Skip to content

chamberm/wuconnectomes

 
 

Repository files navigation

wuconnectomes

dipy based connectomes and statistical analyses to be expanded to include traits Wenlin Wu & Alex Badea #requires https://github.com/zhengwu/TensorNetwork_PCA.git.

  1. Streamline reconstruction is done using either of the two scripts: – CSA: CSA_revised_eg_sk_ww_binary_ALLanimal.py – CSD: CSD_det-reco.py takes about 25 mins/subject; radius =20

  2. Connectivity matrix generation

  • Save_connec.pyn or Save_connec.ipynb -output is saved : /Users/alex/code/Wenlin/preprocessing/results/
  1. combine connectivity csv to tensor /Users/alex/code/Wenlin/preprocessing/ConcentrateMatrix2Tensor.py /Users/alex/code/Wenlin/preprocessing/ConcentrateMatrix2Tensor.ipynb

  2. Tensor PCA: relies on /Users/alex/code/Matlab/popNet_HOPCA rom Dr Zhang /Users/alex/code/Matlab/popNet_HOPCA/ourscripts/TNPCAScripts.m

takes as input the tensor files from above go to lines 73 or 76 and uncomments to generate TNPCA for select genotypes or ages change nme of output files in line 105 saves 'U – subject mode','V – network mode','D-lambda for principal components','percent_store-top 3 reflects x pwercent of variation ','obj-obj value function','connectivity-as tensor'

TNPCA_Clustering_plot to save figs TNPCA identify subnet and copy results to xls sheet for LDA (fisher_training)

  1. register bundles and do stats on bundles using:

Bundle based registration-for loop- Gen3&4-2way-WithFA.ipynb

Bundle based registration-for loop- Old&Young-2way-WithFA.ipynb

About

dipy based connectomes

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Jupyter Notebook 88.2%
  • Python 5.3%
  • MATLAB 3.3%
  • R 3.2%