RANDECOD fMRI Study

This repository contains the analysis code and behavioral dataset from our fMRI study investigating the neural correlates of human randomization.

Dataset

The mat file behavioralData.mat contains each participant's age, sex and responses during the random sequence generation task across all six runs.

Description of columns in behavioralData.response of each participant:

• runNumber, blockNumber, trialNumber
• timingData: Indicates whether the button was pressed during the allotted time period ('on time') or not ('too late')
• keys : 1/2=left button press, 3/4=right button press
• fixationStart, fixationEnd: Timestamps of start and end of fixation phase in each trial
• choiceStart, choiceTimestamp: Time stamps of beginning of choice phase and actual button press respectively
• feedbackStart, feedbackEnd: Timestamps of start and end of feedback phase in each trial
• dur_choice: Time elapsed between stimulus onset (choiceStart) to button press (choiceTimestamp)
• condition: "ER": Explicit Randomness, "FC": Free Choice, "MC": Mental Coin Toss
• headsPosition: Position of heads, left or right (didn't change throughout the experiment within subject)
• focusResponse and instructionsResponse: Responses to the 5 point likert-scaled questions pertaining to the focus and instruction adherence (respectively) in each run

This mat file also includes a matrix of pseudorandomly generated binary sequences (randseq), which was created with the createPseudoRandSeq.m function. It creates 28 sequences of length 540, replicating the sequence length in the experiment, 28 representing the average samples size per condition.

The excel sheet FreeFormAnswers lists all the answers to the post-experiment questionnaire. The internalID corresponds to the internalID in behavioralData.mat. The second sheet in this file shows the text used to prompt the free-form responses.

Behavioral Analysis

Scripts include:

• LikertAnalysis.m: Descriptive Analysis of responses to focus and instruction adherence questions
• ProportionValueScript.m: Calculation of proportion values and comparison between conditions
• ReactionTimes.m: Calculation of reaction times and comparison between conditions
• RunLengthScript.m: Calculation of run lengths and comparison between conditions
• RandomnessMeasures.m;markovOrderBIC.m: Calculation of conditional entropy and optimal Markov order values and comparison between conditions
• createPseudoRandSeq.m: Generates pseudorandom sequences (see above)

Imaging Analysis

This folder contains the preprocessing script of the imaging data as well as the code for the univariate analysis (first&second level) and multivariate analysis (first level, searchlight decoding, second level).