Emergent computation in microtubule model networks

The cytoskeleton, a class of protein polymers including microtubules (MTs), intelligently organizes the interiors of living cells. Interconnected microtubules have been modeled as adaptive networks in which emergent computation occurs. The model comprises parallel networks of MTs variably interconnected by filamentous proteins by means of MT-associated protein 2 (MAP/sub 2/). MAP/sub 2/ can also mediate input into MT from other MTs, by cyclic AMP and/or calcium from membrane events. Optimized MT-MAP/sub 2/ networks form computationally efficient connection topologies, with feedback among topology, internal automata dynamics and automata rules in an MT structural model. Thus MT network models exhibit learning and other emergent effects.<
>