6 92 6. Gaze-contingent processing improves mobility performance 6.1. Introduction Worldwide, approximately 40 million people are affected by blindness (Bourneet al., 2021). In the past few decades promising progress has been made towards the development of visual prosthetics for the blind that aim to support basic visuallyguided activities of daily living, increasing the users’ autonomy and overall quality of life (Fernández et al., 2020; Lewis et al., 2015; Mirochnik & Pezaris, 2019; Niketeghad & Pouratian, 2019; Shepherd et al., 2013). Visual neuroprostheses can elicit a rudimentary form of vision that consist of a pattern of localized light flashes called ’phosphenes’. The aim of visual neuroprosthetics is to create a functional form of phosphene vision that can support basic visually-guided activities of daily living, increasing the users’ autonomy and overall quality of life. The technology of prosthetic vision is still in the early stages of development, and currently the clinical potential is explored for various designs. Even given the rapid recent developments (for instance, see Chen et al., 2020; Fernández et al., 2021) it is clear that the quality of the prosthetic vision will be relatively elementary compared to natural vision. This forms the motivation for an active line of research studying the expected functionality and utility in relation to specific design choices in the development (Chen et al., 2009b; Wang et al., 2022a). A particularly challenging question is how visual prosthetics will deal with the intricate neural interactions between visual processing and oculomotor behavior. In the human brain, eye movements are continuously monitored to preserve visual constancy, mapping a constantly changing stream of retinotopic input onto an internal world-centered representation: a process referred to as ‘spatial updating’ (Burr,2004). In principle, from a neurobiological perspective, there are indications that spatial updating in the brain can be exploited by visual prosthetics in a subpopulation of blind individuals who have intact eyes and residual gaze control. However, this might require active compensation with eye tracking and currently there is scarce literature that systematically compares the relative benefits of such compensation strategies. When considering the biological requirements, it is relevant to note many individuals with late-acquired blindness have intact eyes and some level of gaze control (Kömpf & Piper, 1987; Leigh & Zee, 1980). In many cases of vision loss-related oculomotor difficulties the saccadic accuracy can be improved through training (Ivanov et al., 2016; Kuyk et al., 2010), as the oculomotor system is under constant adaptive visual control (Schneider et al., 2013). Similarly, it is speculated that improved oculomotor control can be regained with the reintroduction of visual feedback through phosphene vision. This suggestion is endorsed by exploratory clinical research. For instance, users of the Alpha IMS retinal photodiode implant (Retina Implant AG, Reutlingen, Germany) are found to display adequate scanning eye movements and object localization with little training and soon after implantation (Hafed et al., 2016). Although these results may not directly translate to other types of prostheses, they suggest that spatial updating in the brain can be successfully exploited to enable natural-like visual scanning behaviour in visual prosthetics. A crucial issue, however, is that visual scanning and spatial updating in the brain work under the assumption that the visual input changes when making eye-movements, but in most contemporary visual prostheses this is not the case. Unlike implants with retinal sensors, most prosthetic designs include a head-mounted camera that samples head-
RkJQdWJsaXNoZXIy MTk4NDMw