levylab.RNA

This package provides reusable code for the analysis of risk & ambiguity task data, as developed by the Yale Decision Neuroscience Lab.

How to use this toolbox

Simple invoke library(levylab.RNA) in your R session or your R code.

Installation

1. install.package('devtools')
2. library(devtools)
3. devtools::install_github('YaleDecisionNeuro/levylab.RNA')

For extracting model-based features, you will also need to install nlopt.

Expected input

The analysis scripts in this repository expect a tidy CSV file with particular columns. Here's how you can get it from our current projects:

Any task based on PsychTaskFramework

1. From the root folder of PsychTaskFramework, run exportTaskData(nameOfYourTask, outputFile). The script expects that your data are saved in tasks/nameOfYourTask/data/; if they aren't, put them there.
2. Run importFromPTF(outputFile).

PsychToolbox-based study of veterans (VA Risk, VA_fMRI_PTB)

1. Get all the raw .mat files together in a single directory - let's call it originDirectory.
2. Run importFromRawMat(file) on each .mat file in the repository.

Tasks based on E-Prime

1. Merge all the .edat2 files in the task folder with E-Merge. (This will require a prior installation of E-Prime, even though you should be able to do this without an activated license.)
2. Open the newly created merged file with E-DataAid and Save as/Export as "SPSS and StatView".
3. Read in the exported file with R: importFromEprimeSPSSExport(filename, choiceColumnName = "choice", discardOriginalColumns = FALSE).
4. Drop the columns that you don't need manually, or import with discardOriginalColumns = FALSE.
5. Use or export to CSV as needed.

Functionality

Extract model-free features

Run getModelFreeEstimates(decision_data) on a clean R&A data frame.

Compute model-based fits in R with nloptr

Run getModelBasedEstimates(decision_data) on a clean R&A data frame.

Compute model-based fits in Matlab with fmincon

1. Save the clean R&A data frame as CSV with write.csv(clean.df, paste0('clean/', filename), row.names = FALSE).
2. Locate the matlab files with system.file("matlab", "fit_matlab_model.m", package = "levylab.RNA").
3. Change the clean choice file location in that file accordingly and run in Matlab.