6.4. Discussion 6 105 versus 9% and 14%, respectively, ranking the gaze-contingent or the gaze-ignored trial as the least comfortable. In terms of subjective functionality ("How well can you navigate with this condition?") the gaze-contingent vision trial was ranked the highest in 48% of the responses, versus 14% for the gaze-locked condition and 38% for the gaze-ignored condition. In 27% of the responses, the participant indicated to experience motion sickness in the trials with gaze-locked vision, versus 6% for both the gaze-contingent and gaze-ignored vision. When asked about specific strategies, 57% of the participants reported that they used exaggerated head movements, 40% of the participants reported that they estimated depth by moving sideways and 61% of the subjects indicated that they estimated the object boundaries by scanning with the center-most phosphenes. After completing Experiment 2, 37% of the participants reported that they had felt disoriented in the gaze-locked study condition, versus 11% in the gaze-contingent condition and 0% in the gaze-ignored condition. 11% of the subjects experienced nausea in the gaze-locked condition, 5% in the gaze-contingent condition and 0% in the gaze-ignored condition. 10% of the subjects report that they experience motion sickness regularly in everyday life. 68% of the subjects reported that they purposely tried to keep their eye still in some of the trials. 47% percent of the subjects reported that despite trying they found it difficult to suppress their eye-movements. 74% of the subjects reported that their eye-movement strategy depended on the eye-tracking condition. When asked about other specific strategies throughout the second experiment, 53% of the subjects indicated that they used exaggerated or frequent head movements, 21% of the subjects indicated that they estimated depth by moving sideways, 47% of the subjects indicated that they used the center-most phosphenes to estimate object boundaries. 6.4. Discussion In this simulated prosthetic vision study, we evaluated the functional benefits of gazecontingent image processing with eye-tracking in several mobility and orientation tasks. Across both experiments, we found, in line with our hypothesis, a reduced number of collisions, lower trial durations and higher subjective ratings for the gaze-contingent study condition compared to the gaze-locked condition. These results suggest that the inclusion of a compensatory eye-tracking system in head steered visual prostheses may improve the functional quality of head-steered visual prostheses in mobility and orientation. Furthermore, in the second experiment, the gaze-ignored control condition generally outperformed both other simulation conditions in all tasks. We speculate that this may be partly owed to adverse effects of an imperfect eye-tracking system. In addition, our findings point to the relevance of considering the effects of gaze in phosphene simulation studies, as neglecting gaze can produce overoptimistic outcomes. 6.4.1. Benefits of gaze contingency with eye tracking There was no significant improvement of the classification accuracy in the scene recognition task with gaze-contingent versus gaze-locked phosphene vision. However the lower number of collisions in the mobility task and the higher speed and subjective ratings across the various tasks in both experiments are in line with the results of previous work (Caspi et al., 2018; Dagnelie et al., 2006; McIntosh et al., 2013; Paraskevoudi & Pezaris, 2021; Sommerhalder et al., 2004; Titchener et al., 2018; Vurro et al., 2014). Several prior
RkJQdWJsaXNoZXIy MTk4NDMw