Clarification on GROUP_LABEL in PET-Volume Pipeline

[Shahzad ALi]

Dear All,

I hope you're doing well. 

I have a quick question regarding the GROUP_LABEL parameter in the pet-volume pipeline.

• GROUP_LABEL is the label of the group that is associated to the DARTEL template that you had created when running the t1-volume pipeline.

When I ran the t1-volume pipeline, I specified three groups (AD, MCI, CN) based on diagnosis. Now, for pet-volume, the subjects in each group are not identical in number (though the labels are the same).

Since pet-volume references the same DARTEL template from t1-volume, does the mismatch in group sizes affect the analysis? Or is it sufficient to use the same GROUP_LABEL as before, even if some subjects are excluded in the PET data?

Thank you for your help!

Best regards,

-----------------------------------------------------------------------

[Alice Joubert]

Hello Shahzad, 

Do you know how many subjects are missing PET for each group ?

Best,

--
You received this message because you are subscribed to the Google Groups "Clinica" group.
To unsubscribe from this group and stop receiving emails from it, send an email to clinica-user...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/clinica-user/CAPaAoQSt8TOJtv5WrGaDpuot5EXUvQS99Hr2mcQw0Av-z3P-Zg%40mail.gmail.com.

[Shahzad ALi]

Hi, Dear Alice, 

Thank you for your response.

Regarding your question: in the PET dataset (subjects with both T1 and PET), the number of subjects per group (AD, MCI, CN) is roughly 60-70% of what was processed in the t1-volume pipeline.

Furthermore, I’d like to clarify a few things before proceeding with the pet-volume pipeline:

1. Does the definition of GROUP_LABEL impact the validity of inter-group comparisons (e.g., AD vs. CN or MCI vs. AD, etc.)? 

Since I created separate templates per group (AD, MCI, CN) using t1-volume, I’m concerned that using different DARTEL templates for each group of subjects may affect the reliability of comparisons across them.

• clinica run t1-volume-create-dartel BIDS_Dir CAPS_DIR CN -tsv CN_Group.tsv 
• clinica run t1-volume-create-dartel BIDS_Dir CAPS_DIR AD -tsv AD_Group.tsv 
• clinica run t1-volume-create-dartel BIDS_Dir CAPS_DIR MCI -tsv MCI_Group.tsv 

2. Is it acceptable, or even preferred, to create a single DARTEL template from healthy controls (CN) and then use that template for all subjects in the PET pipeline (including AD and MCI)?

• clinica run t1-volume-create-dartel BIDS_Dir CAPS_DIR CN -tsv CN_Group.tsv 
• clinica run t1-volume-register-dartel BIDS_Dir CAPS_DIR CN -tsv MCI_Group.tsv 
• clinica run t1-volume-register-dartel BIDS_Dir CAPS_DIR CN -tsv AD_Group.tsv 

I’d really appreciate your expert opinion before running or adjusting the pipeline. Looking forward to your guidance.

Thanks & Best regards,
Shahzad

-----------------------------------------------------------------------

To view this discussion visit https://groups.google.com/d/msgid/clinica-user/CAP_epFt1RF0Ca2fVAaUhfiJJsBh%2BPuOAKrENTQb%3DK6a6ca6ceQ%40mail.gmail.com.

[ninon....@gmail.com]

Dear Shahzad,

When studying the Alzheimer's disease spectrum, it is usual to create a DARTEL template using the baseline session of all the subjects (CN, MCI & AD). This avoids introducing biases related to the use a specific template per diagnostic group, and using the baseline session avoids given more weight to a subject that would have many sessions. Another valid strategy, if you have enough CN subjects, is to create the DARTEL template using only the CN subjects and reusing it for the other diagnostic groups.

Best,

Ninon