Regarding task fMRI group analysis

[Elsie Lin]

Dear HCP experts,

I'm new to the analysis pipelines and I have some questions regarding task fMRI analysis, particularly in relation to the EMOTION task.

 
I am aiming to conduct a group analysis, and my plan involves first parcellating individual contrast maps and then use PALM for group results. I would greatly appreciate if you can help me on the following questions:

(1) Could you please confirm if this planned analysis workflow is appropriate? Additionally, I am curious about the impact of parcellation on PALM analysis. Would it make a significant difference if I run PALM without parcellation?

(2) HCP provides task fMRI analysis files for the S1200 dataset. Can I use these task analysis files to do parcellation, or do I have to rerun the individual task analysis with parcellation myself? Any guidance or resources you can provide would be very helpful.

(3) What would be the best approach to examine the activity in both cortical and subcortical regions with parcellated analysis? I would be grateful for any recommendations or materials on this topic.

Thank you for your time and assistance. I appreciate your patience if my question seems too basic. I look forward to your response.

Best regards,

Yi-Hsuan Lin, Ph.D.

Research Center for Mind, Brain, and Learning

National Chengchi University, Taiwan 

[Glasser, Matt]

1. Seems like a good idea.  Parcellation will improve statistical sensitivity and power.
2. You would rerun with the task analysis pipeline, but it would be very fast to do.
3. You can review the results of PALM with Connectome Workbench.

 

Matt.

--
You received this message because you are subscribed to the Google Groups "HCP-Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hcp-users+...@humanconnectome.org.
To view this discussion on the web visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/6ae34ee6-e70a-4f8b-8e87-5391fe131ae7n%40humanconnectome.org.

 

---

The materials in this message are private and may contain Protected Healthcare Information or other information of a sensitive nature. If you are not the intended recipient, be advised that any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this email in error, please immediately notify the sender via telephone or return mail.

[Elsie Lin]

Hi Matt, 

Thank you very much for your assistance last time. However, I encounter another issue. 

I want to use an ROI mask to target specific brain regions. But when I use the -m flag to provide the mask to PALM, it shows an error message: 

The size of the data does not match the size of the mask:

- Data file Y.dtseries.nii
- Mask file mask.nii

Could you provide possible solutions for this issue? Any suggestions would be appreciated.

Thank you!

Best,

Yi-Hsuan Lin, Ph.D.

Research Center for Mind, Brain, and Learning

National Chengchi University, Taiwan

glas...@wustl.edu 在 2024年5月6日 星期一晚上7:03:39 [UTC+8] 的信中寫道：

[Glasser, Matthew]

The mask needs to be in the same space as the data.  Not CIFTI for data and NIFTI for mask.

[Tim Coalson]

Try wb_command -file-information -no-map-info on the mask.nii and Y.dtseries.nii files - as Matt says, only one of those filenames looks like a cifti file.

Tim

To view this discussion on the web visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/3BC584FF-93AA-4DC5-9212-07F3C3A4FEBE%40wustl.edu.

[Elsie Lin]

Hello Matt and Tim, 

Your suggestions have been very helpful! However, I still encounter some issues when I tried to create the ROI mask in cifti format. 

My goals is to focus on specific cortical areas (e.g. ACC and Insula) and subregions in subcortical areas (e.g. BLA and CA1). 

My plan to make the ROI mask is to first selecting the cortical areas from the MMP 1.0 atlas, and then replace the subcortical volume with ROI files selected from other .nii based atlas.  

I have a fwe questions: 

1. How to select specific areas from the MMP 1.0 atlas to make a new mask?

2. I've tried to create the subcortical volume label file with the following step:

• Resample subcortical ROIs to MNI 2mm space using wb_command -volume-resample with the MNI152_T1_2mm.nii.gz from the HCP template file
• Merge subcortical ROIs using wb_command -volume-merge
• Reduce with wb_command -volume-reduce using INDEXMAX

However, reduction with INDEXMAX didn't result in correct integer keys for my ROIs. 

For the second and third ROIs, the keys are correct, but the rest of the voxels are all denoted as "1"(as shown in the attached screenshot).

Can you help me with these issues?

Many thanks!

 

Best, 

Yi-Hsuan

tim.c...@gmail.com 在 2024年6月14日 星期五凌晨3:49:15 [UTC+8] 的信中寫道：

[Tim Coalson]

If all the maps are equal (for instance, 0), INDEXMAX returns 1.  Do a second -volume-reduce with MAX, and use the result to mask your INDEXMAX file appropriately (probably threshold at 0.5).

Tim

[Tim Coalson]

As for selecting parts of the MMP, one possibility is -cifti-label-to-roi.  Or, you can -cifti-label-export-table, delete all the labels you don't care about, and do -cifti-label-import with the -discard-others option.

However, if you plan to do parcellated analysis against all other areas, you might instead just generate the complete parcellated matrix, and only look at the rows for the areas you are interested in, rather than making an asymmetric parcellated matrix file.

Tim