Aberrant cerebellar-default-mode functional connectivity underlying auditory verbal hallucinations in schizophrenia revealed by multi-voxel pattern analysis of resting-state functional connectivity MRI data

Abstract

Past neuroimaging studies have reported that aberrant functional connectivity (FC) underlying auditory verbal hallucinations (AVHs) in schizophrenia is highly distributed over multiple functional networks. There is thus a need for exploratory approaches without limiting analysis to particular seed regions or functional networks, to identify FC alterations underlying AVH. We applied a multi-voxel pattern analysis (MVPA) of FC together with a series of post-hoc FC analyses to resting-state fMRI data acquired from 25 patients with schizophrenia and 25 matched healthy controls. First, the MVPA revealed multiple clusters exhibiting altered FC patterns in schizophrenia. Subsequent multiple linear regression analysis using scores of these clusters identified that FC alteration in the right cerebellum crus I was significantly associated with the severity of AVH. Furthermore, post-hoc FC analysis with the right crus I as a seed revealed significant FC alterations with regions distributed across multiple functional networks, including speech, default-mode, thalamus, and cerebellum. Subsequent linear regression analyses further demonstrated that, among these regions, only reduced FC in the left precuneus was significantly associated with the severity of AVH. Our unbiased exploratory analysis of FC data revealed a novel evidence for the crucial role of FC between cerebellar and default-mode networks in AVH.