Embracing the fuzziness of cognition and behavior and its relationship with environmental factors: A multi-cohort study.
Authors: Anthony Gagnon1, Virginie Gillet1, Anne-Sandrine Desautels1, Jean-François Lepage1, Andrea Baccarelli2, Jonathan Posner3, Maxime Descoteaux4, Marie Brunet1, Larissa Takser1.
Affiliations:
1 Department of Pediatrics, University of Sherbrooke, Québec, Canada
2 Department of Environmental Health, Harvard T. H. Chan School of Public Health, Boston, MA, USA
3 Department of Psychiatry and Behavioral Sciences, Duke University, Durham, NC, USA
4 Sherbrooke Connectivity Imaging Lab (SCIL), University of Sherbrooke, Quebec, Canada
The present repository contains all relevant code and scripts to reproduce the results found in Gagnon et al. 2024.
In the models/ folder, you can find the PCA model used for dimensionality reduction of the ABCD cohort. In addition, you will also find the centroids needed for the projection of your data in the profiles' space. To use it in your research, simply follow the steps regarding the GESTE and BANDA cohorts in the jupyter notebooks located in the notebooks/ folder. Briefly, you should follow the residualization, reduction (EFA/CFA), harmonization, and imputation (if needed) steps found in notebooks/2-DataPreprocessing.ipynb, then you can proceed with the projection by following the steps in notebooks/3-FCMeansClustering.ipynb. Please raise an issue if you have any questions.
If you do not have the required packages installed, it is recommended that you setup a new python virtual environment in which we will install the required dependencies.
Installing virtualenv
Please run the following in your terminal window:
pipx install virtualenvSetting up a new virtualenv with python 3.11
virtualenv --python 3.11 /path/to/your/destination/folder/
# Activate your newly created environment
source /path/to/your/destination/folder/bin/activateInstalling NeuroStatX
Important
Run the following commands from within your virtual environment.
pip install neurostatx==0.1.0
# Test the installation by calling the help of a CLI script.
AddNodesAttributes -hInstalling GraphViz
For visualization of semplots from the semopy python package, GraphViz is required. If you do not have it installed, please run the following if you are on Linux sudo apt get graphviz or brew install graphviz if you are on MacOS.
By design, the notebooks are made to be run in a particular order (denoted by the numbering in the file name). If you run all of them, you should obtain this specific output file structure:
<output_directory>
├── awp
│ ├── AWP_ABCD
│ ├── AWP_ABCD_youth
│ ├── AWP_BANDA
│ ├── AWP_GESTE
│ └── AWP_GLOBAL
├── datagathering
│ ├── abcd_data.xlsx
│ ├── abcd_dx_labels.xlsx
│ ├── banda_data.xlsx
│ ├── banda_dx_labels.xlsx
│ └── geste_data.xlsx
├── demographicstable
│ └── demo_table.xlsx
