Characterization of Retinal Ganglion Cell Responses to Electrical Stimulation Using White Noise

Abstract

Retinitis pigmentosa and age-related macular degeneration are two leading causes of degenerative blindness. While there is still not a definitive course of treatment for either of these diseases, there is currently the world over, many different treatment strategies being explored. Of these various strategies, one of the most successful has been retinal implants. Retinal implants are microelectrode or photodiode arrays, that are implanted in the eye of a patient, to electrically stimulate the degenerating retina. Clinical trials have shown that many patients implanted with such a device, are able to regain a certain degree of functional vision. However, while the results of these ongoing clinical trials have been promising, there are still many technical challenges that need to be overcome. One of the biggest challenges facing present implants is the inability to preferentially stimulate different retinal pathways. This is because retinal implants use large-amplitude current or voltage pulses. This in turn leads to the indiscriminate activation of multiple classes of retinal ganglion cells (RGCs), and therefore, an overall reduction in the restored visual acuity. To tackle this issue, we decided to explore a novel stimulus paradigm, in which we present to the retina, a stream of smaller-amplitude subthreshold voltage pulses. By then correlating the retinal spikes to the stimuli preceding them, we calculate temporal input filters for various classes of RGCs, using a technique called spike-triggered averaging (STA). In doing this, we found that ON and OFF RGCs have electrical filters, which are very distinct from each other. This finding creates the possibility for the selective activation of the retina through the use of STA-based waveforms. Finally, using statistical models, we verify how well these temporal filters can predict RGC responses to novel electrical stimuli. In a broad sense, our work represents the successful application of systems engineering tools to retinal prosthetics, in an attempt to answer one of the field’s most difficult questions, namely selective stimulation of the retina.