The University of Massachusetts Amherst
University of Massachusetts Amherst

Search Google Appliance


Computational Aspects of Complex Pattern Formation

Monday, May 4, 2015 - 3:00pm
Computer Science Building, Room 151 
Mike Levin - Computational Aspects of Complex Pattern Formation: How Endogenous Bioelectric Signaling Among Cells Implements Goal-Driven Remodeling of Shape Memories.

Biological organisms are capable of remarkable feats of pattern regulation. Oak trees or octopi reliably self-assemble from single cells.  Salamanders are able to fully regenerate eyes, whole legs, jaws, and hearts as adults.  Many other experiments show that bodies are able to dynamically remodel from un-anticipated perturbations to a goal state - the target morphology. The current focus of molecular genetics has identified numerous components necessary for this process, but has not revealed the algorithmic steps sufficient to coordinate cell activity towards a large-scale anatomical outcome. Our lab uses numerous model species to study endogenous bioelectrical signaling - voltage-based communication among cells that is instructive for organ identity, size, and topological arrangements. In this talk, I will first illustrate examples of pattern control that require us to move beyond gene expression data towards understanding of top-down, information-centered processes. I will then show methods and sample data illustrating how all cell groups, not just neural networks, process electrical signals to store pattern memories and recall them during remodeling. Finally, I will briefly mention our efforts in artificial intelligence approaches to discovering models of patterning pathways, and speculate on novel approaches for the use of deep concepts from cognitive neuroscience and computer science to help tame patterning. A more complete understanding of the large-scale dynamics of patterning systems will have transformative implications for regenerative medicine, as well as perhaps suggest novel computational architectures to be implemented in synthetic bioengineering.