Neural network models of chemotaxis in C. elegans
T. C. Ferrée, B. A. Marcotte, J. T. Pierce, and S. R. Lockery
Institute of Neuroscience, University of Oregon, Eugene, OR 97403.
C. elegans provides an excellent opportunity to study the cellular basis of neural computations which underly behavior. This is mainly because the morphology of every cell and the location of most electrical and chemical synapses are known precisely. We focus on chemotaxis, the ability to move up (or down) a gradient of chemical attractants (or repellants). Laser ablations of identified neurons in C. elegans have shown which cells are likely to be important for chemotaxis (Bargmann, unpublished), making an anatomically correct model of the entire chemotaxis circuit an attainable goal. Preliminary electrophysiological recordings indicate that ring neurons in C. elegans do not fire classical all-or-none action potentials, but appear to rely mainly on graded signal propagation (Lockery and Goodman, unpublished). To understand how a small network of such neurons might control chemotaxis, we constructed a mathematical model of the body and chemotaxis network of C. elegans. For simplicity, the model used a small number (6-12) of idealized, graded potential neurons, in which any neuron could be connected to any other neuron. An optimization procedure was used to determine sets of synaptic strengths which produce chemotaxis. Many such sets were found. This suggests that the biological network may be one of many different solutions for chemotaxis in graded neural networks. Like real worms, some of the simulated worms produced intricate, seemingly random tracks, despite the fact that the model included no random processes. This suggests that the complexity of real worm tracks could be due, in some cases, to the dynamical properties of graded potential networks alone. In the future, we will use such models to identify the functional role of anatomical and physiological properties of neurons in the biological network. These efforts should help us understand how genetic regulation of neuronal properties alters the function of neural networks in C. elegans.
Supported by NIMH MH11373, NIMH MH51383, NSF IBN 9458102, ONR N00014-94-1-0642, the Sloan Foundation, and the Searle Scholars Program.