Automated Diagnosis of Encephalopathy Based on Empirical Mode EEG Decomposition

In this study, an empirical mode decomposition (EMD) technique has been applied for EEG signals to identify a neurological disease state qualified as encephalopathy. The EMD technique is an efficient method for decomposing nonstationary and nonlinear signals, which makes it suitable for biosignal processing. This technique generates various components of the signal called intrinsic mode functions (IMFs) whose features are examined for the diagnosis of the disease. We found significant differences between the healthy and patient groups for both statistical and nonlinear parameters of IMFs of the recorded EEGs, which makes those suitable for the diagnosis of encephalopathy. Statistical values, namely minimum, maximum, mean, and standard deviation, and nonlinear parameters, namely approximate entropy and sample entropy of the IMFs, were calculated. Both these features were fed to a Support Vector Machine (SVM) classifier, and their performance parameters were evaluated. It is concluded that statistical parameters, as well as nonlinear parameters of the EEG IMFs, are prospective potential features for automated diagnosis of encephalopathy.

from Physiology via xlomafota13 on Inoreader https://ift.tt/2Rnqo68
via IFTTT