Can Language Disturbances Act as a Biomarker for Schizophrenia?

Certain language disturbances may serve as biomarkers of schizophrenia pathology and symptomatology, according to study data published in Nature Partner Journals Schizophrenia. Language disturbances in patients with schizophrenia were significantly associated with negative symptoms, and white matter integrity in language tracts was predicted by quantifiable speech patterns.

The researchers sought to quantify speech disturbances in schizophrenia. Between 2015 and 2018, 26 patients with schizophrenia spectrum disorder (mean age, 26.7±5.4 years; 76.9% men) and 22 healthy controls (mean age, 24.3±4.4 years; 86.4% men) underwent language processing and magnetic resonance imaging (MRI) scans. All participants underwent a semistructured interview designed to elicit spontaneous speech. Interviewers asked a series of questions about “neutral” general life experiences, although emotional subject matter was avoided and patients were not asked to describe their health or quality of life. The researchers examined the following language measures: articulation rate, average pause duration, speaking turn duration, proportion of interview time spent speaking, mean length of utterance, lexical diversity, clauses per utterance, noun–verb ratio, open-closed ratio, and pause-to-word ratio.

Schizophrenia symptom severity was evaluated with the Positive and Negative Syndrome Scale (PANSS). Diffusion tensor imaging (DTI) was performed, and fractional anisotropy (FA) and mean diffusivity (MD) maps were created to ascertain mean values for preselected regions of interest. Patient-control language differences were assessed using multivariate analysis of covariance. Logistic regression was performed to identify language-based predictors of schizophrenia symptoms and white matter aberrations.  

Compared with controls, patients spoke more slowly, spoke during a smaller proportion of the interview, produced shorter utterances, had lower lexical diversity, and used fewer clauses per utterance (all P <.05). When mean length of utterance and clauses per utterance were used as explanatory variables, a logistic regression model predicted group status with 88.5% sensitivity and 81.8% specificity.

Among patients, articulation rate (P =.036), speaking turn duration (P =.033), proportion of interview spent speaking (P <.001), and mean length of utterance (P =.047) were significantly and negatively associated with PANSS negative subscale scores. Patients and controls did not significantly differ on DTI measures of the language tracts and whole brain. However, patients had significantly decreased clusters of voxels in all regions of interest and the corpus callosum, cingulum and corona radiata. Language measures were highly explanatory of mean FA of the language tracts and mean MD of the whole brain in both patients (adjusted R² =.467) and controls (adjusted R² =.483).

The researchers noted that “language analyses are a promising tool in schizophrenia” since they are noninvasive, low cost, and performed quickly, and they concluded that the findings “make an important contribution to recent initiatives such as the Research Domain Criteria project.” However, the study may be limited by issues related to the influence of intelligence factors.

Overall, these results suggest that quantitative language measures may have utility as a biomarker of schizophrenia severity and symptomatology. In both patients and controls, quantified language aspects were highly predictive of the integrity of language-associated white matter tracts.

Reference

de Boer JN, van Hoogdalem M, Mandl RCW, et al. Language in schizophrenia: relation with diagnosis, symptomatology and white matter tracts. NPJ Schizophr. 2020;6(1):10.