Application of Resting Brain Frontal Lobe Complexity in Depression Screening

The increase in the number of depressed people worldwide has put forward higher requirements for higher accuracy and efficiency of depression screening. In this study, the wearable EEG devices were applied to improve screening efficiency, and the complexity attenuation rate (CAR) based on the large-scale and small-scale indexes of multivariate multiscale entropy (MMSE) were proposed, in order to improve the accuracy of screening. The EEG resting state recordings including 22 depressed patients and 20 healthy people was collected using a four-channel frontal lobe portable device. The results showed that, compared with healthy people, depressed patients had higher small-scale complexity and lower large-scale complexity, which means that depressed patients have a greater CAR. The study also verified depressed patients had lower alpha and higher beta power. Compared to other features, CAR had the highest correlation with depression scale scores. The leave-one-subject-out classification results showed that the accuracy of combined features (CAR, MMSE, multiscale entropy, and power spectral density (PSD)) reaches 88.63%, which was much higher than the traditional PSD accuracy of 79.60%. To further verify the reliability and robustness of the above results, the proposed method was verified in the public depression dataset, and the accuracy rate was increased to 87.05%. The conclusions proved that the depression screening method based on portable EEG devices proposed in this study is universal and has great practical application value.