Introduction to Depression
Depression is one of the most common mental disorders, affecting millions of people worldwide. Diagnosis and treatment of depression often rely on subjective scales and questionnaires, which can lead to inaccurate diagnoses and ineffective treatment.
Deep Learning Model for Depression Prediction
A multidisciplinary team of Chinese researchers developed a deep learning model for automated and accurate prediction of depression severity. The model is based on electroencephalography (EEG) data in a resting state, collected from 70 patients with confirmed depressive disorder and 30 healthy individuals from a control group.
Model Architecture
The proposed architecture is called PLI_GE_gMLP. Within a single framework, the developers synergistically combined three advanced technological methods: phase lag index (PLI), graph embedding (GE), and gated multilayer perceptron (gMLP).
Research Results
The AI model PLI_GE_gMLP demonstrated a mean absolute error (MAE) of prediction of disease severity at the level of 4.30. This result qualitatively surpassed the performance of traditional machine learning algorithms and known deep learning architectures.
Conclusions and Prospects
Developing a deep learning model for predicting depression severity can improve diagnosis and treatment of depressive disorders. The authors of the study emphasize the importance of model interpretability and the possibility of using the model to create accessible and affordable hardware-based express-diagnosis systems for mental health.