gms | German Medical Science

Objective: Demonstrate that the electrical responsiveness of the mouse retina can be probed subthreshold, white noise pulse sequences.

Methods: To categorize mouse retinal ganglion cell (RGC) visual responses into ON, OFF, and ON-OFF types, full-field flashes (20 cycles of 2 s ON and 2 s OFF) and full-field temporal Gaussian white noise at 10 Hz were presented. An electrical ‘white noise’ stimulus with 1 ms, voltage-controlled, epiretinal, cathodic pulses at a frequency of 25 Hz was presented with amplitudes randomly drawn from a Gaussian distribution (mean –800 mV, SD 280 mV). Spike-triggered averaging of noise responses yielded the average visual or electrical stimulus sequence that activated each cell.

Results: Spike-triggered averages (STAs) demonstrated the integration of sequences of ‘subthreshold’ voltage pulses. Significant reduction in the STA integration width, when calculated for degenerate (P84, rd10) compared to healthy (C57Bl/6) mouse retina, showed that healthy STAs result primarily from activation of photoreceptors. Such STAs had upward deflections in the cathodic pulse sequence for ON RGCs and downward deflections for OFF cells.

Discussion: White noise stimulation paired with reverse correlation (STA) can be used to estimate the best pulse sequence to elicit an RGC response in the context of ongoing electrical stimulation – as used in retinal prostheses. Such STAs show that he retinal network has a memory for electrical stimulation that allows it to integrate multiple (subthreshold) pulses across time. Different STAs for ON vs. OFF RGCs suggest that selective stimulation of these visual pathways may be possible.

Acknowledgments: BMBF: 031 A 308; Bernstein Centre for Computational Neuroscience, Tübingen BMBF: 01GQ1002; Deutsche Ophthalmologische Gesellschaft; PRO RETINA; Tistou and Charlotte Kerstan Stiftung; Werner Reichardt Centre for Integrative Neuroscience DFG: EXC 307