Machine Learning Techniques for Schizophrenia Detection Using EEG Signals

Recent advances in the intersection of neuroscience and artificial intelligence have driven the development of machine learning systems capable of detecting schizophrenia from electroencephalography (EEG) recordings. EEG offers a non-invasive window into brain dynamics, capturing electrical oscillations that reflect neural synchrony and connectivity. Machine learning pipelines typically begin with signal acquisition and preprocessing to remove artefacts, followed by feature extraction in time, frequency or entropy domains. Classical methods such as support vector machines and k-nearest neighbours evolved into ensemble classifiers and, more recently, deep neural networks incorporating convolutional and recurrent layers. Hybrid architectures exploit spatial representations of EEG features and temporal dependencies, yielding classification accuracies often exceeding 95 per cent. Emerging approaches also explore quantum-inspired algorithms for high-dimensional pattern recognition. Across global research groups, these methodologies promise earlier diagnosis, more objective biomarkers and streamlined clinical workflows, with potential to complement traditional symptom-based assessments and improve long-term outcomes for individuals at risk.

Research from Nature Portfolio

A 2021 investigation introduced a hybrid deep neural network that transforms preprocessed EEG time-series into spatially encoded red–green–blue images, capturing both time-domain and frequency-domain features. By extracting fuzzy entropy and fast Fourier transform metrics and feeding them into a combined convolutional neural network (CNN) and long short-term memory (LSTM) framework, researchers achieved an average classification accuracy of over 99 per cent. This work underscores the value of combining topographical mapping with sequence modelling to enhance discriminative power between healthy controls and patients with schizophrenia, while demonstrating robustness across feature types.

Machine Learning Techniques for Schizophrenia Detection Using EEG Signals publication trend

The graph below shows the total number of articles in machine learning techniques for schizophrenia detection using eeg signals across all publications each year (not limited to Nature Index journals).

Technical terms

Electroencephalography (EEG): A non-invasive technique for recording electrical activity of the brain via scalp electrodes.

Convolutional Neural Network (CNN): A deep learning architecture that applies convolutional filters to extract spatial features from input data.

Long Short-Term Memory (LSTM): A recurrent neural network unit designed to capture long-range temporal dependencies in sequential data.

Support Vector Machine (SVM): A supervised learning algorithm that separates classes by finding the optimal hyperplane in feature space.

Quantum Support Vector Machine (QSVM): A quantum-enhanced version of SVM that encodes data into qubit states and leverages quantum kernels for classification.

Fuzzy entropy: A measure of signal complexity reflecting irregularity and unpredictability in time-series data.

References

1. Quantum Machine-Based Decision Support System for the Detection of Schizophrenia from EEG Records. Journal of Medical Systems (2024).
2. A hybrid deep neural network for classification of schizophrenia using EEG Data. Scientific Reports (2021).
3. CGP17Pat: Automated Schizophrenia Detection Based on a Cyclic Group of Prime Order Patterns Using EEG Signals. Healthcare (2022).