Multichannel convolutional transformer for detecting mental disorders using electroancephalogrpahy records.

Area of Science:

• Neuroscience
• Artificial Intelligence
• Psychiatry

Background:

• Mental disorders pose a significant global health burden, necessitating early and accurate detection for effective intervention.
• Electroencephalography (EEG) offers a non-invasive method to monitor brain activity, crucial for identifying potential mental health conditions.
• Deep learning models have shown promise in analyzing complex EEG data for disorder classification.

Purpose of the Study:

• To introduce a novel deep-learning architecture, the multichannel convolutional transformer, for classifying mental disorders using EEG data.
• To enhance EEG signal processing through advanced filtering and time-frequency transformations for improved feature extraction.
• To evaluate the proposed model's performance against existing methods on multiple benchmark datasets.

Main Methods:

• EEG data was pre-processed using common spatial pattern, signal space projection, and wavelet denoising filters.
• Continuous wavelet transform was applied to obtain time-frequency representations of EEG signals.
• A multichannel convolutional transformer model, integrating CNNs and transformers, was developed for EEG data classification.

Main Results:

• The proposed model achieved high classification accuracies: 87.40% on the EEG and Psychological Assessment dataset, 89.84% on the MODMA dataset, and 92.28% on the EEG Psychiatric dataset.
• The model demonstrated superior performance compared to all concurrent approaches across the evaluated datasets.
• No signs of over-fitting were observed, indicating robust generalization capabilities.

Conclusions:

• The multichannel convolutional transformer architecture shows significant potential for accurate and reliable mental disorder detection via EEG analysis.
• This approach can pave the way for advancements in the early diagnosis and treatment strategies for mental health conditions.
• The study highlights the efficacy of combining advanced signal processing with deep learning for psychiatric disorder classification.