A layered processing model for knowledge-based contextual interpretation of multichannel EEG/EOG signals

A layered processing model designed for a knowledge-based expert system for automated classification of multiple channel sleep EEG/EOG (electroencephalogram/electrooculogram) signals is described. A three-layer model is formulated to process and modify the contextual information present in the signal. Each layer is associated with a different processing time frame. It uses a hierarchical architecture that reflects the human expert´s information and data abstraction process for the visual inspection of the signal. The model incorporates a certainty handling scheme which uses symbolic representation of certainty levels. The certainty-level information does not participate directly in the rule-execution mechanism, but is manipulated separately and provides augmented information to use effectively for the contextual interpretation in the layered model