Natural language processing methods are sensitive to sub-clinical linguistic differences in schizophrenia spectrum disorders

This study explores the use of natural language processing (NLP) methods to detect sub-clinical linguistic differences in schizophrenia spectrum disorders (SSD). The researchers compared speech samples from 20 SSD participants and 11 healthy controls, analyzing linguistic features at three levels: individual words, parts-of-speech (POS), and sentence-level coherence. NLP features were compared to clinical ratings using the Scale for the Assessment of Thought, Language and Communication (TLC). Results showed that SSD participants used more pronouns and incomplete words but fewer adverbs, adjectives, and determiners. Sentence-level analysis using BERT revealed increased tangentiality in SSD. While there was no significant difference in TLC scores between groups, NLP measures showed greater ability to discriminate between SSD and healthy controls than clinical ratings alone, suggesting NLP's potential for identifying sub-clinical language disturbances in SSD.

Sunny X. Tang, Reno Kriz, Sunghye Cho, Suh Jung Park, Jenna Harowitz, Raquel E. Gur, Mahendra T. Bhati, Daniel H. Wolf, João Sedoc, Mark Y. Liberman