Matched Meyer neural wavelets for clinical and experimental analysis of auditory and visual evoked potentials
暂无分享,去创建一个
The wavelet transform provides a time-scale analysis that permits flexible pattern recognition, component identification, and detection of transients for time-varying neural signals such as the EEG, event-related potentials, neuromagnetic signals, and other neural signals and images. Many future applications to neural signals will benefit from choosing a mother wavelet that mimics neural waveform features. We use a recently developed algorithm to design physiologically realistic orthonormal Meyer wavelets, including 1) a wavelet that matches the prominent IV-V complex of the auditory brainstem evoked response used widely for clinical evaluation of hearing loss, and 2) a wavelet that matches ERPs containing prominent P300 components from control and alcoholic subjects. We also compare the relative naturalness of dyadic decompositions that use matched Meyer wavelets, the Haar wavelet, and Daubechies D4 wavelet. Designer neural wavelets have broad potential to customize and improve neurometric imaging and clinical neurodiagnosis of sensory and cognitive dysfunction.