J Korean Med Sci.  2017 Jun;32(6):900-907. 10.3346/jkms.2017.32.6.900.

Amplitude Modulation-based Electrical Stimulation for Encoding Multipixel Spatiotemporal Visual Information in Retinal Neural Activities

• Ryu SB
• Choi JW
• Ahn KN
• Goo YS
• Kim KH

• KMID: 2377718
• DOI: http://doi.org/10.3346/jkms.2017.32.6.900

Abstract

Retinal implants have been developed as a promising way to restore partial vision for the blind. The observation and analysis of neural activities can offer valuable insights for successful prosthetic electrical stimulation. Retinal ganglion cell (RGC) activities have been investigated to provide knowledge on the requirements for electrical stimulation, such as threshold current and the effect of stimulation waveforms. To develop a detailed "˜stimulation strategy' for faithful delivery of spatiotemporal visual information to the brain, it is essential to examine both the temporal and spatial characteristics of RGC responses, whereas previous studies were mainly focused on one or the other. In this study, we investigate whether the spatiotemporal visual information can be decoded from the RGC network activity evoked by patterned electrical stimulation. Along with a thorough characterization of spatial spreading of stimulation current and temporal information encoding, we demonstrated that multipixel spatiotemporal visual information can be accurately decoded from the population activities of RGCs stimulated by amplitude-modulated pulse trains. We also found that the details of stimulation, such as pulse amplitude range and pulse rate, were crucial for accurate decoding. Overall, the results suggest that useful visual function may be restored by amplitude modulation-based retinal stimulation.

Keyword

Degenerated Retina; Reginal Ganglion Cells; Microelectrode Array; Electrical Stimulation; Spike Train Decoding; Retinal Implant

MeSH Terms

Brain
Electric Stimulation*
Heart Rate
Retinal Ganglion Cells
Retinaldehyde*
Retinaldehyde

Figure

• 

  Fig. 1 Encoding and decoding of visual information. Pulse trains of which amplitudes were modulated according to the brightness of 2 (or 4) pixels were delivered to RGCs via the stimulation electrode on the MEA. The visual information was reconstructed (i.e., decoded) from the evoked RGC activities. Then, the “goodness-of-fit” between the original and decoded pulse amplitude time-series was calculated to evaluate the effectiveness of the stimulation. MEA = microelectrode array, RGC = retinal ganglion cell.

• 

  Fig. 2 Temporal patterns of the neural activities of RGCs. Rhythmic bursting patterns of (A) spontaneous activity and (B) electrically evoked response (pulse amplitude: 20 μA, pulse rate: 1 Hz). (C) PSTH derived from (B). As the arrows indicate, both spontaneous and electrically evoked activities show rhythms of bursts at similar frequencies. (D) ISIH of spontaneous spikes derived from (A). (E) Modulation of response strength by pulse amplitude. The response strength of the RGCs was measured by counting the number of poststimulus spikes within approximately 100 ms of the stimulus onset, which corresponds to the first peak in the PSTH shown in Fig. 2C. RGC = retinal ganglion cell, PSTH = post-stimulus time histogram, ISIH = interspike interval histogram.

• 

  Fig. 3 Dependence of evoked RGC activities on the distance to the stimulation site. (A) Typical temporal patterns of evoked RGC activities. Stimulation pulse trains were applied independently to one of the 2 stimulation electrodes (pulse amplitude: 20 μA). (B) Response strength as a function of the distance between the stimulation and recording electrodes. (C) Spatial profile of the well-modulated RGCs. The number of well-modulated RGCs displayed at the location of the recording electrodes. (D) The percentage of the electrodes with well-modulated RGCs as a function of the distance between the stimulation and recording electrodes. The percentages were calculated by the ratio between the number of all the electrodes and the number of electrodes where the well-modulated RGCs were observed. This result was obtained from 19 retinal patches. RGC = retinal ganglion cell.

• 

  Fig. 4 Examples of original and decoded pulse amplitude time-series, obtained from 2 pixels of a natural scene. (A) Pulse amplitude range: 1–20 μA, pulse duration: 300 μs, pulse rate: 8 Hz. (B) Pulse amplitude range: 1–10 μA, pulse duration: 300 μs, pulse rate: 8 Hz. The location of stimulation electrode of the MEA is described in each panel (open circle: stimulation electrode, closed circle: ground electrode). (C, D) Decoding accuracy vs. simulation parameters. Comparison of the decoding accuracies of (C) 2 different pulse amplitude ranges (obtained from 17 retinal patches, pulse rate: 8 Hz, SVM) and (D) 2 different pulse rates (obtained from 6 retinal patches, pulse amplitude range: 1–20 μA, SVM). The decoding accuracy significantly changed in both cases (t-test, P < 0.001**). MEA = microelectrode array, SVM = support vector machine.

• 

  Fig. 5 Examples of original and decoded pulse amplitude time-series, obtained from 4 pixels of a natural scene. (A) Pulse amplitude range: 1–20 μA, pulse duration: 400 μs, pulse rate: 8 Hz. Pulse amplitude time-series applied to 4 stimulating electrodes could also be successfully decoded from a group of RGCs, just as it was with 2-channel stimulation. (B) Decoding accuracy vs. pulse amplitude ranges. Four-pixel decoding accuracy was significantly different at 2 different pulse amplitude ranges (t-test, P < 0.001**, obtained from 7 retinal patches, pulse rate: 8 Hz, SVM). RGC = retinal ganglion cell, SVM = support vector machine.

