Common and stimulus-type-specific brain representations of negative affect

Summary and requirements

This repository contains code and data related to Ceko et al. 2022 to develop, evaluate and validate in new individuals common and stimulus-type specifc predictive brain models of negative affect, using 4 types of aversive stimuli (mechanical pain, thermal pain, aversive sound, aversive pictures)

• code tested on MATLAB 2019b

• code requires CANLAB core tools and a few other tools, all loaded using a2_mc_set_up_paths.m

• code is in /scripts, unless otherwise noted

• Full sample unthresholded PLS pattern maps for use in independent samples

• input data for some of the analyses is on Dropbox (to discuss access, contact Marta at marta.ceko@gmail.com):

  • data/data_behavior/ -> excel spreadsheet containing ratings, to load with import_Behav_MPA2.m (code in /scripts) Dropbox Link

  • results/ -> various other inputs Dropbox Link

    • results/data_objects.mat -> 1st-level GLM beta images used for PLS
    • results/PLS_crossvalidated_N55_gm.mat -> stats related to cross-validated and full sample PLS models
    • results/PLS_bootstats10000_N55_gm.mat -> bootstrapped stats
    • results/patterns/PLS_CV_patterns -> CV image files (5 per model representing the 5 folds) for use within sample
    • results/patterns/PLS_patterns -> Full sample bootstr. PLS pattern maps - like the link above but with more options: unthr, unc01, unc001,fdr-05)

Overview of code to create main Figures (/scripts):

Figure 2

• Fig2a_behavior_plots.m
• Fig2b .... Model evaluation
• Fig2c .... Crosspred matrix
• Fig2d .... Variance decomposition

Figure 3

• Fig3a3b_display_maps.m
• Fig3c3d_riverplots_roi.m
• Fig3e_plot_common_specific_importance.m

Figure 4

• Fig4_plot_figure_architecture.m

Extended Data Figure 1

• scripts2/EDFig1b_plot_normPLS_signatures.m
• scripts2/EDFig1c_display_maps.m
• scripts2/EDFig1d_roiplots_univariate.m
• scripts2/EDFig1d1e_roiplots_and_3D_encoders.m