Implementation of Microbe-Based Neurocomputing with Euglena Cells Confined in Micro-Aquariums

摘要

Using real Euglena cells in a micro-aquarium as photoreactive biomateri-als, we demonstrated Euglena-based neurocomputing with two-dimensional optical feedback using the modified Hopfield-Tank algorithm. The blue light intensity required to evoke the photophobic reactions of Euglena cells was experimentally determined, and the empirically derived auto-adjustment of parameters was incorporated in the algorithm. The Euglena-based neurocomputing of 4-city traveling salesman problem possessed two fundamental characteristics: (1) attaining one of the best solutions of the problem and (2) searching for a number of solutions via dynamic transition among the solutions (multi-solution search). The spontaneous reduction in cell number in illuminated areas and the existence of photo-insensitive robust cells are the essential mechanisms responsible for the two characteristics of the Euglena-based neurocomputing.