gms | German Medical Science

Artificial Vision 2015

The International Symposium on Visual Prosthetics

27.11. - 28.11.2015, Aachen

Linear Input Filters in Retinal Prosthetics

Meeting Abstract

  • Daniel L. Rathbun - Center for Integrative Neuroscience, Bernstein Center for Computational Neuroscience, Tübingen, Germany
  • S. Sekhar - Center for Integrative Neuroscience, Bernstein Center for Computational Neuroscience, Tübingen, Germany
  • A. Jalligampala - Center for Integrative Neuroscience, Bernstein Center for Computational Neuroscience, Tübingen, Germany
  • E. Zrenner - Center for Integrative Neuroscience, Bernstein Center for Computational Neuroscience, Tübingen, Germany

Artificial Vision 2015. Aachen, 27.-28.11.2015. Düsseldorf: German Medical Science GMS Publishing House; 2016. Doc15artvis07

doi: 10.3205/15artvis07, urn:nbn:de:0183-15artvis074

Veröffentlicht: 7. März 2016

© 2016 Rathbun et al.
Dieser Artikel ist ein Open-Access-Artikel und steht unter den Lizenzbedingungen der Creative Commons Attribution 4.0 License (Namensnennung). Lizenz-Angaben siehe http://creativecommons.org/licenses/by/4.0/.


Gliederung

Text

Introduction: Retinal prosthetics have made major progress over the past decade despite an incomplete understanding of how electrical stimulation activates the retina. A useful expansion of our understanding would be characterization of the linear input filter applied by the retinal circuit to extracellular electrical stimulation. Accordingly, we investigated whether the spike-triggered averaging of random electrical stimulus-activated retinal ganglion cell spikes would reveal signals indicative of temporal integration.

Methods: We recorded wild-type mouse retinal ganglion cell (RGC) action potentials with a 60-channel microelectrode array. Electrical stimuli consisted of a 25 Hz series of 1 msec duration voltage-controlled, square-wave pulses whose amplitudes were determined by a random draw from a Gaussian distribution with mean of -800 mV and sigma of 280 mV. All of the stimuli immediately preceding a spike were averaged to produce the spike-triggered average (STA) which estimates the linear input filter of the stimulated retinal network presynaptic to the recorded RGC. Direct RGC spikes were excluded with a 10 msec exclusion period after each pulse. The tendency for single bursting responses to misleadingly produce a broad peak of temporal integration was accounted for with a burst-triggered averaging technique.

Results: In 28 RGC spike trains, we found that the majority of cells had broad STA peaks on the order of a few hundred msec – indicating that extracellular, indirect stimulation of RGCs can be integrated across time. We show that such stimuli lie below the single-pulse stimulation threshold.

Conclusion: The use of STAs in characterizing extracellular, indirect RGC stimulation reveals a novel sub-threshold stimulation paradigm for retinal prosthetics.

Acknowledgements: Core funding provided by BMBF (FKZ: 031 A 308; 01GQ1002). Additional support: Deutsche Ophthalmologische Gesellschaft, PRO RETINA, Tistou and Charlotte Kerstan Foundation, and DFG (EXC 307).