Figure 1.

The left-hand vertical distribution represents memory strength or perceptual strength (depending on the task). Three confidence criteria divide the distribution into four areas—A, B, C, and D—and the sizes of the areas determine the drift rates for the four accumulators, as shown by the arrows. The middle column shows paths of evidence accumulation in the accumulators, the red paths for one test item, the black for another, and the blue for another. The right hand column shows heat maps for 2,000 simulated decision processes for the constant summed evidence model when the top process wins. The histogram on the top plot shows the distribution of finishing times. The top and bottom processes have low decision boundaries (b₁ andb₂ in Table 1), and the middle two plots have higher criteria (b₃ and b₆ in Table 1). The mean of the memory evidence distribution was 0.0, and the other model parameters were those from the averages of the fits shown in Table 1. conf. = confidence.