Hi Anderson!
Thank you for the suggestion on those parameters. Things are running much faster. There are a lot of output files. I just wanted double check which file is the correct one to check my results and if my methodology is even close?
Would I be looking at my prefix_vox_tstat_mfwep_m*_c*.nii.gz files? I assume that I would be using the fslstats command to find out if there are any significant voxels with values 1.301. Would the best way to do this be fslstats –t prefixvox_tstat_mfwep_m**_c4.nii.gz –R after merging the ICs to get a list of 36 values(I have 36 ICs)?
I feel like I am a bit off but I don’t know if its due to incorrect file type of just different parameters for my fslstats command.
Thank you so much for your time. I really appreciate it.
Please advise,
Noelle Dalin
From: "[log in to unmask]<mailto:[log in to unmask]>" <[log in to unmask]<mailto:[log in to unmask]>> on behalf of "Anderson M. Winkler" <[log in to unmask]<mailto:[log in to unmask]>>
Reply-To: "[log in to unmask]<mailto:[log in to unmask]>" <[log in to unmask]<mailto:[log in to unmask]>>
Date: Thursday, November 10, 2016 at 12:55 AM
To: "[log in to unmask]<mailto:[log in to unmask]>" <[log in to unmask]<mailto:[log in to unmask]>>
Subject: Re: [FSL] Statistical Analysis for Resting State FMRI- Clustering?
Hi Noelle,
The command 2 looks right. To accelerate things, consider adding the options "-n 500 -accel tail -nouncorrected".
Also, I recommend strongly that you use the option "-logp", that will show p-values in log-scale. It will be easier later to check the results (the significant voxels will have values above -log10(0.05) = 1.301).
All the best,
Anderson
On 9 November 2016 at 23:00, Noelle Dalin <[log in to unmask]<mailto:[log in to unmask]>> wrote:
Hi Anderson,
I am testing out commands for palm. We have 36 components and 4 contrasts for two groups. As a new user, I really have no idea what I am doing so I wanted to see if the commands looked correct to you? I am trying two different commands. Also do you know how long PALM should run for? Is it correct to use the stage2 files or should I be using stage3? Any other suggested parameters?
I am just trying to correct for multiple testing problem 2.
Please advise,
Noelle
COMMAND 1:
palm -i dr_stage2_ic0000.nii.gz -i dr_stage2_ic0001.nii.gz -i dr_stage2_ic0002.nii.gz -i dr_stage2_ic0003.nii.gz -i dr_stage2_ic0004.nii.gz -i dr_stage2_ic0005.nii.gz -i dr_stage2_ic0006.nii.gz -i dr_stage2_ic0007.nii.gz -i dr_stage2_ic0008.nii.gz -i dr_stage2_ic0009.nii.gz -i dr_stage2_ic0010.nii.gz -i dr_stage2_ic0011.nii.gz -i dr_stage2_ic0012.nii.gz -i dr_stage2_ic0013.nii.gz -i dr_stage2_ic0014.nii.gz -i dr_stage2_ic0015.nii.gz -i dr_stage2_ic0016.nii.gz -i dr_stage2_ic0017.nii.gz -i dr_stage2_ic0018.nii.gz -i dr_stage2_ic0019.nii.gz -i dr_stage2_ic0020.nii.gz -i dr_stage2_ic0021.nii.gz -i dr_stage2_ic0022.nii.gz -i dr_stage2_ic0023.nii.gz -i dr_stage2_ic0024.nii.gz -i dr_stage2_ic0025.nii.gz -i dr_stage2_ic0026.nii.gz -i dr_stage2_ic0027.nii.gz -i dr_stage2_ic0028.nii.gz -i dr_stage2_ic0029.nii.gz -i dr_stage2_ic0030.nii.gz -i dr_stage2_ic0031.nii.gz -i dr_stage2_ic0032.nii.gz -i dr_stage2_ic0033.nii.gz -i dr_stage2_ic0034.nii.gz -i dr_stage2_ic0035.nii.gz -corrmod -noniiclass -o try6_liststg2
COMMAND 2: with different parameters in red
