Figure 1.
Example stimuli for the A) scene memorization task, B) reading task, C) scene search task and D) pseudo-reading task.
Figure 2.
Classification accuracies for identifying one of the four tasks based on eye-movements are shown for different classification models.
Accuracies for the 12 participants for each type of classification are summarized in a boxplot. On each box, the central mark is the median, the edges of the box are the 25th and 75th percentiles, the whiskers extend to the most extreme values not considered outliers, and values beyond the 1.5 interquartile ranges are marked with pluses. The mean classification accuracies across the 12 participants are shown as dots.
Figure 3.
Panel A: Confusion matrices for identifying one of the four tasks from one day to another for each participant, ordered by classification accuracies (shown above each matrix). The value of each element denotes the proportion of trials identified as the corresponding label to the total number of trials in the actual category. For example, the first row in a confusion matrix indicates the proportions of all the pseudo-text reading trials that were classified as pseudo-text reading, text-reading, scene memorization, and scene search. A perfect classification results in a confusion matrix with 1 s on the diagonal and 0 s on off-diagonal elements. Panel B: Averaged confusion matrix across the participants.
Figure 4.
Classification accuracies for identifying two scene-related tasks based on eye movement patterns are shown for different classification models.
Accuracies of 12 participants for each type of classification are summarized in a boxplot as in Figure 2.