Table 2.

Bayesian learning model parametersa

Model parameterBaselinetDCS − Baseline
AnodalCathodalShamAnodalCathodalSham
β0.0846 ± 0.03180.0707 ± 0.03580.0808 ± 0.0403−0.0031 ± 0.03320.0115 ± 0.02670.0107 ± 0.0480
ϕ0.0304 ± 0.03490.0133 ± 0.01620.0700 ± 0.0770−0.0285 ± 0.0313−0.0018 ± 0.0114−0.0638 ± 0.0699
λ̂0.94890.95370.9406−0.0073−0.0083−0.0127
θ̂31.18919.27912.326−2.34061.951911.465
σ̂d0.00100.00280.00100.0025−0.00010.0053
  • aThe first three columns show the parameter estimates across all participants and trials within each stimulation group (anodal, cathodal, or sham) in the baseline condition. Columns 4 to 6 show the change in the parameter estimates under tDCS relative to baseline for each stimulation condition. Parameters β and ϕ were allowed to vary individually across participants and were fit in a hierarchical manner (see Materials and Methods). For these parameters, the table reports the means (± standard deviations) of the median values of each subject's posterior distribution. The remaining parameters were fit across the population; for these parameters, the table reports the medians of the population posterior distribution.