Code and processed data for: Neural codes track prior events in a narrative and predict subsequent memory for details

This repository contains all code used for (Collin et al., neural codes track prior events in a narrative and predict subsequent memory for details, on biorxiv: https://www.biorxiv.org/content/10.1101/2024.02.13.580092v3), commented but non-runnable, as well as processed data. The raw data can be found on openneuro (https://openneuro.org/datasets/ds005050). 

behav_PredictionsTask-Day1.ipynb, behav_PredictionsTask-Day2.ipynb, behav_schematest.ipynb This contains all code for analyzing the behavioral data (schema prediction questions during video viewing for day 1 and 2 (as used for figure 4A), schema test at end of day 2 (as used for figure 4B).

brainToBehavLink.ipynb This script calculates the within-subject brain to behavior (memory for details, memory for rituals) correlations and compares this to a null distribution (as used for table 1 and figure 11).

tval_evaluate_coloredBrain_fingerprints.ipynb This script takes the calculated p-values for each searchlight as calculated before and plots that as a whole-brain .nii file (as used for figure 5). It saves both a brainmap with all p-values as well as an fdr-corrected one

get_distilled_neural_measures_SHPC.py This script averages over searchlights to calculate average values for the brain to behavior link. RSA_day2_2021march1_E2.py, RSA_day2_2021march1_E3.py and RSA_day2_2021march1_E4.py These scripts are used to calculate the neural similarity timelines as used for the violinplots (described below).

paperFigures.ipynb and paperFigures-GrandAverages.ipynb These scripts take the rsa neural similarity TR-by-TR timelines, averages over time for each event, and create 3-by-3 plots (paperFigures.ipynb) or within vs across-stage averages (paperFigures-GrandAverages.ipynb), separately for each of other  neural codes in the study (schema, rotated, path, event). Scripts as used for figure 6 to 10 and A1, A2, A3.

Processed data data_for_all_behavTests 
This repository has processed data csv files used for analyzing the schema learning during day 1 (day1_SchemaPrediction_coinTorchPrediction.csv and day1_SchemaPrediction_eggPaintingPrediction.csv), day2 schema test (day2_test_for_schemalearning.csv), day2 stop-and-ask prediction questions (day2_SchemaPrediction.csv), and data of the recall task (day2_recall_EpisodicDetails.xlsx, day2_recall_Rituals.xlsx). data_for_brainToBehaviorCorrelation

This repository has processed data used for the brain-to-behavior correlations (within subject) for each of the neural codes assessed in the study (saved as 6 json files). data_for_brainToBehaviorCorrelation

This repository has processed data used to test significance for the brain-to-behavior correlations (all correlations to memory for details as well as to memory for rituals was compared against a null distribution), for each of the neural codes assessed in the study (saved as csv files that end with *nullDistribution_NEW.csv). data_for_rsa_brain_maps 

This repository has processed data related to the RSA analyses in the form of nii files, one file for each of the neural codes, with an fdr 0.05 threshold (saved inverted, i.e. 1 minus pvalue). data_for_rsa_timelines 

This repository has processed data concerning the rsa violin plots for each of the neural codes assessed in the study. Each csv file has neural similarity values TR-by-TR for each participants. Each file used either event2, event3 or event4 template to compare it to, and either for other-event/other-schema, other-event/same-schema, same-event/same-schema or same-event/other-schema. Each of the 5 neural codes has 12 csv files saved for creation of violin plots, one for same-event-same-schema, one for same-event-other-schema, one for other-event-other-schema, and one for other-event-same-schema, and additionally all 4 of those separately for event2, event3 and event 4 templates.