Skip to content

multinetlab/HCP_TDA_Minne

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

47 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

HCP_TDA

This repository contains code for topological analysis of HCP data. This readme contains, step by step, the codes we used for our project. The codes are still under construction. We focused on the HCP aging database.

1 - Create file for cognitive/behavioral data

We combined all behavioral data files from the HCP project into a single behavioral data file:

  • Make_cog_file_HCP.ipynb

2 - Split data into exploration/training set and testing set.

We created two groups, training and validation sets. We performed feature selection in the training data set, and tested our findings in the validation dataset. We splitted the database based on age and on years of education. Data check based on visual inspection of distributions of these variables in both groups, as well as Mann Whitney U test.

Files:

  • Split_HCP_data_Females.ipynb
  • Split_HCP_data_Males.ipynb

3 - From the fMRI-timeseriesm we created 4 types of functional connectivity matrices:

  1. Correlation matrices - i.e., Pearson connectivity matrices: o Connectivity_matrices.py
  2. Filtered pvalued correlation matrices. Is created in three steps:
  • Calculate correlation matrices (is bit different than in Connectivity_matrices.py, so is a necessary step) Connectivity_matrices_pvalued.py
  • Calculate 100 randomized correlation matrices for comparisons: Connectivity_matrices_pvalued_random.py
  • Create filtered correlation matrices with only significant links: Connectivity_matrices_pvalued_filtered.py
  1. Mutual information matrices: Creates correlation matrices with mutual information scores Connectivity_matrices_MI.py
  2. Filtered mutual information matrices Creates mutual information matrices with only significant links To be created

Produce TDA features from correlation matrices:

Input is correlation matrices, as well as files for the region names and subnetworks. Output is csv file with all TDA features for all persons, and npz files for numpy arrays, which can then be plotted afterwards using the HCP_Data_checking_plotting.ipynb file. Files:

  • TDA_features_HCP.py

Create dataframes with TDA and cognitive data

Files:

  • Construct_dataframe.ipynb

Data exploration

Look for significant predictors of WM and EF using automatized codes from the Functions.ipynb file. Files:

  • Functions.ipynb
  • HCP_exploratory_Analysis_xxx.ipynb

About

Contains code for analyzing HCP data

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Jupyter Notebook 99.3%
  • Python 0.7%