| Autor: |
Thilakavathi, B., Devi, S. Shenbaga, Bhanu, K. |
| Zdroj: |
International Journal of Biomedical Engineering and Technology; 2018, Vol. 28 Issue: 1 p18-37, 20p |
| Abstrakt: |
The objective of this work is to identify schizophrenia using electroencephalogram spectral feature. This work proposes two oddball paradigms to stimulate mental activity for cognitive process and EEG is recorded during both rest and mental activity. EEG from 51 schizophrenia and 26 normal subjects are used in this work. The maximum power in various bands and its corresponding frequency are estimated from the EEG power spectrum and are called as peak power and peak frequency, respectively. Statistical evaluations are made using Student's t-test and ANOVA. Theta peak frequency, beta peak frequency and beta peak power, all with a low value of p (p < 0.001) and alpha peak power with p < 0.01 during mental activity are identified as most significant features. SVM classifier designed with these features has a classification accuracy of 91.75% which shows that EEG can be used as a clinical tool to detect schizophrenia. |
| Databáze: |
Supplemental Index |
| Externí odkaz: |
|
