Phosphene evaluation in a visual prosthesis with artificial neural networks

The electrical stimulation of the optic nerve is investigated, as an approach to the development of a microsystems-based visual prosthesis. Non-linear prediction models, i.e. artificial neural networks, for phosphene localisation are used and show promising results. The prediction error is reduced up to 20% in comparison with linear statistical models and the error standard deviation is restrained such that the prediction accuracy is similar to the measurements accuracy. In addition, it is shown that linear methods do not suit this kind of problems