May 2008
Volume 49, Issue 13
Free
ARVO Annual Meeting Abstract  |   May 2008
Implementation of a Retina Implant With Tuning of Temporal Pattern Presentation and Selective Electrode Stimulation
Author Affiliations & Notes
  • S. Borbe
    Computer Science, University of Bonn, Bonn, Germany
  • R. E. Eckmiller
    Computer Science, University of Bonn, Bonn, Germany
  • Footnotes
    Commercial Relationships  S. Borbe, None; R.E. Eckmiller, None.
  • Footnotes
    Support  University of Bonn
Investigative Ophthalmology & Visual Science May 2008, Vol.49, 3018. doi:https://doi.org/
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      S. Borbe, R. E. Eckmiller; Implementation of a Retina Implant With Tuning of Temporal Pattern Presentation and Selective Electrode Stimulation. Invest. Ophthalmol. Vis. Sci. 2008;49(13):3018. doi: https://doi.org/.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose: : To develop novel selective tuning properties for a Retina Encoder (RE) regarding temporal pattern presentation (TPP) and selective electrode stimulation (SES) and to explore their influence on simulated visual perceptions.

Methods: : A: Optical input patterns (P1) were composed of 16 x 16 pixels. A partial pattern module (PM) subdivided P1 into tiles (e.g.: 9) and generated up to 3 partial patterns (Pj) on non-adjacent tiles.B: Each of the 10 x 10 tunable spatio-temporal (ST) filters of the filter module (FM) could be alternately used with specific parameter vectors to simulate the receptive field properties of up to 3 ganglion (G) cells.C: For selected electrodes (Ei), the G-cell selection module (GM) (Eckmiller, Borbe: ARVO, 2008) generated specific stimulation time courses (Si) with stimulus durations (Ts) of 10-20 [ms] within a given time interval (Ti) during the pattern presentation time (Tp). Accordingly, each Ei could selectively stimulate up to 3 neighboring G-cells during Ti.D: The timing coordination module (TM) specified Tp for the consecutive presentation of Pj and the corresponding sequences of Ti with individual Ts. Thus, TM determined the spatial and temporal distribution of electrode stimulation and selective G-cell stimulation.E: The dialog module (DM) used a combination of learning algorithms to incorporate the TPP and SES properties. Specifically, the human subject interacted with the DM to define and gradually change the optimization focus (governed by TM) of the tuning process.F: The function of the inverter module (IM) to mimic the central visual system, was performed based on rectangular inverter tiles (e.g.: 9). Each tile achieved a local inversion by using a combination of classification, decision tree, and simulated miniature eye movements to resolve the ambiguity of the ST filter output signals.

Results: : (1) During dialog-based tuning, spatial and temporal parameters for TPP and for SES could be identified that met the IM requirements.(2) Following successful RE-tuning, the presentation of a pattern P1 as sequence of Pj could elicit a visual percept P2 that appeared similar to P1.(3) Due to selective G-cell stimulation with GM and SES, the resolution of P2 was 3 times higher than expected from the number of electrodes (e.g.: 100).

Conclusions: : A. Selective G-cell stimulation makes it possible to turn the disadvantage of large sized electrodes into the advantage of increased visual resolution.B. The novel approach of selectively tuning the temporal parameters of both pattern and stimulus presentations promises benefits for neural encoding but poses a challenge for perceptual reintegration.

Keywords: pattern vision • retinal connections, networks, circuitry • computational modeling 
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×