Abstract:
Electrical retinal implants aim to restore some artificial vision to patients, e.g.,
suffering from Retinitis Pigmentosa (RP), by electrically stimulating (e-stim) the
remaining retinal network. Although implanted patients have demonstrated
improvements in daily life activities, the implants face significant limitations in
both temporal and spatial resolution. Over the last decade, optimizing stim
paradigms has been the subject of several studies. However, many underlying
mechanisms of retinal e-stim still remain unclear. Therefore, this project seeks
to understand the effects of e-stim on the retina and to identify optimized stim
strategies. By systematically investigating different stim paradigms in an
appropriate RP mouse model, profound new insights were found.
A newly methodical protocol was established. Degenerated mouse retinal
explants were stimulated subretinally, and the evoked ganglion cell (GC)
responses were recorded using Ca2+-imaging and MEA. This allowed a precise
spatiotemporal analysis of the stim dependent GC activity and the investigation
of the role of the neuromodulator Ca2+ in shaping GC responses. First, applying
e-stim using the advanced chip layout of the RetinaSensor (dense electrode
array; macro-electrodes are replaced by 6 circular arranged micro-electrodes)
and activating the electrodes biomimetically could confine the GC activation
spread to ~60 µm. Regarding optimized stim parameters, the systematic
investigation of e-stim parameters showed the advantage of a biphasic
cathodic-first pulse with the target- (stimulating) and counter- (opposite polarity
as target) electrode being nearby in evoking GC responses. The significance of
the target and counter-electrode positioning for future implants was further
validated, as it was demonstrated that the cathode (negative electrode) elicited
stronger GC responses compared to the anode (positive electrode). Moreover,
applying different stim frequencies revealed a correlation between frequency,
response strength and stim depth. Combined Ca2+-imaging with MEA spike
recordings revealed that high-frequent stim (>5 Hz) clamped the intracellular
Ca2+ electrogenically at elevated levels and lead to an outage of action
potentials throughout sustained e-stim. Finally, the disintegration of the
conventional 1 ms pulse into several 100 µs pulses lead to an increase in
response variation within a voltage range of 0.4 V compared to e-implants,
yielding higher variety in translating light intensities into e-stim. Overall, the stim
strategies elaborated in this thesis combined with advanced implant design like
RetinaSensor with biomimetic electrode activation will guide future studies in
enhancing retinal responses spatiotemporally and be the next step to enhanced
e-mediated artificial vision.
CalciumImaging