At the UCL Institute for Neurology, researchers are creating AI-based tools able to characterize faint signatures in patients’ speech, who are diagnosed with schizophrenia, paving the way for new breakthroughs of artificial intelligence in psychiatry.
Published in PNAS journal, the research aims to acknowledge that the sequence of automated analysis of language can aid scientists and doctors assess and diagnose psychiatric conditions.
For the time being, psychiatric diagnoses are entirely based on one-on-one sessions with patients and with people who are close to the patients, with the patient’s consent.
Lack of precision in this case prevents a richer understanding of the cause of the mental illness and the proper treatment.
Methodology & Findings of Automate
Researchers requested 26 people with schizophrenia and 26 control participants to complete two verbal fluency tasks in which they were asked to name as many words as they could from the category “animals” or beginning with the letter “p” in a period five minutes.
In order to analyze answers given by participants, an AI language model was used by the team that had been trained on various amounts of internet text in order to represent the meaning of words in a similar manner to humans. Researchers tested whether the people spontaneously recalled words that could be predicted by the AI model, and whether patients with schizophrenia have reduced predictability.
They discovered that the AI model predicted more answers given by control participants than those given by those with schizophrenia, and that this difference was greatest in patients with more severe symptoms.
Brain Activity and Cognitive Maps
Researchers predict that the difference may have a link with the way the brain processes relationships between ideas, memories, and information storage of what is known as ‘cognitive maps’. Support for this theory is achieved in the second part of the study where the authors applied brain scanning to measure brain activity in several parts of the brain that are involved in learning and storing ‘cognitive maps.’
According to Dr. Mathew Nour, lead author, he stated “Until very recently, the automatic analysis of language has been out of reach of doctors and scientists. However, with the advent of artificial intelligence (AI) language models such as ChatGPT, this situation is changing. This work shows the potential of applying artificial intelligence in psychiatry – a medical field intimately related to language and meaning.”
The only issue with AI usage is what if AI imitated characteristics of people with schizophrenia and stopped giving accurate results.
Inside Telecom provides you with an extensive list of content covering all aspects of the Tech industry. Keep an eye on our Medtech section to stay informed and updated with our daily articles.