General Learning Ability in Perceptual Learning

Published: 21 May 2020| Version 1 | DOI: 10.17632/hdn9tc3sv9.1
jia yang


All data and associated protocols, including code and scripts described in "General Learning Ability in Perceptual Learning ", are available to readers.


Steps to reproduce

1. Please unzip and copy all files and subfolders (data) into one directory and run "ini_subject_task.m" from there (MATLAB code). This will reproduce all results of "Effects of Initial Performance, Task, and Subject on Learning Rates". 2. running "fixedlm" from there (MATLAB code) could produce results for all 8 models fitted from "fixedlm" and examples for how to perform model comparison also represented. 3. LASSO regression was conducted in R. inputting data "feature_d2.csv" into R and run lasso_shrinkage and lasso_aic produced results of LASSO feature selection and change of R2 and AIC after removing each predictor. running "lasso_dat.m" (MATLAB code) produced the probability for each predictor being selected, the mean and standard deviation for coefficients of each predictor, the averaged goodness of fit in training and testing dataset, the change of R2 and AIC after removing each predictor, and model comparison results for removing each predictor using both F test and Likelihood Ratio Test.


Chinese Academy of Sciences, Institute of Psychology Chinese Academy of Sciences


Perceptual Learning