In vivo electrical stimulation of rabbit retina with a microfabricated array: strategies to maximize responses for prospective assessment of stimulus efficacy and biocompatibility.

PURPOSE: Our primary goal was to assess the effects of varying stimulus parameters on the electrically evoked cortical potentials (EECPs) in rabbits, which we intend to use as one measure of biocompatibility of implanted retinal prosthetic devices. We also sought to exclude contamination of waveforms recorded over the occipital cortex from electrical activity from the retina and the degree of reproducibility of EECP recordings.
METHODS: A concentric bipolar platinum electrode or microfabricated 5x5 electrode array delivered current to the retina of 43 Dutch-belted rabbits while the EECP was recorded from extradural electrodes over the occipital cortex. Electroretinogram (ERG) and visual evoked cortical potential (VECP) recordings were routinely obtained. Verification that occipital cortical recordings were not heavily contaminated by electrical potentials from the retina (i.e. the "validity" of the cortical recordings) was made by recording retinal and brain responses before and after intravitreal injection of tetrodotoxin. Electrical stimulation of the retina was performed with monopolar (with distant return) or bipolar electrode configurations. Cortical responses were computer-averaged over 100-500 stimulations. The effect of variation in stimulus current, charge, duration, frequency, polarity and spatial orientation of stimulating electrodes on cortical responses was studied.
RESULTS: Progressive reduction of responses toward the anterior skull and abolition of posterior recordings by tetrodotoxin indicated that retinal activity did not significantly contaminate EECP recordings. Reproducibility testing revealed that inter-animal variability within the first hour of testing across all animals was not significantly greater than that found during prolonged testing of a single animal. The lowest current that yielded a reproducible EECP with monopolar stimulation was 75 microA (total current through 21 electrodes) using 200 microsec pulses, which yielded a 45 microV cortical response. Strength-duration curves were generally flat for fixed charge stimulation and linear for fixed current stimulation, at least up to a saturation point, which occurred at very high charge. Over 0.5-16 Hz stimulus frequencies, ERGs varied little but evoked potential responses showed a monotonic decline in amplitude at higher frequencies. Large negative-going initial pulses of a biphasic pair yielded the largest cortical amplitudes. EECP amplitudes varied significantly with the orientation of stimulating electrodes on the retina.
CONCLUSIONS: This study provides novel data on the reproducibility of EECP recordings, and insight into stimulation parameters that affect retinal and cortical responses. This information can be used to improve the yield of retinal and evoked potential recordings, which will enhance the prospective assessment of the efficacy of stimulation and health of the stimulated tissues following.