Script/data accompanying the paper "Abstract spatial, but not body-related, visual information guides bimanual coordination" by Brandes, Rezvani, & Heed (2017) published in Scientific Reports.
2017-12
written by Janina Brandes
=====================================
FILES NECESSARY TO RECREATE THE ANALYSIS/RESULTS FROM THE PAPER:
jb_bimanualBayesMain: main script, installs necessary packages (may take a while), creates folder structure and calls subfunctions to recreate results reported in the paper.
jb_utilityFunctions: functions to recreate the results (called from the main script).
jb_bimanualBayesModel: Bayesian model fit to the data (with JAGS); hierarchical logistic regression with all possible main effects and interactions in the fixed effects structure (4 factors, 2 levels each) and random intercepts per participant; see paper for details.
jb_mcmcSamples: MCMC samples on which the reported results are based. To re-sample the model, install JAGS and delete this file from the "models" folder (re-sampling takes about 3 days using parallel computation on a machine with 16 cores).
jb_bimanualData: .txt file containing the data.
=====================================
INSTRUCTIONS: Download the files that start with "jb_" and put them to the the same folder. Source the main script ("jb_bimanualBayesMain") from RStudio. Note that the raw movement data (see below) is not necessary to re-create the results from the paper.
=====================================
RAW MOVEMENT DATA:
sXX_dataPub.mat: contains the raw movement data per participant in a MATLAB struct.
Fields in the struct are:
dataPub(1): data of behavioral session 1.
dataPub(2): data of behavioral session 2.
dataPub(x).trial: data of each trial.
dataPub(x).trial(x).condition: description of the experimental setup.
dataPub(x).trial(x).mirrorView: codes whether there was a mirror or not in this trial (dataPub(x).mirrorCodes contains txt descriptions).
dataPub(x).trial(x).palmPosition: codes the palm position in this trial (dataPub(x).palmCodes contains txt descriptions).
dataPub(x).trial(x).movementInstruction: codes the movement instruction in this trial (dataPub(x).moveInstrCodes contains txt descriptions).
dataPub(x).trial(x).status: empty field if this trial is included in the analysis, 66 if it is excluded.
dataPub(x).trial(x).beepTime: index of the first (column 1) and last (column 2) beep of a particular movement speed (index the field "PTITime" to get the time in s).
dataPub(x).trial(x).LHTrace: 3D raw movement trace of the left hand.
dataPub(x).trial(x).RHTrace: 3D raw movement trace of the right hand.
dataPub(x).trial(x).PTITime: the time stamp from the PTI movement tracking system (in seconds). One trial is about 22 seconds long with one data point every 10 ms (sampling rate = 100Hz).
dataPub(x).trial(x).interpolated.LH: 3D interpolated movement trace of the left hand.
dataPub(x).trial(x).interpolated.RH: 3D interpolated movement trace of the right hand.
dataPub(x).trial(x).filteredLowPass.LH: 3D low pass filtered version of the interpolated movement trace of the left hand.
dataPub(x).trial(x).filteredLowPass.RH: 3D low pass filtered version of the interpolated movement trace of the right hand.
dataPub(x).trial(x).demean.LH: 3D demeaned version of the low pass filtered movement trace of the left hand.
dataPub(x).trial(x).demean.RH: 3D demeaned version of the low pass filtered movement trace of the right hand.
Movements are recorded in 3D with the following dimensions:
(1,:) = left-right/x-dimension with negative going to left.
(2,:) = front-back/y-dimension with negative going away from the body.
(3,:) = up-down/z-dimension with negative being up.
The paper is based on analysis of demeaned movements in the x-dimension.
