Autistic traits influence the strategic diversity of information sampling

To replicate the results and figures of the article, try to run

1. first main.R,
2. then figures.R and figures_sup.R

R (>= 3.5.3) are required; R packages mentioned in the codes also have to be installed first. Running codes in RStudio is optional.

Data

In Data folder, there are four .rds files, including anonymous subject information and choice behavior results:

• subject_data.rds: Basic participant information
• choice_data.rds: Participants' choices and experiment conditions of each trial
• rt_data.rds: Participants' decision times of each key press and corresponding stimuli (blue or pink bead)
• in_trial_sim.rds: Pre-generated stimuli that a participant would encounter during the experiment, which was used to do simulation

Other files and sub-folders are for choice models and corresponding RT¹ models and for Bayesian model comparisons.

• choice_rt_combined_model_(aicc|bic)(_output).csv: Choice + RT models' model comparisons and outputs
• choice_model folder: Estimations of choice models' parameters and log-likehood
• rt_model folder:
  • (ts|os)_.+_rt.csv: RT-related data calculated by choice models (i.e., RT and probability), which was used to estimate RT models
  • twelve_ts_plus_four_os foler: Modeling results (mr_) for corresponding RT models

Codes

Contained in Codes folder are mostly helper functions or codes for simulation used in main.R, which are either less necessary or too much too long to be written in the main code file.

Several code files were written in MATLAB; thus they need to be run in MATLAB environment:

• group_level_MBS_choice_rt_combined.m: Group-level Bayesian model comparison for Choice + RT models
• choice_model folder: Codes that fit various choice models
• rt_model folder: Codes that fit different RT models

1: "RT" here and in codes and data refers to decision time (DT) that appears in the main text.