David Zipser
David Zipser did research for many years in molecular biology before becoming a cognitive scientist.



David Zipser
UCSD

Cognitive Scientist

I am retired, but still doing research. Recently I started collaborating with Howard Poizner, who works on Motor Control Deficits in Parkinson's Disease. My interest in theory and modeling complements his experimental approach in a  constructive way.

For example, the work that Elizabeth Torres and I have done has led to a mathematical model that accounts for many aspects of normal 3D reaching and grasping movements of a seven degree of freedom arm model.  Since some of these movements are disrupted in PD it will be informative to see what components of the model have to be altered to account for the behavior observed in PD. In particular it would be interesting to see if the normal pattern of co-articulation of translation and rotation is disrupted in PD. The model accounts for this co-articulation by combining rotation and translation information in a single cost function which is minimized. If it were disrupted this might be accounted for by a failure to be able to integrate different sources of sensory information. This model can also be used to investigate many other factors that effect motor behavior.

In addition I am working on a number of demonstration models designed to illustrate the effects of disruption of various basal ganglia functions on motor behavior. One of these models is concerned with the effects of the correlated firing in BG output that accompanies PD. It could be the source of some PD symptoms. I am training a neural network to move a simple arm model to targets specified by an uncorrelated input format. When I add correlation to this input I expect to see errors in the arm movements. It would be interesting to compare these errors with those made in PD. In any event the model will clearly illustrate that correlated inputs to a system that doesn't generally receive them can disrupt movement. I hope to be able to construct a whole series of  models of this sort. The one I already did on the different effects of visual target location and initial arm posture errors on arm paths, although over-simplified, has been suggestive.

One of the most important aspects of our collaboration is our continuing discussions about what role the BG plays on brain function. These have been particularly stimulating and I have become very interested in what the BG really does, something that is not at all obvious.