Research Abstracts Online
January 2010 - March 2011
University of Minnesota Twin Cities
PI: John H. Anderson
Neural Network Model for the Vestibulo-Ocular Reflex
These researchers are investigating neural networks that mediate horizontal eye movements that serve to stabilize gaze during head movement (the vestibulo-ocular reflex, VOR). Neural networks that represent distinct populations of physiological neurons (100s to 1000s) will be developed. The aims are: to identify those neurons which might be involved in a disease process causing a specific pattern of abnormal behavior of the VOR; to predict the extent to which the network could compensate for or adapt to the abnormal behavior and which neurons/pathways would be involved; and to test models for motor learning involving sensory-motor transformations and the VOR. The initial work considers horizontal eye movements and semicircular canal inputs (due to angular acceleration of the head). Future modeling will consider the three-dimensional VOR with canal-otolith and visual-vestibular interactions and physiological motor learning and adaptation (as intrinsic properties of the network), including the role of the cerebellum. The dynamic characteristics and spatial-temporal properties of the network inputs and desired output eye movements will be specified, based on experimental and theoretical results of previous studies. The connectivity and input-output characteristics of single neurons will be based on what is known about the neurophysiology of the VOR. The researchers use software available through MSI to develop custom programs for training the neural networks and simulating eye movement responses to different profiles of head movement.
Chris Fuller, Staff