Wavelet based detection of epileptic seizures using scalp EEG

Epilepsy is one of the most common neurological disorder in the world. It is characterized by sudden and recurrent neuronal firing in the brain. Electroencephalogram is a major tool used for the detection of seizures. Feature extraction is one of the important aspects in epileptic seizure detection using electroencephalogram. In this paper, scalp Electroencephalogram is used for seizure detection. Wavelet transform based features have been used for the detection of seizures. Eight features viz. Standard deviation, Mean Absolute Deviation, root mean square value, minimum, interquartile range, skewness, entropy and maximum were extracted over wavelet coefficients. Ranking of features was done using T-test class separability criterion. Classification was done using Support Vector Machine classifier using six most significant features. A specificity of 100%, sensitivity of 97.2% and an accuracy of 98.6% were obtained. Results shows an improvement compared to the related works.