Depression signal correlation identification from different EEG channels based on CNN feature extraction.

Area of Science:

• Neuroscience
• Computer Science

Background:

• Depression is a serious mental illness requiring accurate diagnosis and treatment.
• Electroencephalography (EEG) is used for depression diagnosis, but extracting features from multimodal channels is complex.
• Identifying channels with strong depression information can simplify the diagnostic process.

Purpose of the Study:

• To propose a method for identifying EEG channels strongly correlated with depression.
• To simplify the clinical depression diagnosis process by focusing on relevant EEG channels.

Main Methods:

• A depression signal correlation identification method based on convolutional neural network (CNN) was developed.
• Labeled multi-channel EEG data was used, with signals divided into training datasets.
• The AlexNet network was employed for training, followed by correlation classification of each channel for depression.

Main Results:

• The study found that the correlation between EEG channels and depression is not consistent.
• Specific channels (e.g., 13, 17, 28, 40, 46, 66, 69) showed a strong correlation with depression.
• These identified channels can be selected for more efficient depression diagnosis.

Conclusions:

• A CNN-based method effectively identifies EEG channels with strong depression correlation.
• Focusing on these specific channels can streamline the diagnosis of clinical depression.
• This approach aids in the early detection and treatment of depression.