Purchase this article with an account.
H. Christiaan Stronks, Gislin Dagnelie, Michael P. Barry, Argus II Study Group; Electrically Evoked Electroretinograms of ArgusTM II Retinal Prosthesis Wearers. Invest. Ophthalmol. Vis. Sci. 2011;52(14):4969.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To characterize electrically evoked retinal activity in retinal prosthesis wearers.
Four ArgusTM II retinal prosthesis users participated in this study. Electroretinogram (ERG) recordings were performed with a standard ERG setup (Espion e2, Diagnosys LLC) using a Burian-Allen contact lens electrode on the implanted eye. ERG recordings were averaged using up to 250 sweeps per recording. Electric stimuli consisted of current-balanced biphasic pulses. Several parameters were investigated, including electric current level, polarity (cathodic- versus anodic-first pulses) and the number of stimulated electrodes. Raw ERG signals were decontaminated off-line by applying stationary wavelet transformation (SWT) using the "symlet 5" wavelet (Brychta et al., 2007, IEEE Trans. Biomed. Eng., 54, 82-93).
ERG recordings were contaminated by large electric artifacts arising from the transmission link between the external (transmitter) and the epi-scleral (receiver) coils. Signal-to-noise ratios (SNR) could be as low as -20 dB. SWT effectively reduced the artifact, and for two of the tested subjects sufficiently decontaminated ERG responses were obtained. ERG waveforms typically consisted of negative deflections with a peak latency of 200 to 400 ms. ERG responses increased at higher current levels up to amplitudes of 30 µV. Cathodic-first pulses typically resulted in greater amplitudes. Disabling 50% of the stimulation electrodes with the highest subjective thresholds, under equal total charge injection, did not substantially affect the ERG. ERG amplitudes tended to decrease over time during recording, which corresponded to the subjective fading of the phosphenes reported by most subjects. Decontamination of the ERG was not successful in two subjects. One subject had nystagmus causing large movement artifacts in the ERG. The other subject’s ERG showed large, undefined, low-frequency sinusoidal artifacts that could not be sufficiently eliminated by wavelet transformation.
Electrically evoked ERGs in retinal prosthesis wearers can be obtained using conventional ERG recording systems. ERG signals with a SNR of -20 dB can be successfully decontaminated using SWT. ERGs were probably the result of retinal neural activity, as opposed to e.g., eye movements, because ERG amplitudes (1) increased with current level, (2) depended on pulse phase and (3) tended to decrease over time.
This PDF is available to Subscribers Only