Table 1.

List of all predictors in the four different types of the model

PredictorReduced modelBase modelDrop modelFull model
Online performance monitoring
    Error or error*direction√*√*√*√*
    Difficulty or difficulty*direction
    Error*difficulty√*√*√*√*
Task-related confounds
    Target configuration(√)
    Dot direction(√)
    Dot speed(√)
    Reward context(√)
    Action value(√)
    Chosen value(√)
Choice saccade
    Saccade direction(√)
    Reaction time(√)
    Peak velocity*direction(√)
    Amplitude*direction(√)
    Duration*direction(√)
    Main sequence (velocity/amplitude)*direction(√)
Type 1 saccade
    Number of Type 1 saccades(√)
    Amplitude(√)
    Peak velocity(√)
    Horizontal displacement(√)
    Vertical displacement(√)
Type 2a saccade
    Latency of error saccade(√)
    Latency of correct saccade
    β (memsac)
Type 2b saccade
    NA
Type 3 saccade
    (Number of Type 3 saccades)*direction(√)
  • A check mark indicates that a regressor is part of the model. A check mark with an asterisk indicates that this regressor is only present in models that include error trials. A check mark in parentheses indicates that the predictor was included only if it had a p values of <0.4 in the full model. An asterisk between to predictors indicates that both predictors as well as their interaction was included in the model.