Abstract
Purpose: :
To study simulations of a novel retina implant with perceptual feedback for intra-retinal signal processing, with oculomotor feedback for simulated miniature eye movements (SM), and with consideration of spontaneous nerve impulses (NI) of the stimulated regions.
Methods: :
A: Simulations were performed with a novel retina encoder (RE-3) with a filter module (FMRef) consisting of an input array of 20 x 20 pixels for presentation of a pattern P1 and 100 spatio-temporal (ST) filters for generation of selective stimulation signals for 100 electrodes at the retinal output, a novel inverter module (IM-3) to mimic parts of the central visual system and to map the FMRef -output onto a simulated percept P2, and a dialog module (DM).B: Pattern P1 was selected as: house, circle-cross, letter π, or star.C: A dialog module (DM) simulated the perceptual feedback from a human user.D: SMs were generated on demand for movements of P1 by one pixel in a given direction to mimic typical eye movements during fixation.E: Spontaneous NI were considered by two options to avoid stimulus pattern disturbances: a) stimulation signals with NI-suppression capability and b) neural feedback from neural tissue to RE-3 via bi-directional electrode arrays for stimulation and recording.
Results: :
(1) FMRef was specified as a regular distribution of three ST filter types to mimic receptive field properties of primate retinal ganglion cells.(2) Both P1 and FM filter array were described as multi-dimensional vector matrices, which had several advantages, including easy changes of pixel numbers or filter numbers as well as FM-output calculation as matrix products.(3) The software for IM-3 was designed to process the FM-output matrices with an efficient algorithm and to invert the partly ambiguous FM mapping with the help of SMs. In most cases, at least one SM was necessary to resolve residual ambiguity and to recover a complete P2 identical to P1.(4) Due to the matrix structure, inversion of the FM-output by IM-3 could be processed in a single matrix run through all coefficients. The Moore Penrose Pseudoinverse was more time-consuming and less perfect.(5) A direct NI-suppression versus neural feedback of NI via recording electrodes will be discussed. Given the neurophysiological evidence of spontaneous NI, the best solution may have to be decided individually.
Conclusions: :
Since ‘Gestalt’ perception in humans requires active vision as a continuous interaction between sensory and oculomotor processes, a combination of neural-, oculomotor, and perceptual feedback may be important to optimize the function of bionic visual prostheses.
Keywords: computational modeling • retinal connections, networks, circuitry • perception