Non-invasive electroencephalogram (EEG) test on patients with autism by independent component analysis

Ali Ahmadi¹ Samaneh Hajimohammadbagher²

1) Bachelor of Medical Engineering, Faculty of Electrical Engineering, Alborz University of Knowledge, Qazvin, Iran.
2) PhD in Medical Engineering, Faculty of Electrical Engineering, Alborz University of Knowledge, Qazvin, Iran.

IOI Code : XCCZ-FBEBA
Publication : 10th International Conference on Recent Advances in Industrial Management and Engineering(mieconf.ir)
Abstract :
    An electroencephalogram (EEG) is a non-invasive test that records electrical patterns inside the brain. This test is used to help diagnose complications such as seizures, epilepsy, head injuries, dizziness, headaches, brain tumors and sleep problems. It works. Billions of nerve cells in the brain produce very small electrical signals that produce patterns called brain waves. During an EEG EEG test, small electrodes and wires are attached to the head. Electrodes detect brain waves. And the EEG device amplifies the signals and records them as a wave pattern on graph paper or a computer screen. The application of independent component analysis (ICA) to diagnose autism has been investigated. First, the sources producing EEG signals were extracted with ICA and then time and frequency domain processes were applied to these signal components. EEG signals were obtained from 10 children with autism and 10 healthy children in the age range of 6-11 years. The results were compared by statistical method of t-test. The level of correlation between the left hemisphere sources of the brain, especially the area related to the C3 channel, has been observed in people with autism compared to healthy people. Also, the average energy of theta frequency band in the left hemisphere sources of the brain, especially C3 and F3 channels, is lower for people with autism than healthy people, and this criterion is higher in the gamma frequency band.

Keywords : electroencephalography autism disorders independent component analysis correlation analysis spectrum energy