Figure 2. Temporal evolution of the log-likelihood ratio (Equation 5 – thick black line).

We also show a simplified decision variable described later in the text (dashed dark grey line; section ‘Fixation-dependent drift diffusion model’ in the Results). The decision variables were computed on the basis of identical sequences of sensory evidence generated from Gaussian distributions with means μ_x = 4 and μ_y = 1, and equal standard deviations σ_x = σ_y = 2.0. The sampling strategy used for both variables is identical and can be inferred from the background of the figure with grey indicating alternative X is fixated and white alternative Y, each block of fixation is equally sized. For easier comparison the simplified decision variable is scaled by c(1-q), where c is the scaling factor in front of square bracket in Equation 6. The simplified decision variable (Equation 10) in this example assumes time is allocated equally to the two sources () and the points marked by black circles on the figure indicate the times when this assumption holds ().