Analyzing the presence of chirp signals in the electroencephalogram during seizure using the reassignment time-frequency representation and the Hough transform

This paper presents an algorithm to analyze chirp-like signals that occur during seizures. It is based on a reassignment time-frequency representation and image processing technique. Firstly, the reassignment smoothed pseudo Wigner-Ville distribution (RSPWVD) was calculated, and then the Hough transform was executed on the RSPWVD in the time-frequency plane. The algorithm was applied to traditional EEG (electroencephalogram) signals and to the recently proposed Laplacian EEG recorded using tripolar concentric ring electrodes (TCREs), referred to as (tEEG). Preliminary results indicate that this algorithm is well suited for chirp-like activity associated with seizures and that the tEEG gives better performance than the EEG.