-
Notifications
You must be signed in to change notification settings - Fork 5
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Merge pull request #2 from AyaKabbara/main
Pull request
- Loading branch information
Showing
31 changed files
with
1,200 additions
and
1,849 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -1,43 +1,124 @@ | ||
| # StageEEGpre | ||
| # Successful reproduction of a large EEG study across software packages | ||
|
|
||
| dependencies: | ||
| - eeglab : https://sccn.ucsd.edu/eeglab/download.php eeglab current v2021.1 version avec matlab | ||
| - article : | ||
| on trouve les 4 dependances sur https://github.com/Neuro-Tools/ | ||
| il faut telecharger : https://github.com/Neuro-Tools/MATLAB-EEG-preProcessing | ||
| https://github.com/Neuro-Tools/MATLAB-EEG-fileIO | ||
| https://github.com/Neuro-Tools/MATLAB-EEG-timeFrequencyAnalysis | ||
| https://github.com/Neuro-Tools/MATLAB-EEG-icaTools | ||
| by | ||
| Aya Kabbara, | ||
| Nina Forde, | ||
| Camille Maumet, | ||
| Mahmoud Hassan. | ||
| Contact: [email protected] | ||
|
|
||
| data: | ||
| https://osf.io/qrs4h/ | ||
| telecharger Raw data Part 1.zip et l'extraire dans le dossier ./data | ||
| telecharger https://osf.io/ztw8u/ ChanlocsMaster.mat et l'ajouter dans le dossier Raw data part 1 ci dessus | ||
| > This is a guide for researchers to reproduce the results and figures published in the paper "Successful reproduction of a large EEG study across software packages". | ||
| >This paper has been submitted for publication in NeuroImage Reports. | ||
| ## Reproducing Fig 2A of (Williams et al, 2021) | ||
| This study sheds light on how the software tool used to preprocess EEG signals impacts the analysis results and conclusions. EEGLAB, Brainstorm and FieldTrip were used to reproduce the same preprocessing pipeline of a published EEG study performed on 500 participants. | ||
|
|
||
| ### Using the code provided with the paper and the preprocessed datasets | ||
|  | ||
|
|
||
| 1. Run the analysis and create the intermediatary file that processes the subject results to estimate the means etc [RewardProcessing_Plots.m](https://github.com/Inria-Empenn/StageEEGpre/blob/main/src/eeglabcode/plot/RewardProcessing_Plots.m). | ||
| 2. Create the figures using the R script [RewardProcessing_Plots_and_Statistics.R](https://github.com/Inria-Empenn/StageEEGpre/blob/main/src/graphiques/RewardProcessing_Plots_and_Statistics.R). | ||
| *The full pipeline of the study. (a) EEGs were acquired from 500 participants performing a simple gambling task of six blocks composed of 20 trials. (b) The same dataset was then preprocessed using the different software tools: Reference (using the code published by the original paper (Williams et al. 2021)), EEGLAB, Brainstorm and FieldTrip. The preprocessing steps to be performed in each tool includes: the reduction to 32 electrodes, the re-referencing, the automatic detection of bad electrodes, the band-pass filtering, the interpolation of bad channels, the segmentation into time-locked epochs and the removal of artifactual trials. (c) The preprocessed signals derived from the four preprocessing codes are used to validate and reproduce the reference statistics and hypotheses. A quantitative comparison between the resulting signals is also conducted in terms of signal features* | ||
|
|
||
| We obtain [this figure](https://github.com/Inria-Empenn/StageEEGpre/blob/main/figures/articke%20fig2/preprocessed100.png) | ||
|
|
||
| ### Using the code provided with the paper and the raw datasets (without ICA manual) | ||
| ## Abstract | ||
|
|
||
| 1. Add path to dependencies with [demarragemaison.m](https://github.com/Inria-Empenn/StageEEGpre/blob/main/src/header%20script/demarragemaison.m). Note: this file has to be edited to include your own pathes. | ||
| 2. Create the figures using the R script [RewardProcessing_Plots_and_Statistics.R](https://github.com/Inria-Empenn/StageEEGpre/blob/main/src/graphiques/RewardProcessing_Plots_and_Statistics.R). Note: the number of participants has to be updated to the number of participants included. | ||
