Linking language features to clinical symptoms and multimodal imaging in individuals at clinical high risk for psychosis

• DOI: 10.1192/j.eurpsy.2020.73
• Publication Date: 2020-08-11T05:48:50Z
• Authors: S. S. Haas; G. E. Doucet; S. Garg; S. N. Herrera; C. Sarac; Z. R. Bilgrami; R. B. Shaik; C. M. Corcoran
• Keywords: Psychiatry; Adult; Male; Brain Mapping; neuroimaging; multimodal; RC435-571; Brain; Prodromal Symptoms; Linguistics; Multimodal Imaging; Young Adult; 03 medical and health sciences; 0302 clinical medicine; Psychotic Disorders; Clinical high risk for psychosis; sparse canonical correlation analysis; Humans; Female; natural language processing; Research Article; Natural Language Processing

Abnormalities in the semantic and syntactic organization of speech have been reported individuals at clinical high-risk (CHR) for psychosis. The current study seeks to examine whether such abnormalities are associated with changes brain structure functional connectivity CHR individuals.Automated natural language processing analysis was applied samples obtained from 46 22 healthy individuals. Brain structural resting-state imaging data were also acquired all participants. Sparse canonical correlation (sCCA) used ascertain patterns covariation between linguistic features, symptoms, measures morphometry related network.In individuals, we found a significant mode features (r = 0.73; p 0.003), negative symptoms bizarre thinking covarying mostly complexity. In entire sample, separate sCCAs identified single linking 0.65; 0.05) network 0.63; 0.01). both models, covaried network. However, contribution diagnosis models negligible.Syntactic complexity appeared sensitive prodromal while brain-language seemed preserved. Further studies larger required establish reproducibility these findings.