Cerebellar Functional Dysconnectivity in Drug-Naïve Patients With First-Episode Schizophrenia

Abstract

Background: Cerebellar functional dysconnectivity has long been implicated in schizophrenia. However, the detailed dysconnectivity pattern and its underlying biological mechanisms have not been well-charted. This study aimed to conduct an in-depth characterization of cerebellar dysconnectivity maps in early schizophrenia.

Study design: Resting-state fMRI data were processed from 196 drug-naïve patients with first-episode schizophrenia and 167 demographically matched healthy controls. The cerebellum was parcellated into nine functional systems based on a state-of-the-art atlas, and seed-based connectivity for each cerebellar system was examined. The observed connectivity alterations were further associated with schizophrenia risk gene expressions using data from the Allen Human Brain Atlas.

Study results: Overall, we observed significantly increased cerebellar connectivity with the sensorimotor cortex, default-mode regions, ventral part of visual cortex, insula, and striatum. In contrast, decreased connectivity was shown chiefly within the cerebellum, and between the cerebellum and the lateral prefrontal cortex, temporal lobe, and dorsal visual areas. Such dysconnectivity pattern was statistically similar across seeds, with no significant group by seed interactions identified. Moreover, connectivity strengths of hypoconnected but not hyperconnected regions were significantly correlated with schizophrenia risk gene expressions, suggesting potential genetic underpinnings for the observed hypoconnectivity.

Conclusions: These findings suggest a common bidirectional dysconnectivity pattern across different cerebellar subsystems, and imply that such bidirectional alterations may relate to different biological mechanisms.

Keywords: cerebellum; genetic risk; hyperconnectivity; hypoconnectivity; schizophrenia.

Fig. 1.

Group main effects across all cerebellar systems. (A) Illustration of the nine cerebellar systems as defined in Ji et al. (B) Increased connectivity was shown between the cerebellum and the sensorimotor area, ventral part of visual cortex, default-mode regions, insula, and striatum in patients compared with controls, while decreased connectivity was present chiefly within the cerebellum, as well as the lateral prefrontal cortex, temporal lobe, and parietal lobe. (C) No significant group by seed interactions were observed, with the identified dysconnectivity present in all cerebellar seeds. Bar graphs indicate mean connectivity of regions in panel B, and error bars indicate standard error. SM, sensorimotor; AUD, auditory; VIS, visual; ATT, attention; CON, cingulo-opercular; DMN, default-mode; FPN, fronto-parietal; LAN, language; MM, multimodal.

Fig. 2.

Correlation between cerebellar-angular hyperconnectivity and PANSS positive symptom scores in patients, after controlling for age, sex, and head motion.

Fig. 3.

Associations between schizophrenia risk gene expression and cerebellar dysconnectivity. Each dot in the scatter plots represents a brain region in table 2 (the corresponding region indexes were labeled). Left panels: No significant correlations were shown across hyperconnected regions in either patients or controls. Right panels: Significant negative correlations were shown across hypoconnected regions in both patients and controls, as evidenced from 10 000 permutations. The dashed lines in the permutation distribution maps indicate the position of the observed correlation coefficients between risk gene expression and cerebellar dysconnectivity.