A Computational System-Level Model of Oculomotor Pathways Accounting for the Representation of Eye Biomechanics in the Cerebellar Vermis

Abstract

In this paper a comprehensive system-level computational model of oculomotor pathways is presented. This model shows the necessity of embedding internal models of muscles biomechanics in the cerebellar Vermis to realize fast saccadic eye movements based on predicting the changes in muscles lengths. First, the eye biomechanics are described by nonlinear equations during 'slow' and 'fast' movements. Afterward, by analyzing these equations, a computational model, is deduced. Furthermore, each part of this model is interpreted as a possible function of an element in the oculomotor pathways based on physiological and anatomical pieces of evidence. In this model, two internal feedback loops compensate two types of error: 1- error between desired and estimated values of eye position, calculated by Superior Colliculus, and 2- error between desired and estimated torque, calculated by Cerebellar pathways. Simulations of this circuit produce signals similar to the actual neuronal activities in the corresponding sites of the oculomotor pathways during saccades. Effects of bilateral lesions of Fastigial nuclei, Vermis, Prepositus Hypoglossi, the stimulation of Omni-Pause Neuron and Superior Colliculus are studied. Furthermore, the model ability in performing smooth pursuit eye movements is investigated. Finally, the 'main sequence' is reproduced. This model is the first one to derive both the cerebellar function and the bilateral connectivity of the oculomotor pathways from calculations based on physical hypotheses. The proposed model is useful to better understand computational functions of different parts of the oculomotor pathways, and also using in robotics application for controlling fast movement inspired by the brain.