Certain investigations on diabetic neuropathy using soft computing techniques and its FPGA implementation

Abstract

This thesis investigates the epilepsy risk level classification in diabetic neuropathy. Softcomputing techniques and Very Large Scale Integration technology are used. Worldwide, 466 million people are affected by the diabetes. Diabetes affects all body systems. Mainly it produces one brain disorder called as epilepsy in which the cluster of nerve cells in the brain functions abnormally. Brain neuron electrical activity are amplified and brought into a visible record format through the Electroencephalogram (EEG) signal. Thus an EEG signal is used to reflect the cerebral activity. Patients with suspected epileptogenic foci in their brain are subjected to an electroencephalography recording in the neurophysiology laboratory. Detection of abnormal EEG signal plays a significant role in the diagnosis of epilepsy. Approximately 50 million people are affected by this epilepsy seizure. An early detection of epilepsy would help the patients to take some precautions to avoid further complications. While finding the diabetic epilepsy risk level it is equally important to identify an appropriate treatment for the same. If there is any wrong diabetic epilepsy level identification, it may lead to further serious health complications. So it is necessary to identify the risk level accurately in the healthcare field. It is not possible to arrive at a correct crisp decision while identifying the epilepsy risk level. Due to the uncertainty levels in the EEG signals classification, its results in an imprecise boundaries between various linguistic levels of membership function. This work proposes performance analysis using soft computing techniques and FPGA (Field Programmable Gate Array) implementation of diabetic epilepsy risk level classification. A group of 200 diabetic patients Cerebral Blood Flow (CBF) were measured and EEG signals were recorded at Sri Ramakrishna Hospital, Coimbatore and Government Hospital, Dindigul. newline