palm -i dr_stage2_ic0000.nii.gz -i dr_stage2_ic0001.nii.gz -i dr_stage2_ic0002.nii.gz -i dr_stage2_ic0003.nii.gz -i dr_stage2_ic0004.nii.gz -i dr_stage2_ic0005.nii.gz -i dr_stage2_ic0006.nii.gz -i dr_stage2_ic0007.nii.gz -i dr_stage2_ic0008.nii.gz -i dr_stage2_ic0009.nii.gz -i dr_stage2_ic0010.nii.gz -i dr_stage2_ic0011.nii.gz -i dr_stage2_ic0012.nii.gz -i dr_stage2_ic0013.nii.gz -i dr_stage2_ic0014.nii.gz -i dr_stage2_ic0015.nii.gz -i dr_stage2_ic0016.nii.gz -i dr_stage2_ic0017.nii.gz -i dr_stage2_ic0018.nii.gz -i dr_stage2_ic0019.nii.gz -i dr_stage2_ic0020.nii.gz -i dr_stage2_ic0021.nii.gz -i dr_stage2_ic0022.nii.gz -i dr_stage2_ic0023.nii.gz -i dr_stage2_ic0024.nii.gz -i dr_stage2_ic0025.nii.gz -i dr_stage2_ic0026.nii.gz -i dr_stage2_ic0027.nii.gz -i dr_stage2_ic0028.nii.gz -i dr_stage2_ic0029.nii.gz -i dr_stage2_ic0030.nii.gz -i dr_stage2_ic0031.nii.gz -i dr_stage2_ic0032.nii.gz -i dr_stage2_ic0033.nii.gz -i dr_stage2_ic0034.nii.gz -i dr_stage2_ic0035.nii.gz -d ../GLM_22.mat -t ../GLM_22.con -n 2000 -corrcon – corrmod -noniiclass -o try7_liststg2glm22_corrcon_2000
MY GLM Looks like so:
Higher-level/non-timeseries design
# inputs: 22
General Linear Model:
EVs
Number of main EVs: 2
Number of additional, voxel depended Evs: 0
Group EV1 EV2
TD DCD
1 1 0
1 1 0
1 1 0
1 1 0
1 1 0
1 1 0
1 1 0
1 1 0
1 1 0
1 1 0
1 1 0
2 0 1
2 0 1
2 0 1
2 0 1
2 0 1
2 0 1
2 0 1
2 0 1
2 0 1
2 0 1
2 0 1
Contrasts & F-tests
Contrasts:4 F-tests: 0
TITLE EV1 EV2
C1 TD>DCD 1 -1
C2 DCD>TD -1 1
C3 TD MEAN 1 0
C4 DCD MEAN 0 1
From: "[log in to unmask]<mailto:[log in to unmask]>" <[log in to unmask]<mailto:[log in to unmask]>> on behalf of Noelle Dalin <[log in to unmask]<mailto:[log in to unmask]>>
Reply-To: "[log in to unmask]<mailto:[log in to unmask]>" <[log in to unmask]<mailto:[log in to unmask]>>
Date: Tuesday, November 8, 2016 at 9:50 AM
To: "[log in to unmask]<mailto:[log in to unmask]>" <[log in to unmask]<mailto:[log in to unmask]>>
Subject: Re: [FSL] Statistical Analysis for Resting State FMRI- Clustering?
Hi Anderson,
Thank you so much for the response! I did mean correction for the number of ICs as I tend to get a large amount ~36. Our data can be quite noisy though even after single session denoising and exclusion of participants based on motion threshold.
I appreciate the direction, I am now looking into the PALM tool.
Cheers,
Noelle Dalin
From: "[log in to unmask]<mailto:[log in to unmask]>" <[log in to unmask]<mailto:[log in to unmask]>> on behalf of "Anderson M. Winkler" <[log in to unmask]<mailto:[log in to unmask]>>
Reply-To: "[log in to unmask]<mailto:[log in to unmask]>" <[log in to unmask]<mailto:[log in to unmask]>>
Date: Tuesday, November 8, 2016 at 2:59 AM
To: "[log in to unmask]<mailto:[log in to unmask]>" <[log in to unmask]<mailto:[log in to unmask]>>
Subject: Re: [FSL] Statistical Analysis for Resting State FMRI- Clustering?
Hi Noelle,
Do you mean correction for the number of voxels (multiple testing problem "I"), or correction for the number of independent components (multiple testing problem "II")? If for the number of voxels, randomise already corrects that, and there is nothing else that you need to do. If for the number of ICs, then either use Bonferroni (yes, can be conservative), or use correction across modalities in PALM.
All the best,
Anderson
On 8 November 2016 at 00:43, Noelle Dalin <[log in to unmask]<mailto:[log in to unmask]>> wrote:
Hi all,
I am looking into doing a multiple comparisons correction after running dual regression with 5000 permutations on pediatric functional connectivity data with a low power. Bonferroni correction is not a great choice for me because it is quite conservative and I only have 22 participants total (11 in each group).
I am looking into cluster-extent based thresholding however, from what I am reading on your support page, it is redundant to try to cluster when using tfce files.
Does anyone have any tips or tricks for me? Is there another option that I should be looking into or should I continue to pursue clustering?
Thank you for your time,
Noelle Dalin
