Abstract
Purpose::
To investigate the dependence of the excitation area on stimulus pulse parameters and current source and sink in epiretinal prostheses using custom microfabricated multielectrode arrays (MEAs).
Methods::
In order to mimic near-term prostheses, MEAs were designed with four, 200 micron diameter electroplated Pt stimulating electrodes (500 micron center-to-center spacing), and 56, 10 micron diameter Pt recording electrodes. Data was collected from 22 larval tiger salamander retinas (Ambystoma tigrinum). 1 mm2 square pieces of retina were gently pressed RGC side down on the array. Biphasic current pulses were delivered, and RGC responses were recorded using a MultichannelSystems stimulus generator and preamplification board, and corresponding software. Finite-element modeling of the electrode electric field distribution was done using FEMLAB software.
Results::
Cells 50 microns (n = 6) from the stimulating electrode edge had a threshold current of 27.75 +/- 8.66 microamps compared with cells 433 microns (n = 5) away that had a threshold of 68 +/- 23.09 microamps using 400 microsec per phase pulses (p < 0.01). Average rheobase values were 6.5, 27.8, and 46.2 microamps and average chronaxie values were 179, 481, and 600 microseconds for cells 100, 150, and 235 microns away from the stimulating electrode edge, respectively (n = 10). For cells equidistant from adjacent electrodes monopolar stimulation required an average charge threshold of 13.3 +/- 1.7 nC, significantly less than the 29.4 +/- 6.6 nC for dual monopolar stimulation (P < 0.02; n = 5). Bipolar stimulation required 10.0 +/- 3.4 nC. FEMLAB modeling indicated that the increase in threshold in the dual monopolar configuration is due to electrostatic electrode-electrode interactions. Charge density thresholds and response latency and frequency dependence of direct ganglion cell vs. presynaptic excitation are consistent with previous studies (Jensen et al., 2005; Fried et al., 2006; Sekirnjak et al., 2006).
Conclusions::
Cells closer to the electrode edge had lower rheobase and chronaxie values compared with sites further away. Both electrophysiological results and electrostatic finite-element modeling show that electrode-electrode interactions may effect the excitation field at the array-retina interface as has been observed in current cochlear prosthesis implants.
Keywords: age-related macular degeneration