Looking at Images During Personal Space Invasion

Eye Tracking Study Using Mild Deception (Dosso, Anderson, Choi, & Kingstone, in prep)

Inspiration

I initially started volunteering at the Brain, Attention, Reality Lab as a research assistant to help with data collection and video coding for this study.

While observing participants during the study & learning more about the effect of social presence on human behaviour, the concept I found the most interesting was how people thought about the reachable space immediately around them (i.e. peripersonal space). After being asked to help with another study using eye-tracking equipment, I became fascinated with the technology and led me to wonder if there would be any effect of peripersonal space invasion on attention with eye movements, or “overt attention”.

When I brought up the idea to my supervisor & the lab director, they generously agreed to collaborate on the study design, execution, and data analysis. My primary role was in co-designing & writing the methodology section, as well as developing the study protocol & eye-tracking stimuli with SR Research Experiment Builder. In the designing process, I came up with the deception scenario of the ‘misbehaving’ eye-tracker, and pushed for the ultimate decision to go with a 2×2 within-subjects design to ensure validity (although between-subjects was suggested due to convenience). My supervisors then took over the data collection & analysis, and we are now currently collaborating on writing up a preprint for potential publication. Since my main contributions were in the design of the study, the following sample of the project will focus primarily on the methodology & procedures.

Research Question

Do nearby others influence on-screen looking behaviour (i.e. overt attention)?

Participants & Stimuli

Twenty-two participants were all students recruited from University of British Columbia, and consented to participate in the study for course credit. All participants were required to have normal or corrected to normal vision, and passed the exclusion criteria of eye contacts as their vision correction method during the experiment. Everyone indicated they were unaware of the purpose of the experiment, as well as the minimal deception used to manipulate the independent variable of social presence. The study was approved by the Behavioural Research Ethics Board at the University of British Columbia, and conducted according to Tri-Council Policy Statement guidelines.

The experiment was compiled using the SR Research Experiment Builder, a drag-and-drop experiment programming environment integrated with the head-mounted eye movement tracking system used for the experiment, SR Research Eyelink 1000. The eye tracker recorded eye movement data from the left eye, and eyes were positioned at a perpendicular visual angle from the screen. The chin rest was used to secure the head at 50 cm from the screen, and to minimize head movements. Eye position was kept consistent across participants with the chin rest, and set up for the eyes to be leveled at the same height as the screen’s vertical midline.

The stimuli were presented centrally on the computer screen. All participants freely viewed 80 images in 2 separate consecutive blocks (each block done alone or with experimenter on one side of the screen), with each image presented on the screen to freely view for two-seconds. All participants viewed a total of 160 images, consisting of 20 unique coloured fractal images and 20 unique natural outdoor scenes at each of their four perpendicular image orientations (i.e. 0, 90, 180, 270 degrees). For each block (80 two-second free-viewing trials), the order of the 10 unique fractals and 10 unique natural outdoor scenic images at each of their 4 perpendicular orientations were interleaved and randomized for every participant. Each unique image was presented at all four orientations for every participant in the same block. The computer-generated fractal images were sourced from an online fractal database used in another eye-tracking study by by Parkhurst et al. (2002) and Peters et al. (2005) (see Foulsham & Kingstone, 2010). The natural outdoor scenic images were also taken from the same study, and originally selected from the collection of outdoor images available from the Internet. 

forest-272595_1280
Natural outdoor scenic image
Computer generated fractal image

All participants were given verbal instructions during the experiment. Calibration and validation of the participants’ eye position with a 9-dot grid on screen was performed at the beginning of each 80-trial block to ensure stable eye movement tracking. At the beginning of every trial, participants were also asked to indicate with a button when their eyes were fixated on the center of the screen, and the onset of each image began when the eye tracker confirmed stable fixation. The 160 trials were broken up into 2 blocks (80 trials per block), one for each experimental condition (i.e. social presence & alone). Each trial presented the stimuli for 2 seconds while the participants’ eye movements were tracked. 

2x2 Within Subjects Design

The experiment employed a 2×2 within-subjects design, with each participant randomly assigned to one of the four following conditions: Alone-SocialPresence [Left], Alone-SocialPresence [Right], SocialPresence-Alone [Left], SocialPresence-Alone [Right]. The order of the social presence and alone conditions were counterbalanced for all participants, and did not show any order effects.

 In the Alone-SocialPresence [Left] and Alone-SocialPresence [Right] conditions, the participants completed the first block of 80 free-viewing trials alone. To conceal the purpose of the experiment, participants were told during the calibration and validation process (after the first free-viewing block of 80 trials) that there seems to be some technical issues with the eye tracking system. The experimenter tells the participant that they will need to sit by the screen to make sure everything is working properly, and takes a seat at either right or left side of the screen while the participant completes the second block of 80 images. 

 In the SocialPresence-Alone [Left] and SocialPresence-Alone [Right] conditions, the eye tracker “misbehaves” during the first block. After the first 80 two-second free-viewing trials, the experimenter tells the participant that the eye tracker seems to be working properly now, and leaves the participant to complete the second block of the experiment alone.  

Results

  • Do nearby others influence on-screen looking behaviour (i.e. overt attention)? – YES, but only for fractal images
  • Consistent with past research, found expected “pseudoneglect” = human’s natural tendency to shift spatial attention towards the left side (Nuthmann & Matthias, 2014)
  • Could say small effect, but swamped by content of images OR lack of content (i.e. fractals) makes eye movements carry larger social signal ???
  • Need more research to make stronger claims & conclusions!

How does personal space invasion affect task performance?

In this research study proposal, we extend upon the aforementioned personal invasion study to understand the relationship between test performance using VR.