Link Search Menu Expand Document

Post Processing

Analyses performed on CUBIC:

  • ModelArray (version; ran linear model on ASL data (n = 209; 5 participants from the full 214 with processed ASL were removed after QA checks - basically those with QEI below 0.6 and rmsd over 0.3 were manually inspected for quality, those who did not pass manual inspection were excluded. Analyses and related files at cbica/projects/GRMPY/project/curation/testing/stats_ModelArray).
    1. Followed documentation from Model Array demo and quick start guide
    2. Used Confixel for conversion from .nii to .h5; final convoxel .csv available at cbica/projects/GRMPY/project/curation/testing/stats_ModelArray/GRMPY_convoxel_meanCBF_n209_finalversion.csv
    3. Ran, which called which called the actual RScript.R. This used the singularity image cbica/projects/GRMPY/project/curation/testing/stats_ModelArray/modelarray.sif
    4. R script used the formula
       CBF ~ age (negligible non-linear effects) + sex + basilQEI + rmsd + als_score/bdi_score/scared_score_total/ari_score

      and function as below:

     mylm <- ModelArray.lm(formula, data = modelarray, phenotypes = phenotypes, scalar = "CBF",
                       element.subset = element.subset, full.outputs = full.outputs,
                       correct.p.value.terms = c("fdr", "bonferroni"),
                       correct.p.value.model = c("fdr", "bonferroni"),
                       pbar = TRUE, n_cores = num.cores) #n_cores = 4, element.subset = all voxels, full.outputs = TRUE
    1. Converted from .h5 to results folder named after date, containing all niftis, again using Confixel.
    2. No significance found for clinical scales: including fdr_correction images, cluster correction via easythresh and viewing via mricron.

    Results directories are pointed to in the file in the stats_ModelArray directory.

    Results further confirmed by vector testing in R:

     GM (GMmeanCBF from ASLPREP_QC.csv) ~ age + sex + basilQEI + rmsd + als_score/bdi_score/scared_score_total/ari_score
  • Did you use pennlinckit? Not yet.