Evaluating Eye Tracking on a Consumer-Grade Device for Research and Healthcare Applications

Background & Aim: Eye movements provide valuable insights into attention, perception, and decision-making processes. They also help identify clinical conditions affecting brain function. However, eye movement research has been limited by the cost and scalability of infrared eye trackers. This thesis evaluates the suitability of eye tracking on consumer-grade devices, specifically an iPad, for potential applications in research and healthcare. Method: We developed an objective test of functional colour vision using reflexive eye movements. This test served as a case study to assess the use of consumer-grade eye tracking in screening/diagnostic applications. This test was validated with 34 participants, using an infrared eye tracker (Chapter 3). The same test was repeated with 33 of the original participants using an iPad and its built-in eye tracking software (Chapter 4). In Chapter 5, we used both systems to estimate the gaze location of 28 observers performing various tasks, such as reading and following a stimulus with their eyes. Results: The infrared eye tracker-based test accurately determined participants’ colour vision status with a sensitivity and specificity of over 90%, and also assessed the severity of deficiency (Chapter 3). The iPad-based test achieved the same classification accuracy but was less effective in determining severity (Chapter 4). When estimating fixation, the iPad’s gaze position estimates were less accurate compared to the infrared eye tracker. However, for tasks that elicited saccades greater than 5 degrees, we observed moderate correlations in the estimated saccade count, saccade velocity and saccade amplitude measured across both devices. However, we noted a complete failure of tracking on the iPad in 5-20% of observers, depending on the task (Chapter 5). Conclusion: Our colour vision test can objectively assess red-green colour vision deficiency and can be administered on an iPad (Chapter 3,4). However, the iPad has limitations in reliably measuring eye movements for tasks ...