Evaluating Ratio Indices Based on Electroencephalogram Brainwaves in Schizophrenia Detection

Extensive research has been conducted by neurocognitive and psychological scientists to diagnose mental and neurological diseases intelligently. Recently, researchers have shown interest in Electroencephalogram (EEG) analysis, a non-invasive method of recording the brain’s electrical activity from the scalp surface. EEG signals contain different frequency bands, each indicating specific brain activities. Although the relative powers of single EEG waves are not all-inclusive indicators to consistently imitate mental involvement, ratio indices should be considered. These indices calculate the ratio of powers (summations) with more than a single frequency band.

This study quantifies the EEG signals of healthy control and schizophrenic groups using thirty-seven ratio indices based on EEG brainwaves. These indicators are examined for the first time in schizophrenia. The study evaluates which index is more suitable and efficient for solving a classification problem.

The results show the potential of (delta + theta)/alpha in the schizophrenia classification with an average accuracy of 97.92%. Additionally, the study investigates the effectiveness of different EEG electrodes in the problem of schizophrenia diagnosis while utilizing the above indicators. T5, the left posterior temporal region, yields a maximum average accuracy of 92.92%.

In conclusion, the fusion of EEG frequency ratio indices and machine learning algorithms provides a potential avenue for improving the detection and diagnosis of schizophrenia.