5.1. Introduction 5 59 fundamental, clinical and computational scientists working on neuroprosthetic vision. All of the underlying models in the simulator are flexibly and modularly implementend allowing for easy integration with external software (seeFigureS1). Although the simulator, by default, assumes round phosphene percepts that are independently activated, it can be tailored to simulate custom alternative scenarios: the phosphene maps can be adjusted to simulate different shapes (seeFigureS2), and it is possible to simulate arbitrary interactions between electrodes with minor adaptations to our code (seeFigureS3). The modular design of the simulator allows for future extensions to simulate brain stimulation in other regions such as lateral geniculate nucleus (LGN) or higher visual areas (Mirochnik & Pezaris, 2019; Murphey & Maunsell, 2007; Panetsos et al., 2011; Pezaris & Reid, 2007). Figure 5.1: Left: schematic illustration of our simulator pipeline. Our simulator is initialized with electrode locations on a visuotopic map of the visual cortex. Each frame, the simulator takes a set of stimulation parameters that for each electrode specify the amplitude, pulse width, and frequency of electrical stimulation. Based on the electrode locations on the cortical map and the stimulation parameters, the phosphene characteristics are estimated and for each phosphene the effects are rendered on a map of the visual field. Finally, the phosphene renderings are summed to obtain the resulting simulated prosthetic percept. Right: Example renderings after initializing the simulator with four 10×10 electrode arrays (indicated with roman numerals) placed in the right hemisphere (electrode spacing: 0.4 mm, in correspondence with the commonly used ‘Utah array’; Maynard et al., 1997). The output is visualized for 166 ms pulse trains with stimulation amplitudes of 40, 80, and 120µA, a pulse width of 170 ms, and a frequency of 300 Hz. In these example frames, we can observe the effects of cortical magnification, thresholds for activation, current-dependent spread (size) and proportion (brightness) of cortical tissue activation. 5.1.1. Background and related work Cortical prostheses Early attempts by Brindley and Lewin, and Dobelle successfully reported the ability to reliably induce the perception of phosphenes (described as localized round flashes of light) via electrical stimulation of the cortical surface (Brindley & Lewin, 1968; Dobelle & Mladejovsky, 1974). More recent preclinical studies demonstrate promising results concerning the safety and efficacy of long-term stimulation in the primary visual cortex, either via surface electrodes (Beauchamp et al., 2020; Niketeghad et al., 2020)orwith intracortical electrodes (Bak et al., 1990; Fernández et al., 2021; Schmidt et al., 1996). Other studies that performed V1 stimulation in sighted subjects (Bosking et al., 2017a; Winawer & Parvizi, 2016) and non-human primates (Chen et al., 2020; Schiller et al., 2011) have shown similar success. Some milestones include the implantation of over 1000 electrodes in a monkey’s visual cortex (Chen et al., 2020), and the testing of a preliminary
RkJQdWJsaXNoZXIy MTk4NDMw