhAQ">I turned my T1 T2Flair and lesion nii files into mat files with the command 'nii_nii2mat'for each of my subjects. I would now like to create a single .mat file with the 'nii_merge' command. However, when I start this command I am not asked for the .mat files, as it is written in the 'introduction to NiiStat', but for the .nii files. When I run the command with the .nii files I get this error; </pre>
Output argument "v" (and maybe others) not assigned during call to "nii_merge>vReadSub".
Error in nii_merge (line 24)     val = vReadSub(nameVal, size(img,4), 1);
 
<pre id="tw-target-text" class="tw-data-text tw-text-large tw-ta" dir="ltr" data-placeholder="Traduzione" data-ved="2ahUKEwib5OHxzoeDAxUmnf0HHYmmAU0Q3ewLegQIBhAQ">I would like to know what my mistake is and how I can fix it. thanks in advance for your help! danilo</pre>
<pre id="tw-target-text" class="tw-data-text tw-text-large tw-ta" dir="ltr" data-placeholder="Traduzione" data-ved="2ahUKEwib5OHxzoeDAxUmnf0HHYmmAU0Q3ewLegQIBhAQ"> </pre>

interpreting results output

sen0n — Fri, 03 Nov 2023 4:15:31 GMT

Hello, I could successfully use Niistat for performing a VLSM on a set of binary lesion masks and continuous behavioral data with higher scores indicating less deficit. I got the result that 1 region on the AAL atlas survived the threshold of z=-1.6 when chosing 3000 premutations and lesion volume correction. If have two questions related to the output.
1. I was expecting a positive z-value based on the instructions. How can the negative threshold be interpreted?
2. When opening the "threshZresults.nii-file" and overlaying it on the AAL atlas, additional regions and overlaying voxel to the "1 region" in the output are listed, even though to a lesser contribution. Is the threshZresults-file showing only the significant cluster or is there another way to report on the lesion locations /voxels of the associated cluster based on the niistat output?

Demo dataset

Ge Dang — Tue, 25 Jul 2023 12:57:33 GMT

The link for the Demo dataset of niistat does not work. Could you please share the demo dataset if you have? Thank you!

de-skew data permutation testing

Selma Lugtmeijer — Thu, 23 Mar 2023 2:57:54 GMT

Hi,

How critical is it to de-skew the behavioural data when you use permutation testing to correct for multiple testing?

Thanks,
Selma

interpretation of power maps

A S — Tue, 15 Nov 2022 16:32:45 GMT

Hi,

I'm looking at the power.nii and powermap.nii that comes with the results, but can't figure out what the values represent.
Any help is much appreciated!

Many thanks, Anouk

Running VLSM with NiiStat: Unexpected image dimensions

Hanne Huygelier — Tue, 15 Nov 2022 10:39:10 GMT

Dear NiiStat forum,

It's my first time performing a VLSM.
I am working on a dataset of binary lesion maps delineated on different clinical scan modalities (i.e., CT, T2, FLAIR).
I first ran the nii_nii2mat conversion to create a folder with all binary lesion maps in .mat format.
I then tried to run a basic VLSM (single modality: lesion map) with Bonferroni correction (just to practice) and considering lesion volume.

However, I got the following error message:

"Unexpected image dimensions vary [with after that a reference to a FLAIR lesion map] and then a number: 18685."

What could be causing this error? // Update: I found cause of issue.
Images still needed to be resliced to match them on dimensions.

With kind regards,
Hanne

One-tailed statistics

wittayer — Tue, 23 Aug 2022 7:10:36 GMT

Dear all,

as I understood NiiStat assumes a score to decrease with lesion size. If I want to assess a score that increases with lesion size (i.e. NIHSS,EDSS) with VLSM, at your opinion is it valid to invert it (so to take the maximum possible score and subtract the true value from that, i.e. NIHSS=11, NIHSSinv=42-11= 31)? Is it even necessary or do I misunderstand the assumption of an inversly proportinal correlation?
Thanks,
Best Regards
Matthias