Özet:

Artificial Intelligence (AI) methods have been generally used in neuroimaging data to identify patients with psychiatric problems/disorders. Schizophrenia (SZ) is generally defined as a mental problem that affects the thinking ability and memory. Manual assessment of SZ participants is sometimes difficult and susceptible to diagnostic mistakes. Thus, we achieved a Computer Aided Diagnosis (CAD) algorithm to analyze and interpretate SZ patients successfully using single channel measurement Electroencephalogram (EEG) signals with Signal Processing and Artificial Intelligence methods. First, the EEG signals of participants were pre-processed (signal enhancement, filtering, noise removal), Then, signals were disseminated into windowing/segmentation process. Then, the EEG signals are separated with wavelet decomposition via seven sub-bands. Next, the feature extraction process was achieved and specific feature parameters were obtained by summing the numerical values of the processed signals. Then, Feature ranking process was achieved to identify the obtained features of the normal and schizophrenia groups. After ranking process, features are fed to AI (SVM), We have obtained the highest accuracy of 99.31% using SVM with five fold and take off one cross validations.

Anahtar Kelimeler: