2023-10-29 13:07:58
Researchers at University College London have successfully used an AI language model to identify disruptions in speech patterns in individuals with schizophrenia. The study, conducted by psychiatrist and neuroscientist Matthew Nour and his colleagues, showcased the potential application of AI in psychiatry.
Schizophrenia is a mental disorder that affects approximately two to four in every thousand individuals in the Netherlands. Those with the condition experience recurring psychoses, which involve severe disturbances in thinking and feeling, often accompanied by hallucinations and delusions.
In addition to these symptoms, many individuals with schizophrenia struggle with disorganized thinking and speaking. This may be attributed to disruptions in their cognitive map, which is a mental representation that illustrates relationships between different concepts. For example, thinking of a pen might lead to thoughts of paper due to the close association in the cognitive map.
Accurate monitoring of symptoms is crucial for effective schizophrenia treatment. AI has the potential to significantly enhance this process. By leveraging an AI language model, the researchers aimed to assess the cognitive map of individuals with schizophrenia.
Participants with and without schizophrenia were given a task to name as many animal-related words as possible within five minutes. The researchers hypothesized that individuals with a strong cognitive map would produce words in succession that were semantically related. To determine the coherence between words, the researchers utilized a Natural Language Processing model, such as ChatGPT, which predicts word associations based on patterns in internet texts.
The study revealed that individuals with schizophrenia, compared to the control group, tended to mention words in succession that were less semantically coherent. The lack of coherence correlated with the severity of certain schizophrenia symptoms. Those with more severe symptoms demonstrated reduced word coherence.
The researchers demonstrated that the AI language model could aid in monitoring schizophrenia symptoms. This approach also aligns with neurological findings. Psychiatry researcher Iris Sommer from the University Medical Center in Groningen expressed that language analysis could play a significant role in understanding schizophrenia and remotely monitoring patients to detect early signs of psychosis.
Although further research is necessary to ascertain the reliability of AI language models in psychiatry, Nour believes they could be integrated into clinical settings within the next ten years. The study also explored the link between cognitive maps and brain activity, focusing on the hippocampal region. The researchers found that disruptions in the cognitive map were associated with weakened brain activity related to unconscious re-experiencing of tasks, suggesting alternative cognitive map formation in affected individuals.]
#monitor #schizophrenia #symptoms