The psychologist will be unemployed? Artificial intelligence assisted mental illness diagnosis typical case sharing

Ordinary psychological diagnosis and treatment, doctors must first diagnose the patient's mental condition, judge the symptoms through several intuitions similar to psychological interviews, diagnose the mental illness types by experience, and then formulate treatment plans for the symptoms, including which drugs and drugs. Quantity and so on. However, subject to the doctor's subjective judgment and experience limitations, there may be errors in the diagnosis, resulting in delay in the diagnosis, or can not correctly determine the type and dose of medication, delay the condition.

In this era of machine learning, artificial intelligence has become more widely used in psychiatry and has been involved in both the diagnosis and treatment stages. Artificial intelligence has shown a high degree of judgment accuracy in experiments, has more than human trends, and is expected to play an important role in the future diagnosis and treatment of mental illness.

However, if artificial intelligence is to be used to assist in the intervention of mental illness, detailed research and verification are required. For example, although the use of linguistic analysis techniques to diagnose schizophrenia has proven to be quite accurate, it is only the result of a trial. Moreover, no one has yet established a model for the symptoms of depression and bipolar disorder.

AI is still a long way to go in practice, especially for accuracy. Siri's total problem in identifying Scottish accents may be harmless, but in the medical field, small mistakes can have devastating consequences. "If a technology has a 20% error rate (or 80% accuracy), I will never use it on patients," said Harvard University professor Vahabzadeh.

Coupled with factors such as age, gender, race, and region, artificial intelligence poses a greater risk. If the data source population of an artificial intelligence system belongs to the same demographic category, and then test other people, even normal samples will be misjudged. "Some specific ethnic groups, the voices of speaking are relatively soft, or some people are limited by the physiological factors, so it is justifiable that artificial intelligence misdiagnoses them as depression." To strengthen machine learning in mental illness The application currently only provides more "training" through a larger database of human behavior. Perhaps in the near future, artificial intelligence technology can solve the shortage of mental illness patients and greatly reduce labor costs.

Despite this, there are some relatively good programs in the market that use AI to assist in the diagnosis of mental illness. The author introduces you to some of the applications and new trends of artificial intelligence in the spiritual field, as well as some of the challenges they face.

1.NeuroLex Diagnostics: Helps diagnose schizophrenia

The speech of people with schizophrenia has very significant characteristics, often manifested as involuntary voices: most of the short sentences in the speech, semantic confusion, the use of fuzzy words such as "this", "that", "one", etc., sentences and sentences The ideology of the connection is ambiguous. In 2015, a group of researchers created an artificial intelligence model based on the linguistic features of schizophrenia. By analyzing the conversation records, it accurately predicted which group of young people might suffer from mental disorders (the main symptoms of schizophrenia).

Jim Schwoebel, CEO of NeuroLex Diagnostics, wants to use this technology to help newly diagnosed patients screen out patients with schizophrenia. NeuroLex's product logic is this: install a recordable device such as a smartphone in the secret room of the treatment room, record the conversation, and then use the artificial intelligence model used in the above study to find the clues. The system automatically uses numbers to characterize the severity of the illness (just like a sphygmomanometer) and provides a diagnostic reference for the doctor.

NeuroLex Diagnostics is dedicated to analyzing words and improving health

The invention of the schizophrenia filter has given Schwoebel the American Psychiatric Association award. In addition, NeuroLex wants to create a product for patients who are undergoing psychiatric treatment in the hospital. In addition to diagnosing mental illness, it can further conduct long-term language analysis of patients to record the progress of the disease.