IT News & Events

News about IT at Indiana University and the world


IU researchers develop brain-body-environment models using Karst

The human brain is a powerful organ, but does it bear the sole responsibility for human behavior? And, how can a 1mm roundworm help answer this question? As Provost Professor Randall Beer explains, “science does not progress by tackling the most complicated examples of a phenomenon of interest,” and humans are nothing if not complicated. Thus, Beer and his colleagues from IU’s Cognitive Science Program, Professor Eduardo Izquierdo, and Postdoctoral Associate Dr. Erick Olivares are querying the relationship between brain, body, and environment through the lens provided by the tiny nematode Caernorhabditis elegans (C. elegans). As one of the best described animals in biology -- its genetics, anatomy, and development have all been well characterized -- C. elegans makes an excellent subject for study. The goal of this project, ultimately, is to construct computer models of the interaction between neural activity, body musculature, and environmental properties underlying the animal’s locomotion. These brain-body-environment models can then be analyzed to determine how behaviors are generated in living organisms.


Dr. Eduardo Izquierdo (left), Dr. Randall Beer (middle), Dr. Erick Olivares (right)

Though C. elegans’ nervous system is well documented, less is known about the detailed properties of individual neurons and synapses. This lack of knowledge translates into many free parameters in the models, and necessitates the use of evolutionary algorithms to find values for these variables that will make the model behave as closely as possible to the actual worm. In effect, Beer and his colleagues use evolutionary algorithms as a semi-automated hypothesis generator to produce possible values for the unknown parameters in their models. The algorithms are computationally intensive, and must be run many times with different random initial conditions in order to achieve a good sample of possible solutions, which inform future experiments. The results of these experiments then feed back into further constraints on future versions of the model.

Given the intensity of these computational models, Professor Beer and his colleagues have turned to IU’s supercomputer Karst, which, he reports, “makes it easy and inexpensive for IU faculty and students to pursue computationally ambitious research.” Computational models will have an essential role to play in developing a full brain-body-environment model of C. elegans in the future. This model will, in turn, mark an important step on the path toward understanding the neural mechanisms underlying human behavior.