Faculty of Electrical, Biomedical and Mechatronics Engineering, Qazvin Branch, Islamic Azad University, Qazvin, Iran
Autism, also called autism spectrum disorder (ASD), is a complicated condition that includes problems with communication and behavior. It can involve a wide range of symptoms and skills. ASD can be a minor problem or a disability that needs full-time care in a special facility. People with autism have trouble with communication. They have trouble understanding what other people think and feel. This makes it hard for them to express themselves, either with words or through gestures, facial expressions, and touch. According to the Centers for Disease Control, autism affects an estimated 1 in 59 children today. Indicators of autism usually appear by age 2 or 3. Some associated development delays can appear even earlier, and often, it can be diagnosed as early as 18 months. Research shows that early intervention leads to positive outcomes later in life for people with autism. In this paper, we describe an Autism detection algorithm that runs over electroencephalography (EEG) signals. Because this technique comprises different parameters that significantly affect the detection performance, we will use genetic algorithms (GAs) to optimize these parameters to improve the detection accuracy. And in the end, the results have been compared statistically by the T-test. In this paper, we describe the GA setup. EEG signals of 20 children with Autism and 20 healthy children aged 6 to 12 years have been obtained. The results have been compared. Lower correlation levels between resources of the left hemisphere of the brain especially C3 channels region in autistic children compared with healthy subjects have been observed. Also, the average energy of theta frequency band in C3 and F3 channels for children with autism was lower than that in healthy people and this criterion was higher in the gamma frequency band.