Application of Deep Knockoffs for fMRI to Generate Surrogate Data

This repository provides a pipeline, and the corresponding code to generate surrogate data using knockoffs for fMRI data.

Accompanying report: https://arxiv.org/abs/1811.06687

Framework of knockoffs: https://web.stanford.edu/group/candes/knockoffs/.

Software dependencies

The code contained in this repository was tested on the following configuration of Python:

• python==3.8.5
• numpy==1.19.2
• scipy==1.5.2
• torch==1.7.0
• cvxopt==1.2.5
• cvxpy==1.1.7
• pandas==1.1.3
• matplotlib==3.3.2
• seaborn==0.11.0
• statsmodels==0.12.1
• sklearn
• jupyter
• Cython

Installation Guide

pip3 install -r requirements.txt
pip3 install fanok
cd deepknockoffs/
pip3 install DeepKnockoffs/
pip3 install torch-two-sample-master/

Some files need to be downloaded here and put into the correct folders manually, since the github file limit forbids us from uploading them. X_tfMRI_WM_LR_Glasser360.mat is not required to run the project, the checkpoints are necessary to run the jupyter notebook. However, the jupyter notebook is already pre-run.

These files are:

• X_tfMRI_WM_LR_Glasser360.mat: data/input/
• DeepKO_tWM_s0_checkpoint.pth.tar_prev.pth.tar: data/output/knockoffs/
• DeepKO_tWM_s0_checkpoint.pth.tar: data/output/knockoffs/
• DeepKO_tMOTOR_s1_checkpoint.pth.tar_prev.pth.tar: data/output/knockoffs/
• DeepKO_tMOTOR_s1_checkpoint.pth.tar: data/output/knockoffs/

Example pipeline

• pipeline.ipynb a usage example on how to use our approach to generate knockoffs from fMRI data and perform non-parametric tests for thresholding.

Brain Plots

In order to obtain the brain plots, a Matlab script is used: our_plots.m This file is located in PlotGraph, which also contains the necessary functions.

File Structure

Here is the file structure of the project:

Project
|
|-- data
|   |-- input
|   |-- output
|       |-- beta
|       |-- img
|       |-- knockoffs
|
|-- deepknockoffs/
|    |-- torch-two-sample-master/
|
|-- implementation
|    |-- __init__.py
|    |-- create_activations.py
|    |-- glm.py
|    |-- knockoff_class.py
|    |-- knockoff_classes_test.py
|    |-- load.py
|    |-- non_parametric.py
|    |-- params.py
|    |-- utils.py 
|
|-- PlotGraph/
|
|-- tests
|    |-- __init__.py
|    |-- test_glm.py
|    |-- test_load.py
|
|-- .gitignore
|-- .gitlab-ci.yml.deactivate
|-- __init__.py
|-- pipeline.ipynb
|-- README.md

• Authors: Alec Flowers, Alexander Glavackij and Janet van der Graaf
• Supervision: Giulia Preti and Younes Farouj