Functional resting-state networks are differentially affected in schizophrenia

Abstract

Neurobiological theories posit that schizophrenia relates to disturbances in connectivity between brain regions. Resting-state functional magnetic resonance imaging is a powerful tool for examining functional connectivity and has revealed several canonical brain networks, including the default mode, dorsal attention, executive control, and salience networks. The purpose of this study was to examine changes in these networks in schizophrenia. 42 patients with schizophrenia and 61 healthy subjects completed a RS-fMRI scanning session. Seed-based region-of-interest correlation analysis was used to identify the default mode, dorsal attention, executive control, and salience networks. Compared to healthy subjects, individuals with schizophrenia demonstrated greater connectivity between the posterior cingulate cortex, a key hub of the default mode, and the left inferior gyrus, left middle frontal gyrus, and left middle temporal gyrus. Interestingly, these regions were more strongly connected to the executive control network in healthy control subjects. In contrast to the default mode, patients demonstrated less connectivity in the executive control and dorsal attention networks. No differences were observed in the salience network. The results indicate that resting-state networks are differentially affected in schizophrenia. The alterations are characterized by reduced segregation between the default mode and executive control networks in the prefrontal cortex and temporal lobe, and reduced connectivity in the dorsal attention and executive control networks. The changes suggest that the process of functional specialization is altered in schizophrenia. Further work is needed to determine if the alterations are related to disturbances in white matter connectivity, neurodevelopmental abnormalities, and genetic risk for schizophrenia.

Introduction

Dysconnectivity hypotheses of schizophrenia postulate that the disorder relates to abnormalities in neuronal connectivity (Bullmore et al., 1997, Andreasen et al., 1998, Friston, 1999, Stephan et al., 2009). Contemporary dysconnectivity theories posit that disturbed neural connectivity results from a combination of genetic and environmental risk factors that impinge upon normal neurodevelopment (Bullmore et al., 1997, Maynard et al., 2001, Karlsgodt et al., 2008). Reduced dendritic length and spine density, altered coherence in brain activity across cortical regions during task performance, and identification of schizophrenia susceptibility alleles and increased copy number variants in genes related to neuronal signaling and neurodevelopment are all consistent with dysconnectivity theories (Walsh et al., 2008, Lewis and Sweet, 2009, Woodward et al., 2009, Glessner et al., 2010, Uhlhaas and Singer, 2010).

Resting-state functional magnetic resonance imaging (fMRI) has revealed that spontaneous neural activity, inferred on the basis of blood-oxygen-level dependence (BOLD) response time-series data, correlates across brain regions and is organized into spatially segregated functional connectivity networks (e.g. Fox et al., 2005). In addition to basic sensory and motor networks, RS-fMRI has identified several, ‘higher-order’ resting-state networks (RSNs) consisting primarily of interconnections between heteromodal association cortical regions (Biswal et al., 1995, Biswal et al., 1997, Li et al., 2000, Greicius et al., 2003, Fox et al., 2005, Vincent et al., 2008). They include; 1) the well known default mode network (DMN), which is comprised of posterior cingulate cortex (PCC)/precunues, ventro-medial prefrontal cortex (vmPFC), lateral parietal cortex, and mesial temporal lobe structures; 2) a dorsal attention network (DAN) consisting of the intraparietal sulcus (IPS)/superior parietal lobule (SPL), frontal eye fields (FEF), and extrastriate visual areas (middle temporal: area MT+); 3) a dorsolateral prefrontal cortex (dlPFC)-parietal executive control network (ECN); and 4) the ‘salience’ network that includes inferior frontal gyrus/anterior insular cortex and the anterior cingulate. The relevance of resting-state connectivity to individual differences in behavior and neuropsychiatric disorders is an area of intense investigation. The DMN, DAN, and ECN have been linked to memory, attention, and executive cognitive functions, respectively (Hampson et al., 2006, Seeley et al., 2007, Carter et al., 2010, Wang et al., 2010a, Wang et al., 2010b). Consequently, these networks may be particularly relevant to schizophrenia given their association with cognitive functions known to be impaired in schizophrenia.

A number of studies have examined RSNs in schizophrenia, especially the DMN. There is strong evidence that the DMN is abnormal in schizophrenia; although, findings are mixed with reports of both increased connectivity between brain regions comprising the DMN (Zhou et al., 2007b, Whitfield-Gabrieli et al., 2009), and even expansion of the DMN to include additional brain regions (Mannell et al., 2010, Salvador et al., 2010, Skudlarski et al., 2010), and decreased connectivity (Bluhm et al., 2007, Camchong et al., in press, Rotarska-Jagiela et al., 2010). Considerably less is known about the integrity of other RSNs (Zhang et al., 2009, Seeley et al., 2007; Carter et al., 2010). Altered connectivity within a fronto-parietal network has been reported in several studies (Zhou et al., 2007a, Rotarska-Jagiela et al., 2010, Skudlarski et al., 2010); although one study did not find abnormal dlPFC connectivity in antipsychotic naïve first episode patients (Lui et al., 2009).

The lack of definitive conclusions may relate to the relatively small number of patients included in most studies (20 or fewer in many cases), limited data on networks other than the DMN, and the diversity of methods used to quantify connectivity. Moreover, it's unclear if some RSNs are differentially affected in schizophrenia as most studies focused on just one network. The purpose of this investigation was to examine the effects of schizophrenia on four canonical RSNs: the default mode, dorsal attention, executive control, and salience networks.