An on-line multiprocessing interactive computer system for neurophysiological investigations

Abstract

The principal dependencies of neurophysiologists upon the computer are for data collection and analysis, experimental control, and the development of theoretical models. One possible system providing these functions is one that allows several investigators to on-line time-share a moderate sized digital computer capable of performing input, output, and computational functions in a simple interpretive language that is easy to understand and use in a fast decision experimental environment. A community of neurophysiologists in the UCLA Brain Research Institute share such a computer in its data processing laboratory (DPL) by means of remote console stations in the investigators' laboratories connected to the DPL by a direct cabling system. A larger computer facility, available to a larger community of health scientists, is used for batch processing where problems do not need continuous interaction with the investigator for on-line control or analysis, or do need greater computational capability. The two facilities possess compatible I/O formats, thus making some problems soluable by the combination of both computers, and giving other problems the flexibility of either approach. Essentially the DPL is a multiprocessing system with an emphasis on I/O functions and an interpretive system appropriate for neurophysiological investigation and with some unique solutions to resource allocation and system integrity in its temporal-spatial (core) algorithm. The economic advantage of such a system is not argued, nor is CPU economy necessarily maximized with present use, though from the standpoint of I/O devices which are so important for such research, a central facility may possess some advantages. Reliability and demands of time-critical users must be realistically estimated for neurophysiological users for whom the on-line aspect may be with reference to the integrity of their experiments.