What is the clinical evidence from this study?

Study design: Other. Population: Schizophrenia (n=81). Intervention: Deep Residual Network (deep ResNet) feature extraction with Support Vector Machine (SVM) classifier vs. Deep ResNet with softmax classifier. Primary outcome: Classification accuracy for schizophrenia detection.

March 22, 2023Open Access

Exploring deep residual network based features for automatic schizophrenia detection from EEG

Key Result

Deep residual network-based feature extraction combined with a support vector machine classifier achieved 99.23% accuracy in automatically detecting schizophrenia from EEG signals.

Structured PICO

Does a deep residual network (deep ResNet) based feature extraction improve the accuracy of automatic schizophrenia detection from EEG signals compared to traditional methods?

Population

81 subjects (49 with schizophrenia and 32 healthy controls, 17.3% female, mean age 39 years) whose EEG data from a basic sensory task were used to evaluate a deep learning classification model.

Intervention

Deep residual network (deep ResNet) based feature extraction from EEG signals combined with machine learning classifiers (e.g., SVM).

Comparator

Traditional machine learning methods and ResNet softmax classifier.

Outcome

Accuracy of schizophrenia detection.surrogate

A deep residual network-based feature extraction method combined with an SVM classifier achieved 99.23% accuracy in automatically detecting schizophrenia from EEG signals.

Main Result

Absolute Event Rate: 99.23% vs 97.48%

Limitations

The deep learning-based model required significantly longer computational time for training (14,749 seconds) and testing (20.50 seconds) compared to traditional machine learning methods.

Abstract

Schizophrenia is a severe mental illness which can cause lifelong disability. Most recent studies on the Electroencephalogram (EEG)-based diagnosis of schizophrenia rely on bespoke/hand-crafted feature extraction techniques. Traditional manual feature extraction methods are time-consuming, imprecise, and have a limited ability to balance accuracy and efficiency. Addressing this issue, this study introduces a deep residual network (deep ResNet) based feature extraction design that can automatically extract representative features from EEG signal data for identifying schizophrenia. This proposed method consists of three stages: signal pre-processing by average filtering method, extraction of hidden patterns of EEG signals by deep ResNet, and classification of schizophrenia by softmax layer. To assess the performance of the obtained deep features, ResNet softmax classifier and also several machine learning (ML) techniques are applied on the same feature set. The experimental results for a Kaggle schizophrenia EEG dataset show that the deep features with support vector machine classifier could achieve the highest performances (99.23% accuracy) compared to the ResNet classifier. Furthermore, the proposed model performs better than the existing approaches. The findings suggest that our proposed strategy has capability to discover important biomarkers for automatic diagnosis of schizophrenia from EEG, which will aid in the development of a computer assisted diagnostic system by specialists.

KI fragen

Bookmark

View Full Paper