Design and development of signal Conditioning system for nano Sensor based health applications

Abstract

Schizophrenia causes hallucinations and delusions. The person newlinesuffering from schizophrenia cannot process reality clearly because of newlineconstant auditory and visual hallucinations. The hallucination and newlinedelusion cause abnormal brain activity. The abnormal behavior detects newlinethrough EEG signal acquisition. The EEG electrodes have low newlinesensitivity to the current distribution of the brain. The distorted and low newlinesensitive EEG signal leads to inaccurate interpretation of brain activity newlinedue to schizophrenia. In this thesis, we propose an EEG electrode made newlineof Graphene Nano powder sensitive to the low electrical current of the newlinebrain. The cold spray technique is used to create graphene EEG newlineelectrodes in order to enhance the chemical and bonding properties of newlinematerials. The sensitivity of the Graphene electrode is evaluated by newlineobtaining EEG signals from schizophrenic patients. The EEG signal is newlineacquired from the person during cognitive tests such as numerical tasks newlineand query sessions. The hallucination and delusion features in EEG newlinesignals identify with different Neural Network (NN) algorithm. The NN newlineshows Graphene electrode provides more information about newlinehallucination and delusion features in EEG signal. The comparative newlineexamination of different NN models reveals BFGS Quasi Newton Back- newlinePropagation Algorithm identifies hallucination and delusion newlinecharacteristics with minimal errors compared to other NN models. newlineThe proposed research work is intended to provide an newlineix newlineautomatic diagnostic system to determine the EEG signal in order to newlineclassify the brain function which shows whether a person is affected newlinewith schizophrenia or not. However, the diagnosis of schizophrenia still newlinedepends on clinical observation to date. Without reliable biomarkers, newlineschizophrenia is difficult to detect in its early phase and hence we have newlineproposed this idea. In this work, the EEG signal series are divided into newlinenon-linear feature mining, classification and validation, and t-test newlineintegrated feature selection process. At last, the Deep Learning (DL) newline