Skip to content

Latest commit

 

History

History
90 lines (59 loc) · 4.73 KB

README.md

File metadata and controls

90 lines (59 loc) · 4.73 KB

pipelines

How to cite ?

[CITATION]

Contents Overview

This repository contains code that was used to analyze and create the figures of the manuscript cited above. The code is organised as follows:

.

├── data: .zip archives downloaded from the Human Connectome Project
├── results: storage of the analysis results
└── src: source code for the analysis

Reproducing full analysis

Raw data

The data used in the study are the raw functional and structural fMRI data from the 1080 HCP subjects who have completed the Motor Task ({subject}_3T_Structural_unproc.zip and {subject}_3T_tfMRI_MOTOR_unproc.zip). Archives were downloaded from https://db.humanconnectome.org/ into the data folder.

Dependencies and setup

Octave 5.1 and SPM12 for Octave were used.

The folder path for SPM12 was specified in the file .src/bash_scripts/spmpath.txt. Additionally, the paths to the folder containing the functions was stored in ./src/bash_scripts/srcpath.txt.

Preprocessing with SPM

Run ./src/bash_scripts/preprocessing_<PARAM_PREPROC>_list_IGRIDA.sh "<SUB1> <SUB2> <SUB3>" in a terminal where <SUB1>, <SUB2>, <SUB3>, etc. are the HCP subject identifiers as found in src/src/list_subjects_ordered_full.txt (this can be used to run preprocessing for part of the subjects to account for reduced storage space) and <PARAM_PREPROC> is taken from the table below.

PARAM_PREPROC Pipeline parameters
5 smoothing 5mm
8 smoothing 8mm

First-Level Analyses with SPM

  1. To perform first-level analysis and obtain the left foot contrast, run ./src/bash_scripts/fla_<PARAM_FLA>_list_IGRIDA.sh "<SUB1> <SUB2> <SUB3>" in a terminal where <SUB1>, <SUB2>, <SUB3>, etc. are the HCP subject identifiers as found in src/src/list_subjects_ordered_full.txt (this can be used to run preprocessing for part of the subjects to account for reduced storage space) and <PARAM_FLA> is taken from the table below.

  2. Similarly to 1., run ./src/bash_scripts/fla_<PARAM_FLA>_list_hand_IGRIDA.sh "<SUB1> <SUB2> <SUB3>" in a terminal to obtain the right hand contrast.

PARAM_FLA Pipeline parameters
5_0_0 smoothing 5mm, no motion regressors, canonical HRF
5_0_1 smoothing 5mm, no motion regressors, canonical HRF with temporal derivatives
5_6_0 smoothing 5mm, 6 motion regressors, canonical HRF
5_6_1 smoothing 5mm, 6 motion regressors, canonical HRF with temporal derivatives
5_24_0 smoothing 5mm, 24 motion regressors, canonical HRF
5_24_1 smoothing 5mm, 24 motion regressors, canonical HRF with temporal derivatives
8_0_0 smoothing 8mm, no motion regressors, canonical HRF
8_0_1 smoothing 8mm, no motion regressors, canonical HRF with temporal derivatives
8_6_0 smoothing 8mm, 6 motion regressors, canonical HRF
8_6_1 smoothing 8mm, 6 motion regressors, canonical HRF with temporal derivatives
8_24_0 smoothing 8mm, 24 motion regressors, canonical HRF
8_24_1 smoothing 8mm, 24 motion regressors, canonical HRF with temporal derivatives

Preprocessing and first-level analysis with FSL

Similarly to the previous SPM scripts, we use scripts ./src/bash_scripts/preprocessing_{parameter values}_list_fsl_IGRIDA.sh and ./src/bash_scripts/fla_{parameter values}_list_hand_fsl_IGRIDA.sh to perform preprocessing and first-level analysis on subject data with FSL.

Second-level analysis and false positive rates

  1. Run ./src/bash_scripts/second_level_analysis_full_hand_spm_IGRIDA.sh to perform the 1000 group analyses for every pair of pipelines.

  2. Run ./src/bash_scripts/false_positive_rate_full_hand_spm_IGRIDA.sh to obtain the empirical false positive rate for every pair of pipelines.

The resulting estimation of the false positive rate for each pair of pipeline is stored in the following file : results/smooth_<SMOOTH1>_reg_<REG1>_der_<DER1>/smooth_<SMOOTH2>_reg_<REG2>_der_<DER2>/mean_hand_50_FWE.mat, with <SMOOTH1>, <REG1>, <DER1> the parameter values for the first pipeline and <SMOOTH2>, <REG2>, <DER2> for the second pipeline, taken from the tables below.

SMOOTH1, SMOOTH2 Pipeline parameters
5 smoothing 5mm
8 smoothing 8mm
REG1, REG2 Pipeline parameters
0 no motion regressors
6 6 motion regressors
24 24 motion regressors
DER1, DER2 Pipeline parameters
0 canonical HRF
1 canonical HRF with temporal derivatives