Recently, the 735th Xiangshan Scientific Conference was held in Beijing under the theme of "Challenges and opportunities of new mental health diagnosis and treatment technologies". Hu Bin, professor of the School of Medical Technology of Beijing University of Technology and head of the Brain Health Engineering Team, said that the existing diagnosis methods of mental diseases include specialist interviews and scale diagnosis, which are highly subjective and lack of quantitative indicators. The new artificial intelligence diagnostic technology based on physiological and behavioral signals such as EEG, voice, expression, etc., can achieve a diagnostic accuracy of 70% - 90%. The National Health Plan for the 14th Five Year Plan proposes that by 2025, it will effectively slow down the rising trend of psychological related diseases, effectively control the incidence rate of serious mental disorders, and improve the standardized management rate to more than 90%. The experts at the meeting believed that in order to improve the accuracy and efficiency of the diagnosis of mental diseases as soon as possible, it is urgent to raise the priority of AI related application research and development. Mental disease diagnosis requires artificial intelligence. "Mental science is the field most in need of artificial intelligence technology." Lu Lin, an academician of the CAS Member, said that new research continues to prove that AI models have the potential to outperform traditional diagnosis and treatment models in terms of prediction, diagnosis, intervention and treatment of mental diseases. "EEG, ECG, EMG and other electrophysiological signals are related to people's mental state, and the current acquisition technology has been able to achieve 'noise reduction' acquisition." Hu Bin said that even snoring, microexpression, gait and other behavioral signals that are difficult for human senses to detect laws or accurately capture can be used as the basis for artificial intelligence to judge mental health. For example, based on the principle of "brain gut axis", a wearable bowel sound acquisition and monitoring device can be applied to mental health assessment. Hu Bin said that information technology allows many "clues" that were difficult to detect before to be identified, mastered and used efficiently. In addition to solving the problem of "no trace to find", AI also has unique advantages in treatment. Lu Lin said that doctors need to get feedback in real time to adjust the treatment plan in time. The timely diagnosis and treatment "feedback" provided by AI assisted diagnosis technology can improve the effect of doctors' treatment. Finding the law of objective quantitative evaluation is still a difficult problem because of the complex pathological mechanism and high clinical heterogeneity of mental diseases (each person has different symptoms and effects). "Only through the cross integration of biosensors, artificial intelligence technology and clinical basic research such as medicine and psychology can we make breakthroughs in the objective stratification, classification standards and non drug intervention technologies of mental diseases." Hu Bin said that at present, interdisciplinary research continues to be carried out, but the gap between "language families" needs to be further bridged to form a systematic interpretive paradigm, so that mental health diagnosis and treatment technology "can be identified accurately, explained clearly, and treated correctly". "The portable EEG acquisition device developed by our team innovatively uses physiological signals as objective indicators to assess mental status." Shen Jian, a postdoctoral fellow in the Health Engineering Laboratory of Beijing University of Science and Technology, told the Science and Technology Daily that with the portable EEG mental state assessment technology, researchers can obtain the subject's EEG signals in 180 seconds, during which the system will interact with the subject through different audio sounds