Code and data for: Baranger DAA, Halchenko YO, Satz S, Ragozzino R, Iyengar S, Swartz HA, Manelis A. Aberrant levels of cortical myelin distinguish individuals with depressive disorders from healthy controls. NeuroImage: Clinical (2021) 32:102790. doi:10.1016/J.NICL.2021.102790
MethodX_data/:
data/
mri_derivatives/
primary/ses-01 ** surface-level T1w/T2w ratio and cortical thickness images for subjects in primary analyses
followup/ ** surface-level T1w/T2w ratio and cortical thickness images for the subject with follow-up data who converted
ses-01/ ** follow-up subject session 1
ses-02/ ** follow-up subject session 2
other_input/ ** clinical, demographic, and parcellated T1w/T2w ratio data
outputs/
cvs/ ** output of posthoc analyses varying the number of cross-validation folds
loocv_inneronly ** output of posthoc analyses varying the number of inner cross-validation folds
glmnet/ ** output of primary glmnet analysis
permutations/ ** output of permutation analyses
preprocessing/ ** output of preprocessing outlier detection
regressions/ ** output of regression analyses
scripts/
analyses/ ** scripts for primary analyses, including glmnet, permutations, and regressions
figures/ ** scripts to create figures in the paper
followup/ ** scripts for post hoc analyses
preprocessing/ ** scripts for parcellating mri derivative files and outlier detection
MethodX_data/data/mri_derivatives/primary/ses-01/sub-*/
sub-*.L.midthickness.32k_fs_LR.surf.gii ** left cortical thickness file
sub-*.R.midthickness.32k_fs_LR.surf.gii ** right cortical thickness file
sub-*.SmoothedMyelinMap_BC.32k_fs_LR.dscalar.nii ** cortical myelin cifti file
MethodX_data/data/other_input/
converted_participant_parcels_bothsessions.xlsx ** Cortical myelin values for the follow-up subject who converted
data_360parcels_Glasser32K.csv ** Cortical myelin for 360 Glasser parcels, output of scripts/preprocessing/parcellate.R
data_clinical_and_parcels_all.csv ** Participant demographics, clinical variables, and cortical myelin values
data_dictionary.csv ** Description of columns in data_clinical_and_parcels_all.csv
ElasticNet_variables.csv ** All variables used for elastic net analyses
glmnet_performance.csv ** Performance metrics for glmnet/LDA classifier
MethodX_data/outputs/cvs/
glmnet_leave-one-out_nested_1uniquepair_removed_HC_UD_[x]_outer_[y]_internalfolds_2021-07-02.txt
** output of follow up analyses, varying both the internal [y] and outer [x] cv folds
MethodX_data/outputs/cvs/loocv_inneronly/
glmnet_leave-one-out_nested_1uniquepair_removed_HC_UD_[i]_internalfolds_2021-07-02.txt
** output of follow up analyses, varying the number of inner cv folds [i] (retaining 2 pairs held-out)
MethodX_data/outputs/glmnet/
glmnet_variable_selection.csv ** frequency of variable selection in true and permutation analyses
glmnet_with_age_sex_iq_leave-one-out_nested_1uniquepair_removed_HC_UD_2020-11-26.txt ** main results, output of scripts/analyses/glmnet_with_LDA_myelin_paper.R
predict.followup.txt ** predicted class for followup participant who converted mid-study
MethodX_data/outputs/permutations/
split[i]_glmnet_permuted_labels_10times_leave-one-out_nested_with_age_sex_iq_1uniquepair_removed_HC_UD_2021-02-22 ** Outputs of permutation analyses (100 permutations per file)
split[i]_glmnet_permuted_sets_10times_leave-one-out_nested_with_age_sex_iq_1uniquepair_removed_HC_UD_2021-02-22 ** Outputs of permutation analyses (100 permutations per file) - record of all permutation combinations
glmnet_permuted_labels_100times_for_leave-one-out_nested_with_age_sex_iq_1uniquepair_removed_HC_UD_2021-02-22.txt ** Combined all label files
glmnet_permuted_sets_100times_for_leave-one-out_nested_with_age_sex_iq_1uniquepair_removed_HC_UD_2021-02-22.txt ** Combined all set files
MethodX_data/outputs/preprocessing/
outlier.results.csv ** results of parcel outlier detection
MethodX_data/outputs/regressions/
regression_dd_control_myelin.csv ** results of regression analyses between control/dd and myelin
regression_demo_clin_myelin.csv ** results of regression analyses between clinical/demographic variables and myelin
regression_performance_myelin.csv ** results of regression analyses between lda accuracy and myelin
MethodX_data/scripts/analyses/
glmnet_with_LDA_myelin_paper.R ** Nested cross-validation elastic net regression with LDA (primary analysis)
permuted_glmnet_with_LDA_myelin_paper.R ** Permutation analyses
process_glmnet_output.R ** Compute performance metrics of the primary analysis
regression_lda_performance_and_clinical.R ** Regression analyses between model accuracy and clinical variables
regression_myelin_and_clinical.R ** Regression analyses between cortical myelin and clinical variables
regression_myelin_patients_vs_controls.R ** regression analyses comparing myelin values in patients and controls
MethodX_data/scripts/figures/
brain_plot.R ** Code for Figure 3
performance_plots.R ** Code for Figure 2
plot_antidepressants.R ** Code for Supplemental Figure 3
MethodX_data/scripts/followup/
glmnet_with_LDA_myelin_paper_cvs_outerloop.R ** repeating the glmnet analyses, varying both the number of inner and outer cv folds
glmnet_with_LDA_myelin_paper_cvs.R ** repeating the glmnet analyses, varying the number of inner cv folds
Predict_converted.R ** predict the group (control, DD) of the participant who converted
process_cvs_parcels.R ** process the output of glmnet_with_LDA_myelin_paper_cvs.R & glmnet_with_LDA_myelin_paper_cvs_outerloop.R
MethodX_data/scripts/preprocessing/
outlier_regions.R ** Detects outlier parcels
parcellate.R ** Parcellate mri derivative files to compute mean T1w/T2w ratio value for each parcel
https://balsa.wustl.edu/88mp ** Glasser 360 parcellation atlas