Scientists at the University College London Institute of Neurology have created new tools based on AI language models that can characterize subtle signatures in the speech of patients diagnosed with schizophrenia.
Image Credit: Jamie Jin/Shutterstock.com
The study, published in PNAS, explores how automated language analysis could help clinicians and scientists diagnose and assess psychological disorders.
At present, psychiatric diagnosis heavily relies on discussions with patients and their close associates, with minimal involvement of diagnostic tests such as blood tests and brain scans.
This lack of precision, however, hampers a deeper understanding of the origins of mental illness and the effectiveness of treatments.
The researchers asked 26 participants with schizophrenia and 26 control participants to complete two verbal fluency tasks, where they were asked to name as many words as they could, either belonging to the category “animals” or starting with the letter “p”, in five minutes.
To analyze the responses provided by participants, the team employed an AI language model trained on extensive internet text to interpret the meaning of words akin to human understanding. They conducted tests to determine whether the words spontaneously recalled by individuals could be predicted by the AI model and if this predictability diminished in patients diagnosed with schizophrenia.
They discovered that the AI model predicted more responses given by control participants than those supplied by those with schizophrenia and that this difference was largest in patients with more severe symptoms.
The researchers believe that this distinction could be related to how the brain processes connections between memories and ideas, storing this information within what are referred to as 'cognitive maps.' This hypothesis is also supported in another segment of the study, where the authors employed brain scanning techniques to measure activity in specific brain regions responsible for forming and retaining these 'cognitive maps.'
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 AI language models to psychiatry—a medical field intimately related to language and meaning.
Dr. Matthew Nour, Study Lead Author and NIHR Clinical Lecturer, Department of Psychiatry, University of Oxford
Schizophrenia is a devastating mental condition that affects over 24 million individuals globally and over 685,000 people in the United Kingdom.
Symptoms of the disease include hallucinations, delusions, muddled ideas, and behavioral abnormalities, according to the NHS.
The researchers from UCL and Oxford intend to test this technique on a wider sample of patients in more diverse speech settings to see whether it might be effective in the clinic.
Dr. Nour concluded, “We are entering a very exciting time in neuroscience and mental health research. By combining state-of-the-art AI language models and brain scanning technology, we are beginning to uncover how meaning is constructed in the brain, and how this might go awry in psychiatric disorders. There is enormous interest in using AI language models in medicine. If these tools prove safe and robust, I expect they will begin to be deployed in the clinic within the next decade.”
Wellcome supported the research.