Weak Feature Extraction and Strong Noise Suppression for SSVEP-EEG Based on Chaotic Detection Technology

Brain computer interface (BCI) is a novel communication method that does not rely on the normal neural pathway between brain and muscle of human. It can transform mental activities into relevant commands to control external equipment establish direct pathway. Among different paradigms, steady-state visual evoked potential (SSVEP) widely used due its certain periodicity stability control. However, electroencephalogram (EEG) SSVEP extremely weak companied with multi-scale strong noise. Existing algorithms for classification are based principle template matching spatial filtering, which cannot obtain satisfied performance feature extraction under Especially subjects produce response stimuli in EEG representation, i.e., BCI-Illiteracy subject, traditional difficult recognize internal patterns brain. To address this issue, Chaos theory proposed extract SSVEP. The rule applying peculiarity nonlinear dynamics system detect by judging state changes chaotic systems after adding EEG. evaluate validity method, research recruit 32 participate experiment. All divided two groups according preliminary accuracy (mean acc >70% or < 70%) canonical correlation analysis we define above 70% as group A (normal subjects), below B (BCI-Illiteracy). Then, information transmission rate verified using Chaotic theory. Experimental results show all methods our study achieve good while chaos excellent significant improvements than BCI-Illiteracy.