andersonwinkler / PALM

PALM: Permutation Analysis of Linear Models
70 stars 28 forks source link

exchange block file #50

Open arumajiro1976 opened 1 year ago

arumajiro1976 commented 1 year ago

I'm attempting permutations within-block and whole-block design using an eb file (file was csv file.). I created the eb file with:

The first column containing all values as 1 (all positive values) The second column with IDs (all positive values) The third column with values of either 1 or 2 (all positive values) However, when I look at the palm_permidx.csv, it appears that only the elements in the third column have been shuffled. (I want to shuffle the data based on second and third columns data, in within-block and whole-block design manner)

Am I doing something wrong?

andersonwinkler commented 1 year ago

It should most certainly have worked if you followed the examples in the paper. In particular, Figure 1 (c): https://doi.org/10.1016/j.neuroimage.2015.05.092

arumajiro1976 commented 1 year ago

Thank you for your quick reply.

As you suggested, we read the paper and made the eb file based on Figure1(c), however the palm_permidx.csv showed Figure1(a) manner.

this is my code

palm -i only_fc_data.csv -d design_matrix.csv -t contrast.csv -eb eb.csv -fdr -quiet -savemax -savemetrics -saveparametric -saveperms -savedof

Is there any other Tips to make eb file?

andersonwinkler commented 1 year ago

It could be that the permutations that whould shuffle the 2nd column are not informative for your hypothesis, as defined by design and contrasts. PALM will skip these permutations. If you want to run these nonetheless, try adding the option -cmcx.

arumajiro1976 commented 1 year ago

Thank you for your reply. I will try it. best

arumajiro1976 commented 1 year ago

I have an additional question. I'm looking to test for interaction effects in repeated measures data (Time 1, Time 2) for two groups (Group A and Group B). In this case, as I mentioned earlier, I am considering applying permutation using within-block (Time 1, Time 2) and whole-block (ID). Is this approach to permutation correct?

Wyuuyuan commented 10 months ago

Hello, I've recently been designing a blocks.csv file similar to yours, but I've encountered an error during the execution process. I've already checked for file permission issues and path errors, but have not resolved it. My first column is the number 1, the second column contains the numbers 1 and 2, and the subsequent columns go from 1 to 40. However, my file only contains these three columns of numbers. Does your blocks file also contain only numbers, or are there commas or other textual content? Thank you. image