| As an active field of research, Electroencephalography (EEG) analysis workflow has increasing flexibility and complexity, with a great variety of methodological options and tools to be selected at each step. This high analytical flexibility can be problematic as it can yield variability in research outcomes. Therefore, growing attention has been recently paid to understand the potential of different methodological decisions to influence the reproducibility of results. In this paper, we aim to examine how sensitive the results of EEG analyses are to variations in software tools. We reanalyzed shared EEG data (N=500) previously used in a published task EEG study using three of the most commonly used software tools: EEGLAB, Brainstorm and FieldTrip. After reproducing the same original preprocessing workflow in each software, the resultant evoked-related potentials (ERPs) were qualitatively and quantitatively compared in order to examine the degree of consistency/discrepancy between softwares. Our findings show a good degree of convergence in terms of the general profile of ERP waveforms, peak latencies and effect size estimates related to specific signal features. However, considerable variability was also observed between software packages as reflected by the similarity values and observed statistical differences at particular channels and time instants. In conclusion, we believe that this study provides valuable clues to better understand the impact of the software tool on analysis results of EEG. | ||
|
|
||
| We obtain [this figure](https://github.com/Inria-Empenn/StageEEGpre/blob/main/figures/articke%20fig2/noica73.png) | ||
| ## Software implementation | ||
|
|
||
| ### Using EEGlab and the raw datasets | ||
| > In each software, the preprocessing script includes the reduction to 32 electrodes, the re-referencing, the automatic detection of bad electrodes, the band-pass filtering, the interpolation of bad channels, the segmentation into time-locked epochs and the removal of artifactual trials. | ||
| 1. Run the analysis [eeglab_preprocessing.m](https://github.com/AyaKabbara/StageEEGpre/blob/main/src/eeglab/eeglab_preprocessing.m). | ||
| 2. Create the figures using the R script [fig2Av2.R](https://github.com/Inria-Empenn/StageEEGpre/blob/main/src/codeR/fig2Av2.R). Note: the number of participants has to be updated to the number of participants included, the name of csv files should be changed accordingly. | ||
|
|
||
| We obtain [this figure](https://github.com/Inria-Empenn/StageEEGpre/blob/main/figures/articke%20fig2/100sujetseeglabfinal.png) | ||
| All the preprocessing codes were written in Matlab (Matlab R2019a). The visualizations of ERP waveforms (Figure 2) and the quantitative features (Figure 3) were done in R (R Core Team 2020). Seaborn (Waskom 2021) was used to illustrate the comparison between the software distribution of the quantitative measures (Figure 4), and the similarity matrix between softwares (Figure 5). Other visualizations and statistical assessments were conducted using Matlab. | ||
|
|
||
| All source code used to generate the results and figures in the paper are in | ||
| the `src` folder. | ||
| The figure generation of figure 5 is run inside | ||
| [Jupyter notebooks](http://jupyter.org/). | ||
| The data used in this study is stored in `data` and downloaded from https://osf.io/qrs4h/. | ||
| Results generated by the code are saved in `results`. | ||
| See the `README.md` files in each directory for a full description. | ||
|
|
||
| ## Getting the data | ||
|
|
||
| Download data from https://osf.io/qrs4h/ and store it `data` directory. | ||
| The channel location file should be also downloaded from https://osf.io/ztw8u/ ChanlocsMaster.mat, and stored in `data` folder. | ||
|
|
||
| ## Getting the code | ||
|
|
||
| You can download a copy of all the files in this repository by cloning the | ||
| [git](https://git-scm.com/) repository: | ||
|
|
||
| git clone https://github.com/AyaKabbara/EEG_preprocessing.git | ||
|
|
||
| or [download a zip archive](https://github.com/AyaKabbara/StageEEGpre/archive/master.zip). | ||
|
|
||
| ## Dependencies | ||
| - Reference script: To re-run the script of the reference paper, one should first download the Neuro-Tools package from https://github.com/Neuro-Tools/ including https://github.com/Neuro-Tools/MATLAB-EEG-preProcessing, | ||
| https://github.com/Neuro-Tools/MATLAB-EEG-fileIO , | ||
| https://github.com/Neuro-Tools/MATLAB-EEG-timeFrequencyAnalysis , | ||
| https://github.com/Neuro-Tools/MATLAB-EEG-icaTools. | ||
|
|
||
| - EEGLAB script : This script uses EEGLAB functions available at https://sccn.ucsd.edu/eeglab/download.php eeglab current version | ||
|
|
||
| - Brainstorm script : This script uses Brainstorm functions available at https://neuroimage.usc.edu/bst/download.php | ||
|
|
||
| - FieldTrip script: This script uses FieldTrip functions available at https://www.fieldtriptoolbox.org/download/ | ||
|
|
||
| ## Reproducing the results | ||
|
|
||
| ### Reproducing Figure 2A, Figure 3A using the reference script of (Williams et al, 2021) | ||
|
|
||
| > We used the same preprocessing code provided by the original paper, by eliminating the step of manual ICA employed to detect the arttifactual components related to eye blinks | ||
