Abstract The present study aims to explain how to use the precision nomothetic approach to analyze the interconnections between the negative symptoms, cognitive dysfunctions and biomarkers of schizophrenia. We review our data obtained in different study groups of patients with (deficit) schizophrenia and show, using examples extracted from these studies, how Partial Least Squares (PLS) path analysis should be used to examine these complex associations. PLS path analysis combines factor and multiple regression analysis in mediated models. We show that a single latent trait can be extracted from negative symptom domains and psychosis, hostility, excitation, mannerism, formal thought disorders and psychomotor retardation (PHEMFP). Both the negative and PHEMFP concepts miss discriminant validity whilst a common latent construct may be extracted from the 6 negative and 6 PHEMFP subdomains, dubbed overall severity of schizophrenia (OSOS). A common latent factor may be extracted from neurocognitive test scores including executive functions, and semantic and episodic memory dubbed the general cognitive decline (G-CoDe) index. PLS analysis shows that the effects of neuroimmunotoxic pathways on OSOS are partly mediated by the G-CoDe and indicate that those pathways have also direct effects on OSOS. We explain that the intercorrelations between those features should be assessed in an unrestricted study group combining patients and controls. Moreover, further bifactorial factor analysis with the restricted schizophrenia group may disclose illness-specific covariations among the features. Machine learning discovered a new schizophrenia phenotype characterized by increased severity of AOPs, G-CoDe, and OSOS, dubbed “major neurocognitive psychosis”.
Discussion(0)
No comments yet. Be the first to comment.