A Novel Algorithm for Detection of Social Joint Attention from single-trial EEG signals of Autistic Spectrum Disorder (ASD)

Abstract

Autism Spectrum Disorder (ASD) is a developmental disability that shows its symptoms during second year of life or later. People with ASD have poor performance in different abilities such as social joint attention. In recent years, scientists has developed different methods and technologies to help ASD people to have a better life. Brain-Computer Interface (BCI) is one of these technologies that plays an important role in the rehabilitation of patients with neurological disorders. BCI is a technology that makes a bridge between brain and computer which let brain signals like EEG be recorded for processing. On the other hand, it has been shown that social joint attention can be detected in EEG signals using P300 which is one of the most popular components of Event-Related Potential (ERP). Therefore, recording EEG signals with the use of BCI for processing them with a reliable algorithm can be a great step toward understanding the helping people with ASD to improve their social joint attention. In this study, a novel algorithm based on a Convolutional Neural network will be introduced for detection attention in single-trial EEG signals more precisely. As a dataset, IFMBE MEDICON 2019 challenge dataset will be used in which autistic adults were learning social joint-attention with the help of the BCI system. Results show that it can improve the performance of detecting P300 from single-trial EEG signals effectively in comparison to other algorithms. This method increased final target detection accuracy from 92.37% to 94.85%.