Design and Development of Electroencephalogram EEG based Autirm Spectrum Disorder ASD Classifier

Abstract

Electroencephalogram (EEG) signals have a morphological pattern that distinguishes between normal and autism children. Screening children for autism using EEG signals has the characteristics of good sensitivity and real-time support. Existing EEG screening methods focus on analyzing the data during the resting state of children and their features extracted to have a weak capacity to distinguish and are information redundant. Detection of autism in children in real-time classroom environment and monitoring of progression of autism with intervention and treatment is an important focus of research. Early detection of abnormalities that can identify autism in young children using EEG data is an advantage for doctors treating autism kids. EEG signals are observed to be nonstationary and complex, extraction of features from EEG signals is required to be accurate and non-redundant for detection and classification. Complex wavelets have demonstrated advantages in terms of shift-invariant property, phase information, and directional information. Very few research on the use of complex wavelets for EEG signal analysis is reported in the literature. In this work, a detailed study on complex wavelets for EEG signal analysis and the use of wavelet features for classification is carried out. In this chapter, an introduction to autism detection using EEG signals, a review of existing methods, limitations of existing methods, and new methods developed are discussed newline