Linear and nonlinear analysis of intrinsic mode function after facial stimuli presentation in children with autism spectrum disorder

In this work, a method for classifying Autism Spectrum Disorders (ASD) from typically developing (TD) children is presented using the linear and nonlinear Event-Related Potential (ERP) analysis of Electro-encephalogram (EEG) signals. The signals were acquired during presentation three types face expression stimuli —happy, fearful neutral faces. EEGs are first decomposed Multivariate Empirical Mode Decomposition (MEMD) to extract its Intrinsic Functions (IMFs), which provide information about underlying activities ERP components. sample entropy (SampEn) features, as well standard measurements utilizing maximum (Max.), minimum (Min), deviation (Std.), then extracted each set IMFs. These features evaluated by statistical tests used construct input vectors Discriminant (DA), Support vector machine (SVM), k-Nearest Neighbors (kNN) classifiers. Experimental results show that proposed can differentiate ASD TD happy stimulus dataset with high classification performance all classifiers reached 100% accuracy. This result suggests general deficit in recognizing positive children. Additionally, we found SampEn computed alpha theta bands delta band be considered biomarkers disturbances Emotional Facial Expression (EFE) processing • article new approach based on their EEG A combination was analyse components response emotional facial expression. time-frequency successfully differentiated two groups accuracy 100%.