[Zurück]

H. Schobesberger, S. Resatz, F. Rattay:
"Influence of neural geometry and pulse duration for extracellular stimulation of simulated retinal ganglion cells";
Poster: 5th Forum of European Neuroscience, Vienna; 08.07.2006 - 12.07.2006; in: "FENS Abstr.", 3 (2006), S. A072.21.

Focal excitation of retinal ganglion cells is mandatory in order to achieve better resolution in artificial "light" perception in blind people with retinal implants. Cha et al. (1992) estimated that a minimal grid of 25 x 25 stimulation points across the macula would suffice to let a patient freely roam public places again without requirement of other special aid. Focused stimulation however is constrained by the limit to electrode minimization due to the risk of high charge density or heat damage. To investigate location specificity of epiretinal excitation, we designed a Fohlmeister-Coleman-Miller (1990) model of a retinal ganglion cell from Necturus maculosus derived from morphometric data from Velte & Masland (1999) using combined compartmental and finite element methods. The electrode, modelled as a point source in an infinite homogeneous volume conductor, was moved along the retina surface, essentially in the x-y-plane 30 mikrons z-distance above the soma center. On 25 electrode positions with a distance of 50 mikrons each the cathodic threshold was calculated for 50, 100 and 200 microseconds pulse duration (PD). Threshold minimum (0.27mA, 0.14mA, 0.08mA for 50/100/200microseconds PD) was always located with electrode placement directly above the axonal thin segment, maximal threshold, being approximately 16x higher (2.87mA, 1.55mA, 0.86mA) than the lowest, at the most distant segment from neural tissue. These observations fit very well to extracellular recordings from isolated rabbit retinas (Jensen et al. 2003, 2005). Change in threshold current by doubling PD however was greatly influenced by the exact position of the electrode. Certain neural structures required disproportional high effective stimulation currents to increased PD. Simulations show that this phenomenon depended to a large extent on neural geometry (ramifications, curvature or axon initial segment) and not on the exact membrane properties of the underlying structures (dendrite, soma or axon)