Scientists from Kaunas University of Technology in Lithuania have developed an artificial intelligence capable of diagnosing depression. It accurately analyzes a person's brain activity and speech. According to the researchers, a person's voice can provide more information about their emotional state than other parameters. Changes in intonation, pace of speech and overall vocal energy can serve as indicators of depression, which are harder to fake compared to facial expressions. The experiment utilized EEG data from the MODMA open dataset, as well as audio recordings obtained while subjects were answering questions and reading texts. Thanks to an advanced deep learning model, DenseNet-121 was able to analyze signal changes and classify participants into healthy and depressed with 97.53% accuracy. According to the authors of the project, in the future, this artificial intelligence could speed up the process of diagnosing depression and make it available remotely. However, additional clinical trials and refinement of the model are needed for implementation