Figure 4.

Distribution of t_SDs

Each row represents a coherence level, each column a subject. Ordinate is proportion of responses with the t_SDs from the controlled-duration experiment, signed by the direction of response (positive, right; negative, left). The scale of each row is normalized to fit the row’s height for visualization. Black lines are the data, and blue lines are the drift diffusion fit with collapsing bounds (parameters in Table S3). The data are smoothed in time with a Gaussian kernel (σ = 0.05 s) for visualization. Goodness of fit, quantified by the Jensen-Shannon divergence, is displayed in Figure S4. The more elaborate model also accounts for the reaction time distributions (and slow errors) in the free-response task (data not shown). See also Figure S3.