Short Pulse Epiretinal Stimulation Allows Focal Activation of Retinal Ganglion Cells

Epiretinal implants suffer from a lack of spatial resolution, which is greatly influenced by the undesired co-activation of distal cells with their axons passing close to targeted somas. Short current pulses in the range of $50\mu $ s have been shown to preferentially activate somas, but the low specificity may limit practical applications. In this paper, we explored decreasing pulse durations down to $10\mu $ s for achieving focal activation, i.e., a large differentiation between axonal and somatic activation in epiretinal configuration. We determined thresholds for pulses ranging between 10 and $500\mu $ s in retinal ganglion cells of both wild-type and photoreceptor-degenerated mouse retina. Ex-vivo stimulation using biphasic rectangular pulses was performed using a custom-built modified Howland-type current-controlled stimulator and a microelectrode. We demonstrate reliable direct activation of retinal ganglion cells using $10\mu $ s pulses for both somatic and axonal electrode positions. Cells from wild-type and photoreceptor-degenerated retinas exhibited similar thresholds. Axonal thresholds were significantly higher for all pulse durations, with the ratio between axonal and somatic thresholds strongly increasing with decreasing pulse duration (1.32 and 4.39 for pulse durations of 500 and $10\mu $ s, respectively). Computational modeling points to somatic polarization as the underlying mechanism for lower somatic thresholds. Our results demonstrate focal activation with pulses in the range of $10\mu $ s as a potential strategy to avoid the long-standing problem of axonal co-activation in epiretinal implants.