Machine learning is a type of artificial intelligence that is trained on data and applied to make predictions.

Research has shown that machine learning models can predict cases of schizophrenia within five years.

Machine learning methods work better on patients with schizophrenia compared to bipolar disorder.

Diagnosing serious mental illnesses like schizophrenia can take up to 10 years after the first episode. As one can imagine, this can have very negative repercussions. Catching psychosis early on helps create a better prognosis. Although there are many cases where those who have a late diagnosis go on to lead fulfilling lives, the hope is that future patients can get to that place faster.

Machine Learning Study

Using machine learning to identify early signs of psychosis, a cohort of 24,449 participants were enlisted. The participants’ data were gathered from the Psychiatric Services of the Central Denmark Region between 2013 and 2016. Notes from the hospital in which the participants stayed were analyzed through their text and keywords.

A machine learning model was then developed on the existing text data to train the model to assess whether future patients would develop schizophrenia within five years. By doing so, the machine learning model used 1092 “predictors,” which are factors related to the disorder that are seen in people who already have it. When predictors are identified in a cluster or increased severity (or even a decreased frequency), a machine learning model can make predictions that a person would have schizophrenia.

Some of the major text predictors for identifying schizophrenia included: hearing voices, discharge or admission from an inpatient hospital visit, contact (or no contact) with female friends, inpatient admission, wake or sleep time in the morning, words like “play” and “the game” (like board games and interacting with hospital staff), and the ability to explain their symptoms, to name a few.

The model trained on these data was able to identify those who would progress from a less severe mental illness to receiving a diagnosis like schizophrenia. The machine learning model was trained and applied on the same or very similar demographic. While the study was designed to attempt to identify both bipolar disorder and schizophrenia, the results mainly showed that the machine learning model was more accurate in determining the diagnosis of schizophrenia.

Limitations of the Study

Since this is an early study for the effectiveness of machine learning, there are some things to watch out for before bringing the technology for widespread use, clinically.

These models were used on patients who already exhibited symptoms, and were progressing from less severe mental illness to a full diagnosis of schizophrenia. The model might use different predictors for those who have minimal to no symptoms, and the researchers stressed that it was important for any machine learning models to be tested on the same or similar demographic of the training sample. A general application of a specific machine learning model might provide a poorer ability to make a diagnostic confirmation.

While novel, there is still much research to be done on the applications of machine learning to diagnosing serious mental illness. Schizophrenia is one of the disorders of the Diagnostic and Statistical Manual of Mental Disorders (DSM) that shows evidence that it is a distinct disorder both biologically and psychologically through brain imaging and self-report measures. The brains of patients with schizophrenia show distinct differences, while it is more difficult to find among those with mood disorders or personality disorders, for example.

However, since it is so distinct, there is hope for advancing the diagnostic process, treatment, and care for patients with schizophrenia. While schizophrenia was once severely neglected in research and public image, now there are methods and treatment options that can help people with schizophrenia spectrum disorders lead more fulfilling lives.

Hansen L, Bernstorff M, Enevoldsen K, et al. Predicting Diagnostic Progression to Schizophrenia or Bipolar Disorder via Machine Learning. JAMA Psychiatry. 2025;82(5):459–469. doi:10.1001/jamapsychiatry.2024.4702

© Psychology Today