Training Artificial Neural Network to Classify Correct and Erroneous Interpretations of Visual Stimuli before Behavioral Response

Abstract

We have trained an artificial neural network (ANN) to predict correct and erroneous interpretations of complex visual stimuli, Necker cubes with different levels of ambiguity. For the selected configuration of ANN, the classification accuracy was 88%. These results are prospective for developing the new-generation assistive technologies.