| 1. Add path to corresponding dependencies listed above using: | ||
| ``` | ||
| addpath(genpath('dependencyDir')) | ||
| ``` | ||
| 2. Run the analysis with [RewardsPreprocessing_withoutICA.m](https://github.com/AyaKabbara/StageEEGpre/tree/main/src/article/RewardsPreprocessing_withoutICA.m). This script will store the ERPs of all participants in a .mat file. It will also generate the csv files to be used for plotting the figures and the calculation of the descriptive statistics. | ||
| 3. Create figures 2A,3A and generate the related values in tables 1,2 using the following R script: [RewardProcessing_Plots_and_Statistics.R](https://github.com/AyaKabbara/StageEEGpre/blob/main/src/graphiques/RewardProcessing_Plots_and_Statistics.R). | ||
| Note: The name of csv files should be changed accordingly. | ||
|
|
||
| ### Reproducing Figure 2B, Figure 3B using EEGLAB | ||
|
|
||
| 1. Add path to corresponding dependencies listed above using: | ||
| ``` | ||
| addpath(genpath('dependencyDir')) | ||
| ``` | ||
|
|
||
| 2. Run the analysis [eeglab_preprocessing.m](https://github.com/AyaKabbara/StageEEGpre/blob/main/src/eeglab/eeglab_preprocessing.m). | ||
| 3. Create figures 2B,3B and generate the related values in tables 1,2 using the R script [RewardProcessing_Plots_and_Statistics.R](https://github.com/AyaKabbara/StageEEGpre/blob/main/src/graphiques/RewardProcessing_Plots_and_Statistics.R).Note: the name of csv files should be changed accordingly. | ||
|
|
||
|
|
||
| ### Reproducing Figure 2C,Figure 3C using Brainstorm | ||
|
|
||
| 1. Add path to corresponding dependencies listed above using: | ||
| ``` | ||
| addpath(genpath('dependencyDir')) | ||
| ``` | ||
| 2. Convert the dataset into EEGLAB set using the following functions: [toset.m](https://github.com/AyaKabbara/StageEEGpre/blob/main/src/BST/toset.m) | ||
| 3. Run the analysis with [bsPreprocessing.m](https://github.com/AyaKabbara/StageEEGpre/blob/main/src/BST/bsPreprocessing.m). | ||
| 4. Create figures 2C,3C and generate the related values in tables 1,2 using the R script [RewardProcessing_Plots_and_Statistics.R](https://github.com/AyaKabbara/StageEEGpre/blob/main/src/graphiques/RewardProcessing_Plots_and_Statistics.R).Note: the name of csv files should be changed accordingly. | ||
|
|
||
| ### Reproducing Figure 2D,Figure 3D using FieldTrip | ||
|
|
||
| 1. Add path to corresponding dependencies listed above using: | ||
| ``` | ||
| addpath(genpath('dependencyDir')) | ||
| ``` | ||
|
|
||
| 2. Run the analysis with [ftPreprocessing.m](https://github.com/AyaKabbara/StageEEGpre/blob/main/src/BST/ftPreprocessing.m). | ||
| 3. Create figures 2D,3D and generate the related values in tables 1,2 using the R script [RewardProcessing_Plots_and_Statistics.R](https://github.com/AyaKabbara/StageEEGpre/blob/main/src/graphiques/RewardProcessing_Plots_and_Statistics.R).Note: the name of csv files should be changed accordingly. | ||
|
|
||
| ### Reproducing Figure 4 | ||
|
|
||
| 1. Prepare the csv files containing the feature to be plotted using [fig4_preparation.m](https://github.com/AyaKabbara/StageEEGpre/blob/main/src/graphiques/fig4_preparation.m). This script calls the function [extract_features.m](https://github.com/AyaKabbara/StageEEGpre/blob/main/src/graphiques/extract_features.m). | ||
| 2. Create the figure using [Fig4.py](https://github.com/AyaKabbara/StageEEGpre/blob/main/src/graphiques/Fig4.py) | ||
|
|
||
| ### Reproducing Figure 5 | ||
|
|
||
| 1. Generate the similarity matrix (for both conditions) using the function [similarity_calc.m](https://github.com/AyaKabbara/StageEEGpre/blob/main/src/graphiques/similarity_calc.m) | ||
| 2. Create the figure using [Fig5.py](https://github.com/AyaKabbara/StageEEGpre/blob/main/src/graphiques/Fig5.py) | ||
|
|
||
| ### Reproducing Figure 6 | ||
|
|
||
| Create Figure 6.A using [eeglab_stats.m](https://github.com/AyaKabbara/StageEEGpre/blob/main/src/eeglab/eeglab_stats.m), Figure 6.B using [bs_stats.m](https://github.com/AyaKabbara/StageEEGpre/blob/main/src/BST/bs_stats.m) and Figure 6.C using [ft_stats.m](https://github.com/AyaKabbara/StageEEGpre/blob/main/src/fieldtrip/ft_stats.m) | ||
|
|
||
| ### Using Brainstorm and the raw datasets | ||
|
|
||
| 1. Convert the dataset into EEGLAB set using the following functions: [toset.m](https://github.com/AyaKabbara/StageEEGpre/blob/main/src/BST/toset.m) so that it can be recognized by BS | ||
| 2. Run the analysis with [bsPreprocessing.m](https://github.com/AyaKabbara/StageEEGpre/blob/main/src/BST/bsPreprocessing.m). Note: the paths should be changed |
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Oops, something went wrong.