Abstract:
The physics of the dipole system of a neuron cytoskeleton microtubule is put forward, and a Hamiltonian of the dipole system is constructed. The previously developed microscopic model of the cytoskeleton microtubule dipole system is extended for the case of dipole-dipole bonds when the bonds are not fully ordered and are constrained (exhibit the memory property). Molecular field expressions are derived for a random polarization function and its two moments: mean polarization and rms polarization. An evolutionary equation for a random order parameter is of a relaxation character and describes the pattern recognition process. It is shown that the phase transition nonlinearly transforms (projects) one (nodal) space of higher dimension pattern features to another space of lower dimension attributes (order parameters), this transformation greatly cutting the body of data to be processed.