A data-driven electroencephalography channel expansion method and system

Abstract

The application discloses a data-driven electroencephalogram channel expansion method and system, acquires a low-channel electroencephalogram signal, divides the low-channel electroencephalogram signal along a time dimension, integrates a plurality of divided block signals along a channel dimension to form a block signal matrix, extracts and fuses features of each block signal in the block signal matrix to obtain a fused feature matrix, expands channels of the fused feature matrix to generate an expanded feature matrix, splices the expanded feature matrix and the fused feature matrix, maps time-domain signals through feedforward operation of a feature matrix obtained after splicing to obtain a high-channel electroencephalogram signal. The method applies a state space model to a channel expansion task of an EEG signal, has stronger time sequence modeling capability and feature extraction capability compared with existing methods, and can better capture the space-time features of the EEG signal.

Images