Automated Diagnosis of Encephalopathy Using Fractal Dimensions of EEG Sub-Bands

Abstract

The pattern of fractal dimensions in various EEG sub-bands is analyzed in a particular disease called encephalopathy. Both Higuchi's fractal dimension and Katz's fractal dimension are analyzed in this paper. After extracting EEG sub-bands from the signal using wavelet decomposition, fractal dimensions are found out for all sub-bands and compared between normal and disease groups. Katz's fractal dimension of sub-bands of EEG gave better performance parameters for Support Vector Machine classifier for identifying Encephalopathic cases than Higuchi's fractal dimension.