Abstract

The cognitive capacities of nervous systems have their evolutionary origin in ancient, pre-neural systems for computation. Long before an organism must make decisions to guide behavior in space, its embryonic tissues must navigate morphospace to create a complex anatomy. Indeed, during regeneration, cellular collective must coordinate their activity to rebuild a specific organ or appendage and then stop when the correct pattern is complete. Our lab has shown that one important component of morphological computation is bioelectrical: cell groups drive spatio temporal patterns of resting potentials that process morphogenetic information and regulate gene expression and cell behavior. In the first half of this talk, we will describe developmental bioelectricity and show examples in which we re-write the bioelectric pattern memories that guide growth and form. Understanding the origin and computational capacities of these non-neural circuits is critical for evolutionary developmental biology and the design of new biomedical therapies. Therefore, we have begun efforts to model the biophysics and information processing in bioelectrical networks. In the second half of the talk, we examine and analyze a minimal bioelectric network model (BEN). We show that BEN networks can be trained to function as logic gates (the fundamental units of a digital computer), pattern detectors, and pattern regulators. We show examples of trained BEN networks in each case and phenomenologically map them to biological examples. We further show how decisions are made in these networks using dynamical systems and information theory analysis tools. The results motivate the understanding of non-neuronal phenomena like morphogenesis and embryogenesis from an information-processing point of view, and thus help design novel biological experiments and regenerative medicine strategies.