The prosocial_effort_neural_rep from amberxuqianchen

This folder contains the scripts and functions to run:

i) effort-discounting models (with maximum likelihood estimation; mle) on participant data [subfolder Model_real_data]
ii) model identifiability and parameter recovery [subfolder Model_simulated_data]
iii) analysis on the model parameters, trial by trial behavioural data, and MRI data with plots [subfolder PM_R_code]

Click here to view plots and main results

For analysis of the real participant data (i & iii):

Step 0 (if data not already in MATLAB structure) - Convert data to the required format from Presentation .log / .csv / .xlsx formats

Step 1 - Run_mle_model.m

Script to run and compare models using maximum likelihood estimation fit

Output from script

matlab variables - 's.PL.ml' - contains model results including the model parameters for each participant - 'output.winModel7' - compares two main models (7 & 10), used in model_comp_7_10_relative_BIC.csv
datafiles in specified output directory:
- K_values_PM_fmri_two_k_one_beta.csv - estimated parameters for each participant

Step 2 - create datafile

Match parameters from K_values_PM_fmri_two_k_one_beta with other behavioural and MRI data in a wide data format for analysis (see PM_fmri_questionnaire_wo_excluded_totals_share.csv)

Step 3 - Create data files of trial by trial behavioural data (see lme4_PM_data_6_21.csv)

Step 4 - Prosocial_effort_analysis.Rmd

Run analysis using R project, script, and files from above output (note sections of this script also plot results from simulation experiments - model identifiability and parameter recovery - see below and use results from RSA analysis).

For simulation experiments (ii):

Step 1 - Simulate_PR_MI_data_PM.m

Script to run model identifiability and / or parameter recovery

Step 2 - Prosocial_effort_analysis.Rmd

Plot results using R script

Prosocial effort discounting models

Based on Lockwood et al. (2017), Nature Human Behaviour - test different variations of k and beta parameters

Models compared combine all combinations of single or separate k and single or separate beta parameters:

- one_k_one_beta

- two_k_one_beta

- one_k_two_beta

- two_k_two_beta

and different shapes for the discounting (k) parameter:

- parabolic

- linear

- hyperbolic

Developed using:

MATLAB 2019b - requires Econometrics and Bioinformatics toolboxes

macOS 10.15 Catalina / 11.1 Big Sur

R version 3.6.2 (2019-12-12)

amberxuqianchen / prosocial_effort_neural_rep Goto Github PK

prosocial_effort_neural_rep's Introduction