{"title":"Deciphering language disturbances in schizophrenia: A study using fine-tuned language models","authors":"","doi":"10.1016/j.schres.2024.07.016","DOIUrl":null,"url":null,"abstract":"<div><p>This research presents two stable language metrics, namely Successful Prediction Rate (SPR) and Disfluency (DF), to objectively quantify the linguistic disturbances associated with schizophrenia. These novel language metrics can capture both off-topic responses and incoherence in patients' speech by modeling speech information and fine-tuning techniques. Additionally, these metrics exhibit cultural sensitivity while providing a more comprehensive evaluation of linguistic abnormalities in schizophrenia.</p><p>This research fine-tuned the ELECTRA Pretrained Language Model on a 750 MB text corpus obtained from major Chinese mental health forums. The effectiveness of the fine-tuned language model is verified on a group comprising 38 individuals diagnosed with schizophrenia and 25 meticulously matched healthy controls. The study explores the association between the fine-tuned language model and the Positive and Negative Syndrome Scale (PANSS) items.</p><p>The results demonstrate that SPR is higher in healthy controls, indicating better language understanding by the pre-trained language model. Conversely, DF is higher in individuals with schizophrenia, indicating more inconsistent language structure. The relationship between linguistic features and P2 (conceptual disorganization) reveals that patients with positive P2 exhibit lower SPR and higher DF. Binary logistic regression using the combined SPR and DF features achieves 84.5 % accuracy in classifying P2, exceeding the performance of traditional features by 20.5 %. Moreover, the proposed linguistic features outperform traditional linguistic features in discriminating FTD (formal thought disorder), as demonstrated by multivariate linear regression analysis.</p></div>","PeriodicalId":21417,"journal":{"name":"Schizophrenia Research","volume":null,"pages":null},"PeriodicalIF":3.6000,"publicationDate":"2024-07-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Schizophrenia Research","FirstCategoryId":"3","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0920996424003219","RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"PSYCHIATRY","Score":null,"Total":0}
引用次数: 0
Abstract
This research presents two stable language metrics, namely Successful Prediction Rate (SPR) and Disfluency (DF), to objectively quantify the linguistic disturbances associated with schizophrenia. These novel language metrics can capture both off-topic responses and incoherence in patients' speech by modeling speech information and fine-tuning techniques. Additionally, these metrics exhibit cultural sensitivity while providing a more comprehensive evaluation of linguistic abnormalities in schizophrenia.
This research fine-tuned the ELECTRA Pretrained Language Model on a 750 MB text corpus obtained from major Chinese mental health forums. The effectiveness of the fine-tuned language model is verified on a group comprising 38 individuals diagnosed with schizophrenia and 25 meticulously matched healthy controls. The study explores the association between the fine-tuned language model and the Positive and Negative Syndrome Scale (PANSS) items.
The results demonstrate that SPR is higher in healthy controls, indicating better language understanding by the pre-trained language model. Conversely, DF is higher in individuals with schizophrenia, indicating more inconsistent language structure. The relationship between linguistic features and P2 (conceptual disorganization) reveals that patients with positive P2 exhibit lower SPR and higher DF. Binary logistic regression using the combined SPR and DF features achieves 84.5 % accuracy in classifying P2, exceeding the performance of traditional features by 20.5 %. Moreover, the proposed linguistic features outperform traditional linguistic features in discriminating FTD (formal thought disorder), as demonstrated by multivariate linear regression analysis.
期刊介绍:
As official journal of the Schizophrenia International Research Society (SIRS) Schizophrenia Research is THE journal of choice for international researchers and clinicians to share their work with the global schizophrenia research community. More than 6000 institutes have online or print (or both) access to this journal - the largest specialist journal in the field, with the largest readership!
Schizophrenia Research''s time to first decision is as fast as 6 weeks and its publishing speed is as fast as 4 weeks until online publication (corrected proof/Article in Press) after acceptance and 14 weeks from acceptance until publication in a printed issue.
The journal publishes novel papers that really contribute to understanding the biology and treatment of schizophrenic disorders; Schizophrenia Research brings together biological, clinical and psychological research in order to stimulate the synthesis of findings from all disciplines involved in improving patient outcomes in schizophrenia.