Vergence eye movement with prediction and learning based on dual visual-local feedback

Research results in neurophysiology show the predictive nature of vergence eye movement, vergence eye movement can persistently track a moving target which shifts in distance relative to the head. Few models have attempted to consider prediction of target motion in vergence models. Most models only considered static targets, their input are frozen driving signals. In this paper, a model with estimator and predictor to predict and learn current target vergence is proposed. The internal model is designed to record the learned target dynamics and the selector recognizes the final prediction of the target vergence between predictor and the internal model. Then the final prediction of the target vergence is put to a dual visual-local feedback model of vergence eye movement system proposed by Erkelens. Simulation results are included to suggest the effectiveness of proposed model, especially to the continuous moving targets.