ABSTRACT
The analysis of biological signals involves the extraction of information contained in such features as waveforms and background from several channels of data. Certain events acquire meaning only when information from several channels is confronted; several channels of data therefore must be analyzed concurrently. This type of problem was first faced in a laboratory in which an electroencephalogram (EEG) was being studied. The signal of an EEG is intrinsically multichannelled, and it generates large amounts of data. A single-channel, single-object analysis called microanalysis is performed in the first step. This analysis is primarily self-contained, and It can achieve a prescreening of events to be supplied to the second step. It is in this first stage that real-time performance is most necessary. A multiple microprocessor system called the Hierarchical Instrument for Distributed Real-Time Analysis (HIDRA) maps this strategy of analysis. The HIDRA system is based on two buses that support a number of peripheral processors.
