Participants
Fifty-three participants were recruited, 19 aged between 65 and 85 years old (y/o, mean=70.80, SE=1.31), and 34 aged between 18 and 24 y/o (mean=19.94, SE=0.26). The recruitment of OAs was cut short due to the COVID-19 pandemic. All YAs were recruited at Bournemouth University, UK. Older adults were recruited either from the Bournemouth Ageing and Dementia Research Centre (ADRC) participant pool or from the Wimborne branch of the University of the Third Age. All participants had normal or corrected to normal vision. Participants were screened for mild cognitive impairment using the Montreal Cognitive Assessment (MoCA, Nasreddine et al., 2005). No participants scored below the cut off score of 23 (Luis et al., 2009). Therefore all recruited participants were included in the final analyses. The study was approved by Bournemouth University’s ethics committee. Informed consent was obtained from participants prior to taking part. Participants took part in exchange for course credits or monetary compensation for their time. This study was performed in accordance with all appropriate institutional and international guidelines and regulations, in line with the principles of the Helsinki Declaration.

Executive function tests
To assess the participants’ EF abilities, participants completed the BADS zoo map test (B. A. Wilson et al., 1996), and the Rogers and Monsell attention shift paradigm (RMA; Rogers & Monsell, 1995). The BADS zoo map test assessed the participants’ spatial planning ability by assessing participants’ ability to plan a route around a
zoo. In the first trial participants were given a map of a zoo and instructed to plan a route around a zoo, starting at the entrance and finishing with a picnic. Along the route participants had to visit specified locations in any order while they followed set rules, such as only using specified paths twice and not visiting unspecified locations. Participants’ planning time and time to complete the task was recorded. In the second trial participants had to plan a route around the same zoo, followed the same rules, and visited the same locations but in a specified order. Again, the participants’ planning time and time to complete the task was recorded.
Participants’ performance was assessed based on visiting the correct locations and points were deducted when participants broke the rules and exceed time limits for planning on the second trial. The scores ranged from zero to four, the higher the score the better participants performed on the test.

The RMA assesses participants’ attentional control by getting participants to switch between two similar tasks. Participants were presented with number letter pairs (e.g., 9E) and depending on the position of the stimulus on the screen they either had to identify whether the number was odd or even or whether the letter was
a vowel or consonant. For the RMA task I extracted the global and local switch costs as done by Rogers and Monsell (1995). The global switch costs refer to the difference in performance between a block where participants perform the same task and a block where participants are switching between tasks. Local switch costs refer to the differences in performance between switch and non-switch trials. I also extracted
the participants’ accuracy and response times on each trial of the RMA. Correct responses were scored as one, incorrect responses as zero. Individual performance was then assessed by averaging accuracy over the entire RMA experiment. These tests have previously been linked to road crossing ability (Dommes et al., 2013; Geraghty et al., 2016) and were designed to assess participants’ spatial planning and attention shifting abilities.

 Walking speed
I measured participant’s walking speed by asking participants to walk along a nine meter corridor while measuring their walking time. Participants were asked to walk at their normal day to day walking pace. This was done three times and an average walking time was then calculated. The walking speed was then calculated by dividing the nine meter distance by this average walking time.

 Apparatus
During the experiment participants’ eye movements were recorded at a sampling rate of 250Hz with the SR-Research EyeLink II, which has an average spatial resolution of < 0.005° . Only the dominant eye was tracked. Stimuli were presented across three Samsung monitors, each with a screen resolution of 1920 by 1080 pixels, an
aspect ratio of 16:9, a width of 88.6cm, and a height of 49.8cm. The left and right screens were placed at 120° to each other. Participants were seated at a distance of 100cm (setup shown in Figure, 4.2a). The screens had a combined horizontal viewing angle of 180° and a vertical viewing angle of 32°. The experiment was coded in Worldviz Vizard 5.0 using Python 2.7 and the PyLink Toolbox extensions (Peirce, 2007). Calibrations for eye fixations were conducted at the beginning of the experiment using a nine-point fixation procedure as implemented in the EyeLink API (see EyeLink Manual). Calibrations were then validated with EyeLink software
and repeated until there was less than 1° of error for every calibration point. Head position and orientation were recorded using the Polhemus Fastrak motion tracking system with a sampling rate of 120Hz.

 Experimental Procedure
Both experiments used a virtual road crossing environment created in 3DS Max and Maya (Figure. 4.2b) which was made to simulate the road crossing scene used in Nicholls et al. (2019) and Chapter 3, without the roundabout. Prior to the start of the experiment participants’ eye movements were calibrated using a custom
calibration procedure across all three screens. This procedure involved presenting circles with a break on the left or right side and a dot in the middle (Figure 4.1) at random locations on all three screens. Participants had to look at the circle and indicate whether the break in the circle was on the right or left hand side using the left and right arrow keys on the keyboard. While participants performed this task their eye movements were recorded. Once this was completed participants eye movements were calibrated on one screen using the Eyelink calibration procedure. At the beginning of the experiment participants were informed that they would be presented with a series of road crossing situations on screen and that they would have to indicate by pressing the spacebar on a keyboard when they could cross the road and hold the key pressed for as long as they thought it was safe to cross. At the start of each experimental block participants were informed on which side the cars
would appear from – left hand side, right hand side, or both sides (Experiment 2 only). Vehicles travelled at two speeds – 249 (slow) or 583 (fast) virtual world units per second. This was equivalent to approximately 30 and 70 km/h respectively. Each trial started with the presentation of a central fixation cross. Once the participants
had fixated on the cross, the virtual environment was presented. Each trial was followed by a black screen with text stating the trial had ended and the participant should press the spacebar to continue. Once the participants pressed the spacebar the next trial would start with the central fixation cross.