Abstract:Background and objective: In 2 previous studies, we have shown the ability of special machine learning systems applied to standard EEG data in distinguishing children with autism spectrum disorder (ASD) from non-ASD children with an overall accuracy rate of 100% and 98.4%, respectively. Since the equipment routinely available in neonatology units employ few derivations, we were curious to check if just 2 derivations were enough to allow good performance in the same cases of the above-mentioned studies. Methods: A continuous segment of artifact-free EEG data lasting 1 minute in ASCCI format from C3 and C4 EEG channels present in 2 previous studies, was used for features extraction and subsequent analyses with advanced machine learning systems. A features extraction software package (Python tsfresh) applied to time-series raw data derived 1588 quantitative features. A special hybrid system called TWIST (Training with Input Selection and Testing), coupling an evolutionary algorithm named Gen-D and a backpropagation neural network, was used to subdivide the data set into training and testing sets as well as to select features yielding the maximum amount of information after a first variable selection performed with linear correlation index threshold. Results: After this intelligent preprocessing, 12 features were extracted from C3-C4 time-series of study 1 and 36 C3-C4 time-series of study 2 representing the EEG signature. Acting on these features the overall accuracy predictive capability of the best artificial neural network acting as a classifier in deciphering autistic cases from typicals (study 1) and other neuropsychiatric disorders (study 2) resulted in 100 % for study 1 and 94.95 % for study 2. Conclusions: The results of this study suggest that also a minor part of EEG contains precious information useful to detect autism if treated with advanced computational algorithms. This could allow in the future to use standard EEG from newborns to check if the ASD signature is already present at birth.

Feature Extraction by Nonnegative Tucker Decomposition from EEG Data Including Testing and Training Observations

Classifying Healthy Children And Children With Attention Deficit Through Features Derived From Sparse And Nonnegative Tensor Factorization Using Event-Related Potential

Classification of Normal and Hypoxia EEG Based on Hilbert Huang Transform

Benefits of Multi-Domain Feature of Mismatch Negativity Extracted by Non-Negative Tensor Factorization from EEG Collected by Low-Density Array

[Electroencephalogram Feature Extraction and Classification of Autistic Children Based on Recurrence Quantification Analysis].

Nonnegative Matrix Factorization for EEG Signal Classification

Multi-Domain Feature Extraction for Small Event-Related Potentials Through Nonnegative Multi-Way Array Decomposition from Low Dense Array EEG.

Tensor-driven extraction of developmental features from varying paediatric EEG datasets

Multi-domain Feature of Event-Related Potential Extracted by Nonnegative Tensor Factorization: 5 vs. 14 Electrodes EEG Data

Deep Learning-Assisted Non-Invasive Pediatric Tic Disorder Diagnosis Using EEG Features Extracted by Residual Neural Networks

RESNET34 with Synchrosqueezing Transform for ADHD Disorder Detection Using EEG Signals

Classification of attention deficit hyperactivity disorder using machine learning on an EEG dataset

Detection of an Autism EEG Signature From Only Two EEG Channels Through Features Extraction and Advanced Machine Learning Analysis

A Novel Brain Network Analysis Method for Pediatric ADHD Using RFE-GA Feature Selection Strategy

EEG multi-domain feature transfer based on sparse regularized Tucker decomposition

A new feature extraction and classification mechanisms For EEG signal processing

[Feature Exaction and Classification of Autism Spectrum Disorder Children Related Electroencephalographic Signals Based on Entropy].

An Automatic Framework For Detecting Autism Spectrum Disorder From EEG Signals Using TFD

N-BodyPat: Investigation on the dementia and Alzheimer's disorder detection using EEG signals

Topological Feature Search Method for Multichannel EEG: Application in ADHD classification

Sparse Nonnegative Tensor Decomposition for EEG Data