Reference

1. Ryu SB, Ye JH, Goo YS, Kim CH, Kim KH. Temporal response properties of retinal ganglion cells in rd1 mice evoked by amplitude-modulated electrical pulse trains. Invest Ophthalmol Vis Sci. 2010; 51:6762–6769.

2. Merabet LB, Rizzo JF, Amedi A, Somers DC, Pascual-Leone A. What blindness can tell us about seeing again: merging neuroplasticity and neuroprostheses. Nat Rev Neurosci. 2005; 6:71–77.

3. Zrenner E, Bartz-Schmidt KU, Benav H, Besch D, Bruckmann A, Gabel VP, Gekeler F, Greppmaier U, Harscher A, Kibbel S, et al. Subretinal electronic chips allow blind patients to read letters and combine them to words. Proc Biol Sci. 2011; 278:1489–1497.

4. Barry MP, Dagnelie G, Argus II. Study Group. Use of the Argus II retinal prosthesis to improve visual guidance of fine hand movements. Invest Ophthalmol Vis Sci. 2012; 53:5095–5101.

5. Pérez Fornos A, Sommerhalder J, da Cruz L, Sahel JA, Mohand-Said S, Hafezi F, Pelizzone M. Temporal properties of visual perception on electrical stimulation of the retina. Invest Ophthalmol Vis Sci. 2012; 53:2720–2731.

6. Ryu SB, Ye JH, Goo YS, Kim CH, Kim KH. Decoding of retinal ganglion cell spike trains evoked by temporally patterned electrical stimulation. Brain Res. 2010; 1348:71–83.

7. Ryu SB, Ye JH, Goo YS, Kim CH, Kim KH. Decoding of temporal visual information from electrically evoked retinal ganglion cell activities in photoreceptor-degenerated retinas. Invest Ophthalmol Vis Sci. 2011; 52:6271–6278.

8. Cottaris NP, Elfar SD. Assessing the efficacy of visual prostheses by decoding ms-LFPs: application to retinal implants. J Neural Eng. 2009; 6:026007.

9. Jensen RJ, Rizzo JF 3rd. Activation of ganglion cells in wild-type and rd1 mouse retinas with monophasic and biphasic current pulses. J Neural Eng. 2009; 6:035004.

10. Fried SI, Hsueh HA, Werblin FS. A method for generating precise temporal patterns of retinal spiking using prosthetic stimulation. J Neurophysiol. 2006; 95:970–978.

11. Stett A, Barth W, Weiss S, Haemmerle H, Zrenner E. Electrical multisite stimulation of the isolated chicken retina. Vision Res. 2000; 40:1785–1795.

12. Lewicki MS. A review of methods for spike sorting: the detection and classification of neural action potentials. Network. 1998; 9:R53–78.

13. Bialek W, Rieke F, de Ruyter van Steveninck RR, Warland D. Reading a neural code. Science. 1991; 252:1854–1857.

14. Warland DK, Reinagel P, Meister M. Decoding visual information from a population of retinal ganglion cells. J Neurophysiol. 1997; 78:2336–2350.

15. Hoegaerts L, Suykens JA, Vandewalle J, De Moor B. Subset based least squares subspace regression in RKHS. Neurocomputing. 2005; 63:293–323.

16. Jensen RJ, Rizzo JF 3rd. Activation of retinal ganglion cells in wild-type and rd1 mice through electrical stimulation of the retinal neural network. Vision Res. 2008; 48:1562–1568.

17. Ryu SB, Ye JH, Lee JS, Goo YS, Kim CH, Kim KH. Electrically-evoked neural activities of rd1 mice retinal ganglion cells by repetitive pulse stimulation. Korean J Physiol Pharmacol. 2009; 13:443–448.

18. Greenwald SH, Horsager A, Humayun MS, Greenberg RJ, McMahon MJ, Fine I. Brightness as a function of current amplitude in human retinal electrical stimulation. Invest Ophthalmol Vis Sci. 2009; 50:5017–5025.

19. Stasheff SF. Emergence of sustained spontaneous hyperactivity and temporary preservation of OFF responses in ganglion cells of the retinal degeneration (rd1) mouse. J Neurophysiol. 2008; 99:1408–1421.

20. Eckmiller R, Neumann D, Baruth O. Tunable retina encoders for retina implants: why and how. J Neural Eng. 2005; 2:S91–104.

21. Freeman DK, Rizzo JF 3rd, Fried SI. Encoding visual information in retinal ganglion cells with prosthetic stimulation. J Neural Eng. 2011; 8:035005.

22. Drasdo N, Fowler CW. Non-linear projection of the retinal image in a wide-angle schematic eye. Br J Ophthalmol. 1974; 58:709–714.

23. Legge GE, Ahn SJ, Klitz TS, Luebker A. Psychophysics of reading--XVI. The visual span in normal and low vision. Vision Res. 1997; 37:1999–2010.