Conceptual disorganization (CD), a core symptom of schizophrenia, significantly impacts real-world functioning. While previous research has linked aberrant connectivity to formal thought disorder (related to CD), no studies have examined the whole-brain functional correlates of CD in FEP. This study aimed to identify brain regions associated with CD severity in untreated FEPs compared to healthy controls (HCs). The researchers hypothesized that network-level dysconnectivity would be evident, potentially manifesting as reduced centrality in language processing regions and potentially compensatory increases in other areas in patients with high CD severity. The study also explores the potential difference between acute CD and the persistent form seen in chronic schizophrenia, the latter often showing more negative features and resisting treatment. Existing neuroimaging literature on thought disorders is reviewed, highlighting inconsistencies and a lack of studies directly relating network-level functional connectivity to disorganization, particularly in untreated FEPs. This study seeks to address this gap by employing ultra-high field 7T resting-state fMRI to identify brain regions with aberrant connectivity in relation to both the presence of psychosis and the severity of CD.

Systematic reviews have indicated aberrant connectivity in formal thought disorder (FTD), a construct related to CD, implicating language processing regions like the superior temporal gyrus (STG). However, findings have been inconsistent, with some studies reporting null results or implicating regions outside the language network. A limited number of resting-state fMRI (rs-fMRI) studies have examined network-level dysconnectivity in relation to disorganization in established schizophrenia, reporting mixed findings including reduced and increased connectivity in various brain regions. These studies lacked a voxel-by-voxel whole-brain approach and largely focused on medicated patients with chronic schizophrenia, limiting understanding of the neural basis of acute disorganization in FEP. This study aimed to bridge this gap by using ultra-high-field 7T rs-fMRI to perform a voxel-wise whole-brain investigation on untreated FEPs.

This study involved 38 untreated FEPs and 31 age-matched HCs. 7T rs-fMRI scans were acquired, and whole-brain binarized degree centrality (bDC) maps were computed. Group comparisons (FEP vs. HC, high CD vs. low CD within FEP) were performed using two-sample t-tests, with family-wise error (FWE) correction for multiple comparisons. Seed-based functional connectivity analyses were conducted to further characterize the identified regions' network interactions. The severity of symptoms was assessed using the Positive and Negative Syndrome Scale (PANSS), the Social and Occupational Functioning Assessment Scale (SOFAS), and a modified Digit Symbol Substitution Task (DSST). Patients were categorized into high and low CD groups based on PANSS item P2 scores. Regression analyses were used to test the specificity of findings to CD, controlling for other symptoms, cognitive function, and head motion. Data preprocessing steps included slice timing correction, motion correction, regression of head motion artifacts, white matter and CSF signal removal, linear detrending, standardization, and band-pass filtering. The DARTEL toolbox was used for spatial normalization to the MNI template. A 4 mm FWHM smoothing kernel was applied to the data. A correlation threshold of r > 0.25 was used to generate the adjacency matrix for degree centrality calculation. The study used a conservative cluster inclusion threshold of p < 0.001, with cluster-level significance set at FWE corrected p < 0.05. The researchers used both binarized degree centrality (bDC) and weighted degree centrality (wDC) but only reported bDC results in the main text, with wDC results shown in supplementary materials. They controlled for multiple testing using a family-wise error correction procedure with a conservative cluster inclusion threshold of p < 0.001 and a cluster-level significance set at FWE corrected p < 0.05. The specificity of the findings to P2 scores was tested through Spearman’s correlation and regression analysis.

The study revealed three key findings: 1. **Reduced centrality in the right STG/insula cluster in FEPs compared to HCs.** This cluster showed significant functional connectivity with distributed brain regions, including bilateral insula, middle cingulate cortex, supplementary motor area (SMA), thalamus, calcarine and lingual regions, and cerebellum. No significant increases in centrality were found in FEPs compared to HCs. 2. **Increased centrality in a medial superior parietal (mSPL) cluster in FEPs with high CD scores compared to those with low CD scores.** This cluster (paracentral lobule and precuneus) showed functional connectivity with sensorimotor regions, the middle cingulate cortex, SMA, and, to a lesser extent, language processing regions. 3. **Specificity of mSPL centrality to CD.** Regression analysis revealed that only the PANSS P2 (CD) score significantly predicted mSPL centrality, indicating the finding's specificity to CD and not other positive or negative symptoms or level of functioning.

The observed reduced centrality in the right STG/insula cluster in FEPs aligns with previous research in medicated schizophrenia patients. The lack of increased centrality in other regions in this early-stage sample suggests that the emergence of peripheral hubs may be a feature of chronicity or treatment effects. The increased centrality in the mSPL cluster in highly disorganized patients is intriguing. This region's connectivity to sensorimotor and language regions suggests that the emergence of this peripheral hub might be a compensatory mechanism, potentially leading to aberrant information processing and disorganized speech. This aligns with the concept of cortical reorganization and flexible redistribution of network activity following a pathological insult. The study suggests that distinct pathophysiological processes underlie different symptoms in psychosis. The results do not negate the role of language networks in thought disorders but highlight the importance of system-level dysconnectivity in determining FTD severity. The study’s findings are not directly comparable to those in studies of patients with persistent thought disorder. The increased activity observed in the medial sensory association cortex in this study and increased BOLD activity in ketamine challenge studies support this interpretation. Further research is needed to explore whether this increased centrality reflects a maladaptive compensatory mechanism or a disruption in network integration.

This study provides the first voxel-wise whole-brain investigation of acute CD in untreated FEPs using 7T rs-fMRI. The findings show reduced centrality in a core hub (rSTG/insula) and increased centrality in a peripheral region (mSPL) specifically associated with CD severity. This suggests that inefficient cortical information transfer resulting from the emergence of peripheral hubs might underlie CD pathogenesis. Future longitudinal studies are needed to track the symptom course and clarify the mechanisms underlying these compensatory changes. Further research is also necessary to explore the role of the mSPL in CD and investigate whether these findings generalize to other populations and diagnostic categories.

The study had a relatively small sample size, with a disproportionate number of male participants. The use of a specific analytical approach might influence the results. The assessment of thought disorder was not comprehensive, potentially limiting the interpretation of findings to the broader concept of formal thought disorder. The selection of participants using the PANSS-8 as the primary measure of conceptual disorganization might have led to a selection bias. Although head motion was controlled for and subjects with excessive head motion were excluded, residual effects of head motion might affect the results. The results might not be generalizable to other populations and diagnostic categories. Future studies should use a larger sample size, a more comprehensive assessment of formal thought disorder and include measures to test for selection bias.