Recently, depression detection is mainly completed by some rating scales. This procedure requires attendance of physicians and the results may be more subjective. To meet emergent needs of objective and pervasive depression detection, we propose an EEG based approach for females. In the experiment, EEG of 13 depressed females and 12 age matched controls were collected in a resting state with eyes closed. Linear and nonlinear features extracted from artifact-free EEG epochs were subjected to statistical analysis to examine the significance of differences. Results showed that differences were significant for some EEG features between two groups (
<0.05) and the classification rates reached up to 92.9% and 94.2% with KNN and BPNN respectively. Our methods suggest that the discrimination of depressed females from controls is possible. We expect that our EEG based approach could be a pervasive assistant diagnosis tool for psychiatrists and health care specialists.