The task design largely followed Bohrn et al. (
2012). Each participant completed eight sessions, each consisting of 20 trials. A session started with a five-point calibration followed by a validation, and drift check and correction were performed after every four trials. Each trial comprised three phases: the fixation (1,000 ms), stimulus (4,000 ms), and task (1,500 ms) phase. In the fixation phase, participants focused on a fixation cross (0.8° × 0.8° in VA) presented at the central position of the first word of the stimulus that would follow. Next, the stimulus (center-aligned one-line diction; Arial; 30 pt.; height 1.0°, width varying from 5.2° to 24.9°) was presented at the center of the screen, accompanied by another fixation cross (0.8° × 0.8°) at 3.5° below the text. Participants were instructed to read and understand the stimulus silently and to avoid rereading (and hence regressions) as much as possible. They were also instructed to focus on the fixation cross below the text after finishing the reading. Throughout the manuscript, we call this postreading fixation an “ending fixation.” While the majority of reading-related eye-tracking studies applies a method in which the stimuli disappear immediately after participants stop reading (e.g., Fernández, Shalom, Kliegl & Sigman,
2014), we applied this “ending fixation” method to keep the presentation time of the stimuli constant (4,000 ms). Since we planned to apply the curve-fitting method to the pupillary waveforms (see
Methods, Pupillary data section), we needed to obtain a dataset containing the same number of data points for all trials. In the task phase, participants performed a semantic categorization task for each stimulus. A category cue (“everyday life,” “health and well-being,” “love and relationship,” or “work and success”) was presented on the screen with the choices of no or yes (“N / Y”), and participants indicated whether or not the stimulus fitted into that semantic category by pressing a button within a response period of 1,500 ms, during which the category cue was shown. The results of the semantic categorization task were not analyzed further, since the main purpose of the task was to keep participants engaged with the stimuli and motivate the language processing. A schematic representation of a trial is shown in
Figure 1. Note that, although participants were instructed to read and understand the stimuli, they did not perform any explicit tasks, such as making decisions or giving aesthetic evaluations
while the stimuli were presented (the main time window of the data analysis). Each participant completed a total of 160 trials (20 trials × 8 sessions), responding to the 160 stimuli, which were each presented once and in randomized order. The eye-tracking part took approximately 20 to 25 min (varying by resting time between sessions).