A Practical EEG Study on Autism using Artificial Neural Networks

X. You, N. Teng, M. Ayala, L. Wang, A. Barreto, N. Rishe, and M. Adjouadi (USA)

Keywords

Autism, MMN, artificial intelligence, classification, and neural networks

Abstract

Autism is characterized as a spectrum of neurodevelopment impairments in communicative, social behavioural, and sensory motor skills. Public concerns about autism have grown in recent years due to the prevalence of its diagnosis in 1 out of 150 young children. Though many researches have been carried out to analyse autistic patients’ EEG behaviour, an effective physiological diagnosis for autism does not exist and researchers haven’t found a distinguishing pattern to classify autistic and non-autistic subjects. This preliminary study analyses the EEG data to compare patterns of speech and non- speech sound discrimination between 8 non-autistic and 4 autistic teenagers. An Artificial Neural Networks (ANNs) based classifier has been implemented to determine whether EEG data reflects differences from the two types of responses.

Important Links:



Go Back