Fig. 7: Task and results for Experiment 4 which uses stimuli on a continuous scale.

a Three sets of dots moved in three different directions separated by 120°. Similar to Experiments 1–3, one of the three sets of dots had more dots (i.e., dominant direction) compared to the other two sets (i.e., nondominant directions). Each trial began with a fixation cross followed by the moving dot stimulus presented for 500 ms. The response screen was presented immediately after the offset of the moving dots and randomly assigned a stimulus-response mapping on each trial. Similar to Experiments 1 and 2, subjects picked the dominant direction of motion between all three directions (three-alternative condition) or between the dominant direction and one randomly chosen nondominant direction (two-alternative condition). b Mean accuracy in the two-alternative condition observed in the actual data (white bar) and predicted by the population (light gray bar) and summary (dark gray bar) models. The predictions for both models were derived based on the data in the three-alternative condition. All p values are derived from two-sided paired t tests. Error bars represent SEM, n = 11. c Individual subjects’ differences in the accuracy of the two-alternative condition between the predictions of each of the two models and the observed data. d AIC difference between the population and summary models. Positive AIC values indicate that the summary model predicts the observed data better. The red triangle indicates the average AIC difference. The summary model fits better than the population model for nine out of the 11 subjects.