The Readiness Potential reflects internal source of actions, not decision uncertainty

Eoin Travers & Patrick Haggard

[Data and Code]

Abstract

Voluntary actions are preceded by the Readiness Potential (RP), a slow EEG component generated in supplementary motor area. The RP is usually thought to be specific to internally-driven decisions to act, and reflect post-decision motor preparation. Recent work suggests instead that it may reflect noise or conflict during the decision itself, with internally-driven decisions tending to be more random, more conflictual and thus more uncertain than externally-driven actions. To contrast accounts based on endogenicity with accounts based on uncertainty, we recorded EEG in a task where participants decided to act or withhold action to accept or reject gambles. We found no difference in an RP-related motor component when comparing actions driven by strong versus weak evidence, indicating that the RP does not reflect uncertainty. In contrast, the same RP-related component showed higher amplitudes actions performed without external evidence (guesses) than for those performed in response to equivocal, conflicting evidence. This supports the view that the RP reflects the internal generation of action, rather than decision uncertainty.

This repository contains the data, experiment, and analysis scripts for this manuscript.

It is hosted on GitHub (https://github.com/EoinTravers/ezmc) and the Open Science Framework (https://osf.io/m834c/).

Contents

• Procedure/* Python/PsychoPy scripts used to run the experiment
• Results/* Data and anlalysis scripts
  • src/*: Shared python and R functions
  • data/*: Data files. See notes below if downloading from GitHub.
  • Behaviour.Rmd: Behavioural analyses
  • Behavioural_Supplementary.Rmd: Supplementary behavioural analyses
  • 0.Preprocess.ipynb: EEG preprocessing script. Note that raw EEG isn't included, so this can't be run
  • 1.Merge.ipynb: Merge preprocessed EEG. Also can't be run
  • 2.ERPs.ipynb: Plot simple ERPs. This can be run
  • 3a.PCA_Rotation.ipynb: PCA-CSD stuff
  • 3b.PCA_Supplementary.ipynb: Some extra information about the PCA-CSD approach
  • 6.EasyVsDifficult.Rmd: Comparison of Easy and Difficult decisions
  • 7.GuessVsDifficult.Rmd: Comparison of Guesses and Difficult decisions
  • TraversHaggard2019.Rproj: RStudio project file

Data

We provide the combined, preprocessed EEG data for this experiment. If you're accessing these files on GitHub, you'll need to run the download.sh script in the data folder.

Requirements

Python, R, various packages.

Confusingly, the experiment script requires python 2, but the EEG analysis scripts use python 3.