Exploiting correlated discriminant features in time frequency and space for characterization and robust classification of image RSVP events with EEG data

In this paper, the problem of automatic characterizing and detecting target images in an image rapid serial visual presentation (RSVP) task based on EEG data is considered. A novel method that aims at identifying event-related potentials (ERPs) in time-frequency is proposed, and a robust classifier with feature clustering is developed to better utilize the correlated ERP features. The method is applied to EEG recordings of a RSVP experiment with multiple sessions and subjects. The results show that, the target image events are mainly characterized by 3 distinct patterns in the time-frequency domain, i.e., a theta band (4.3 Hz) power boost 300-700 ms after the target image onset, an alpha band (12 Hz) power repression 500-1000 ms after the stimulus onset, and a delta band (2 Hz) power boost after 500 ms. The discriminate time-frequency features are mostly power boost and relatively consistent among multiple sessions and subjects. These features are visualized for later analysis. For classification of target and non-target images, our LDA classifier was based on the uncorrelated features, which was constructed from original correlated features using clustering method. With feature clustering, the performance (area under ROC) was improved from 0.85 to 0.89 for within-session tests, and from 0.76 to 0.84 for cross-subject tests. Meanwhile, the constructed uncorrelated features were shown more robust than the original discriminant features, and corresponding to a local region in time-frequency.