Spatio-temporal phenomena in distributed neuron-like media have been studied in IAP RAS for a number of years. Neuron-like media are distributed systems or neural network architectures consisting of active elements with several stationary (or quasi-stationary) states with nonlocal spatial coupling between the nonequilibrium elements. Basic models of neuron-like systems describing dynamics of homogeneous systems as well as hierarchical neuron-like architectures used for development of complex image recognition systems are studied. Dynamics of spatio-temporal (autowave) structures is investigated. The obtained solutions are used for interpreting dynamic modes of normal and pathologic transformation of sensor signals in physiological experiments (fig. 1). Algorithms of signal processing belong to the class of parallel algorithms of information processing (fig. 2). Variants of decision-making systems with adaptive algorithms (fig. 3) have been elaborated.