Transdiagnostic types of formal thought disorder and their association with gray matter brain structure: a model-based cluster analytic approach

Abstract

Formal thought disorder (FTD) is a complex syndrome affecting language and thought processes in psychotic and affective disorders. Clustering (i.e., identification of data-driven clinical subtypes) establishes latent (sub-) structures into psychopathological syndromes. A latent profile analysis (LPA) of FTD symptoms was conducted in 1 032 patients diagnosed with Schizophrenia-Spectrum-Disorders (n = 107), Major Depressive (n = 800), and Bipolar Disorder (n = 125). Clusters were compared for cognition and psychopathology. Associations with gray matter volume (GMV) and cortical surface (gyrification, cortical complexity, sucal depth) were explored using T1-weighted MRI data, analyzed with CAT12. Robustness-analyses in an age- and sex-matched subsample (n = 321) with the same n for each diagnosis (n = 107) were applied. LPA revealed 4 transdiagnostic clusters: minimal FTD, poverty, inhibition, severe FTD that remained stable in an age- and sex-matched subsample and in each diagnosis separately. Patients exhibiting severe FTD compared to minimal FTD showed GMV reductions in the right superior and middle frontal gyri. Inhibition showed a GMV reduction in the right inferior and middle temporal gyri, and fusiform gyrus compared with minimal and severe FTD. Sulcal depth was reduced around the left insula, superior temporal sulcus and temporal pole in the poverty cluster, and in the bilateral insula in the severe cluster, both compared to the inhibition cluster. No results for cortical thickness, gyrification, and complexity were found. Results from the total sample could be replicated in the matched subsample. Our results unravel the clinical heterogeneity of FTD psychopathology across affective and psychotic disorders. Associations of FTD clusters with neuroanatomical substrates imply language-related brain structures being involved in thought and language impairment.

Fig. 1. Four-cluster item profiles for FTD symptoms.

Note: Estimated standardized sample means of SANS, SAPS, YMRS and HAM-D items used as indicator variables in the LPA across clusters. Boxes encompass +/− 1SD. Cluster 1 (minimal FTD), cluster 2 (poverty), cluster 3 (inhibition), cluster 4 (severe FTD).

Fig. 2. Distribution of diagnosis and clinical status across identified FTD clusters.

Note: A Relative distribution of diagnoses. Diagnoses were distributed across all 4 clusters with MDD patients constituting the largest proportion of every cluster except for cluster 4 (severe FTD), where SSD patients outnumbered all other patient groups, while MDD patients composed the smallest percentage. B Acute/remitted patients within latent clusters.

Fig. 3. Associations of FTD clusters with gray matter volume.

Note: Significant GMV clusters at p < 0.05 cluster-level FWE-corrected (initial cluster defining threshold of p < 0.0001). A Main effect of FTD cluster; B cluster 1 (minimal overall FTD) > cluster 4 (severe overall FTD); C cluster 3 (inhibition) > cluster 1 (minimal overall FTD); D cluster 3 (inhibition) > cluster 4 (severe overall FTD); FTD formal thought disorder, FWE Family-Wise-Error.

Fig. 4. Associations of FTD clusters with sulcal depth.

Note: Significant sulcal depth clusters at p < 0.05 cluster-level FWE-corrected (initial cluster defining threshold of p < 0.0001). A Main effect of FTD cluster; B cluster 3 (inhibition) > cluster 2 (poverty); C cluster 3 (inhibition) > cluster 4 (severe overall FTD); FTD formal thought disorder, FWE Family-Wise-Error.