2-DIMENSIONAL NEURAL-NETWORK MODEL OF THE PRIMATE SACCADIC SYSTEM

The objective of this paper is to present a distributed model of the spatiotemporal neural processing that underlies the control of two-dimensional saccadic eye movements in the monkey. In this new model the superior colliculus (SC) is represented by two layers of cells. Simulated visual inputs activate the upper layer that is connected lo the lower (motor) layer with feed forward projections. Extensive lateral interconnections exist in the motor layer The weights assigned to these interconnections are established with a recurrent back propagation algorithm and by training on a set of activity patterns obtained from neurons recorded in the monkey SC. A distributed set of connections to horizontal and vertical brainstem saccadic burst generators is trained to allow the model to make accurate saccades to randomly selected target positions. Finally, the model is able to produce accurate eye movements and realistic neural discharge for saccades evoked with a variety of experimental conditions not included in the training set (for example, averaging and express saccades).