Apparatus

The experiment apparatus (Fig. 5a) consisted of an LCD monitor (51 × 25.5 cm; 1920 × 1080 pixels; 60 Hz) and the Tobii TX300 eye tracker (Tobii, Danderyd, Sweden) configured to record eye gaze at 60 Hz. Our reasoning for using 60 Hz was partly pragmatic. Affordable eye trackers with similar sampling rate capabilities are available, and could be used for home monitoring (e.g., screening for VF loss while watching TV) if found to be effective⁴⁹. This sampling rate was also matched to the refresh rate of the screen, and helped to minimize the computational overhead of generating the artificial scotoma (i.e., a higher sampling rate may have added addition delay or imprecision into to the overall system). However, there was a small amount of lag before any changes in gaze could be registered. The manufacturer claims that the overall latency of its eye tracker system is <10 msec. If we further factor in the refresh rate of the screen (60 Hz) and 3D rendering time, the total expected lag was approximately 35–45 msec. Studies⁵⁰ have shown that system delays of up to 60 msec have an insignificant effect in disrupting perceptual processing. Moreover, the effect of system latency of gaze-contingent setups on large scotomas is low when compared with small central scotomas⁵¹. In the Supplemental Material we also report additional control data, collected post-hoc, in which no substantive differences in recorded eye-movements were observed when using a hardware setup with a higher refresh rate (240 Hz).

Apparatus, stimuli, and procedure. (a) The experiment apparatus consisted of an LCD monitor (51 × 25.5 cm; 1,920 × 1,080 pixels; 60 Hz) with an integrated eye tracker. Participants viewed the screen binocularly and were seated at approximately 60 cm away from the eye tracker without a head/chin rest, meaning that the screen subtended at a visual angle of 46° × 24°. The eye tracker allows for moderate head movements of up to 37 cm (width) × 17 cm (height) when operating at the distance of 65 cm. Participants were required only to look at the screen during test trials (no explicit response). A keyboard was used between trials to indicate readiness to continue. Changes in viewing distance were monitored using the eye tracker and the size of the artificial VF loss was dynamically adjusted accordingly. (b) The stimuli consisted of two video clips and three static images (only the images shown here), the latter of which were displayed for 60 seconds each. The two videos were presented for their full duration of 301 seconds and 307 seconds, respectively. The resolution of video and the images were presented at a resolution of 1,280 × 720 pixels and the frame rate of the two videos were 30 frames/second. All stimuli were displayed in a full-screen mode (resolution of 1,920 × 1,080 pixels). Stimulus order was randomised between participants. The first image (beach scene) was downloaded from http://genchi.info/beach-wallpaper-1920×1080 in March 2019.