Fast computer to say something and see if you have schizophrenia

　　I like to play games, I mean, I don't have to play the game is not available, but playing games is fun. You know, on the one hand to play the game and there is no problem. OK, I lied, I was scared to play the game.

　　If you speak of logic and similar to the phrase above, be careful that may be in danger of suffering from schizophrenia. About mental illness, doctors in most cases is by asking a series of questions and then answer according to the doctor to determine if someone is sick. The future, diagnosis of mental illness may be able to let the computer do it.

　　Recently, Colombia University, New York Institute and IBM Watson Research Center published a study report pointed out that computers can be semantic and grammatical analysis to distinguish normal people, people who are at risk for schizophrenia risk and people who suffer from schizophrenia 2.5 years. Predict the accuracy of respondents suffer from schizophrenia to reach 100%, than an optimal level, 79% of higher accuracy. Jet Pack China s first behind the scenes including

　　People rarely tracked how his ideas into language the process, but for Psychiatry, and complex thought process is the most important point. As the thought processes often decides the language mode, so by express language monitor tracks a person will be able to determine whether the person is suffering from mental illness.

　　Researchers is based on this principle, allow the computer to track and identify the prevalence of respondents. From the IBM Watson Research Center, Guillermo Cecchi said in a interview with The Atlantic, for example, they found, respondents if the semantics are not coherent or logical confusion that it would later become schizophrenic. But the visit lasts 45 minutes, during which if the visitor a little God, will miss the respondents part of semantic confusion. Computers are different, it can include the entire dialog.

　　In view of the above reasons, the researchers designed a set of algorithms, so that your computer can automatically recognize speech, followed by judging the speaker of semantic coherence and logic to achieve the forecast.

　　However, the study's sample size is too small (only 34 respondents), the future they will study ways to more of the same on the respondents. Apart from the diagnosis of schizophrenia, the research team said they will further analysis of patients with schizophrenia language and semantics to find targeted treatments.