├── fuzzyclustering
│ ├── ABCDFuzzyCMeans
│ ├── BANDAProjected
│ ├── GESTEProjected
│ ├── GraphNetwork
│ ├── GraphNetwork.gml
│ ├── VizNetwork
│ ├── VizNetworkDxABCD
│ ├── VizNetworkDxBANDA
│ ├── VizNetworkDxGESTE
│ ├── VizNetworkDxGlobal
│ └── merged_fcm_data.xlsx
├── LongitudinalProfiles
│ ├── datagathering
│ │ ├── abcd_data_2y.xlsx
│ │ ├── abcd_data_4y.xlsx
│ │ ├── abcd_data_baseline.xlsx
│ │ ├── abcd_dx_labels_2y.xlsx
│ │ └── abcd_dx_labels_baseline.xlsx
│ ├── datapreprocessing
│ │ ├── ABCD_2y_Apply
│ │ ├── ABCD_2y_CFA
│ │ ├── ABCD_2y_EFA
│ │ ├── ABCD_4y_Apply
│ │ ├── ABCD_4y_CFA
│ │ ├── ABCD_4y_EFA
│ │ ├── abcd_2y_preprocessed.xlsx
│ │ ├── abcd_2y_residualized.xlsx
│ │ ├── abcd_4y_preprocessed.xlsx
│ │ └── abcd_4y_residualized.xlsx
│ ├── demographics
│ │ └── demo_table_followups.xlsx
│ └── fuzzyclustering
│ ├── ABCD_2y_FCM
│ ├── ABCD_4y_FCM
│ ├── GraphNetwork2y
│ ├── GraphNetwork4y
│ ├── RadarPlot2y.png
│ ├── RadarPlot4y.png
│ ├── Tukey_results_2y.xlsx
│ ├── Tukey_results_4y.xlsx
│ ├── VizNetwork2y
│ ├── VizNetwork4y
│ ├── anova_2y.xlsx
│ ├── anova_4y.xlsx
│ ├── fcm_2y.xlsx
│ ├── fcm_4y.xlsx
│ └── silhouette_scores.png
├── preprocessing
│ ├── ABCD_CFA
│ ├── ABCD_CFA_Apply
│ ├── ABCD_EFA
│ ├── BANDA_CFA
│ ├── BANDA_CFA_Apply
│ ├── BANDA_EFA
│ ├── GESTE_CFA
│ ├── GESTE_CFA_Apply
│ ├── GESTE_EFA
│ ├── abcd_data_preprocessed.xlsx
│ ├── abcd_data_residualized.xlsx
│ ├── banda_data_preprocessed.xlsx
│ ├── banda_data_residualized.xlsx
│ ├── geste_data_preprocessed.xlsx
│ └── geste_data_residualized.xlsx
├── profileanalysis
│ ├── ABCD_coef1.png
│ ├── ABCD_coef2.png
│ ├── ABCD_coef3.png
│ ├── ABCD_coef4.png
│ ├── ABCD_coef5.png
│ ├── ABCD_coef6.png
│ ├── ABCD_coef7.png
│ ├── ABCD_coef_heatmap.png
│ ├── ABCD_env_loadings.png
│ ├── ABCD_plsr_coef.xlsx
│ ├── ABCD_plsr_coef_pval.xlsx
│ ├── ABCD_plsr_coef_pval_fdr_corrected.xlsx
│ ├── ABCD_plsr_scatter.png
│ ├── ABCD_plsr_stats.xlsx
│ ├── ABCD_profile_loadings.png
│ ├── GESTE_coef1.png
│ ├── GESTE_coef2.png
│ ├── GESTE_coef3.png
│ ├── GESTE_coef4.png
│ ├── GESTE_coef5.png
│ ├── GESTE_coef6.png
│ ├── GESTE_coef7.png
│ ├── GESTE_coef_heatmap.png
│ ├── GESTE_env_loadings.png
│ ├── GESTE_plsr_coef.xlsx
│ ├── GESTE_plsr_coef_pval.xlsx
│ ├── GESTE_plsr_coef_pval_fdr_corrected.xlsx
│ ├── GESTE_plsr_scatter.png
│ ├── GESTE_plsr_stats.xlsx
│ ├── GESTE_profile_loadings.png
│ ├── abcd_barplot_dx.png
│ ├── all_studies_barplot.png
│ ├── banda_barplot_dx.png
│ ├── circular_ABCD_dx.png
│ ├── circular_BANDA_dx.png
│ ├── circular_GESTE_dx.png
│ └── circular_barplot_all.png
└── viz
├── ANOVA_results.xlsx
├── ANOVA_results_cohorts.xlsx
├── NetworkABCD_AD_youth.png
├── NetworkABCD_DD_youth.png
├── NetworkABCD_PSYPATHO_youth.png
├── NetworkCohort.png
├── RadarPlotABCD.png
├── RadarPlotBANDA.png
├── RadarPlotCombined.png
├── RadarPlotGESTE.png
├── Tukey_results.xlsx
└── Tukey_results_cohorts.xlsx