Application of Machine Learning Models for Identification of Epileptiform Activity

Abstract

Diagnosis of epilepsy primarily involves understanding cautious patient history and assessment of EEG (Electro Encephalography), which is an essential diagnostic support tool. It captures the electrical activity in the brain, which enables the neurologist to look for the presence of epileptiform patterns for which brain waves (Delta, Theta, Alpha, Beta, and Gamma) are studied thoroughly. Visual Analyses of EEG for the presence of interictal discharges is a critical task. It needs the expertise of a practiced neurologist to identify the presence of epileptiform patterns. The morphology of inter-ictal activity supports the newlinediagnosis of epilepsy and hence is an integral part of detecting and understanding the disease. newlineAn Inter-ictal state is a period between convolutions (seizures) that are characteristic of newlineepilepsy disorder. A patient with such a disorder often will have a trace of inter-ictal activity in his EEG (electroencephalogram). This study work towards achieving three goals. Firstly on differentiating between different epileptic states; and identification of inter-ictal activity in to support the diagnoses of epilepsy. Secondly, this study investigates the contribution of Beta (13-35 Hz) and Gamma (36 - 44Hz) waves as they present a grave challenge because of their high-frequency nature. This study investigates if these waves incorporate features essential for the identification of inter-ictal activity. Finally, we have also worked to differentiate newlinecollected data with artifacts incurred during the recording of EEG from inter-ictal epileptiform discharges.For achieving these objectives, publically available benchmark dataset Bonn database and novel data collected from Max Hospital, Saket after data after seeking approval from the scientific and ethical committee. newline