A.Yu. Alekseev - Ph.D. (Philos.), Associate Professor, Department of Informatization of Culture, Moscow State University of Culture and Arts, Coordinator of research programs of the Scientific Council of RAS on the methodology of artificial intelligence (Moscow, Russia). E-mail: аа65@list.ru

The biomedical radioelectronics industry includes development neurointerfaces «brain - computer», «neuronet», «artificial body». For sociocultural proportionality of this industry, for example, for compliance to the bioethical imperatives, need to explicitly prescribe the conceptual links between cognitive phenomena of vital content and engineering, in particular, radioelectronic implementations. The modern philosophy of artificial intelligence (PhAI) sets broad community examine for the links in the format relations «cognition/computation» [1]. The Cognition is personally experienced phenomenal properties or events, which discussed in unity with the ways of their rational description and engineering implementation. Today formed two engineering-based, non-metaphysical paradigm: representationalism and connectionism. The Representative (sumbolical) approach identifies «cognition/computation» with semiotic systems. In contrast, the Connectionism (C.) identifies the relationship with connections of heterarchical network organizations as elements of an arbitrary nature. In today's C. allocated areas: 1) radical C., wich absolutely ignores the representations, 2) moderate (subsimbolic) C.: the connection is an infrastructure of representations, 3) critical C., divided into: a) philosophical C., for example, when discussing the sociocultural perspectives of the «neuronet»; b) methodological C.: for example, in discussions about definitions of neurons and their networks, and 4) comprehensive C.: studying cognition in the integral, connectionist-representational format. Methodology C. carefully designed in anthologies of neurophilosophy (P. Churchland, 1990s) and cognitive philosophy (M. Boden, 2010). But this metodologes of C. have too abstract for the study of biomedical engineering. They are trying to include C. into functionalism - the dominant paradigm of the PhAI. But all modern versions of functionalism essentially reducible to the original - the Turing machine functionalism (Hilary Putnam, 1960). But Turing machine (in versions 1936 and 1950) ? is an ideal representative design. The Turing machine, it in fact, fundamentally, is not able to be the foundation of connectionism. For this reason, for example, the traditional theory of neurocomputing model «formal neural networks» do not correspond to C., as they are not only far from the natural-scientific concepts of neurons and their networks, but also bring harm due to methodological terminological confusion . History methodology C. should start from the Korsakov Machine (S.N. Korsakov, 1832), the world's first perfect («clean») connectionist machine, proto-neuroсomputer [2]. This Machine fully implements the idea of a comprehensive C. In the context of moderate C., the Korsakov-s machine represented connectionist component of a machine of Turing-Babbage-Neumann (1842-1950), as it provides a method for forming heterarchical symbols on the tape of modified Turing machine. The modification consists in the representation of the tape as a stack of punched cards, on which moves the tabulator (the analogue of Turing-s recorder/writer). Needles tabulator, trapped in a hole punched, specify of system connections. In the context of an comprehensive C., the Korsakov machine advisable to combine with the Sechenov machine (Sechenov, 1863). Principles of operation of this machine completely cover the idea of machine functionalism, and, in addition, specify the evolutionary-epistemological mechanisms that seem promising conceptual basis of bioengineering. The Korsakov machine this is the first biomedical computer. The concept machine is clearly spelled out the connection between cognition (complex ideas) and computation (connectionist mechanism functioning). The Korsakov machine successfully used not only medical, in particular, pharmacology problems. It has been applied (in the 1830s!) to study the socio-cultural, for example, the demographic problem. The modern biomedical radioelectronics industry should reach such a level of universality caused by common methodology connectionism.