Emerging evidence suggests schizophrenia to involve widespread alterations in the macroscale wiring architecture of the human connectome. Recent findings of attenuated connectome alterations in unaffected siblings of schizophrenia patients suggest that altered connectome organization may relate to the vulnerability to develop the disorder, but whether it relates to progression of illness after disease onset is currently unknown. Here, we examined the interaction between connectome structure and longitudinal changes in general functioning, clinical symptoms and IQ in the 3 years following MRI assessment in a group of chronically ill schizophrenia patients. Effects in patients were compared to associations between connectome organization and changes in subclinical symptoms and IQ in healthy controls and unaffected siblings of schizophrenia patients. Analyzing the patient sample revealed a relationship between structural connectivity—particularly among central ‘brain hubs’—and progressive changes in general functioning (p = 0.007), suggesting that more prominent impairments of hub connectivity may herald future functional decline. Our findings further indicate that affected local connectome organization relates to longitudinal increases in overall PANSS symptoms (p = 0.013) and decreases in total IQ (p = 0.003), independent of baseline symptoms and IQ. No significant associations were observed in controls and siblings, suggesting that the findings in patients represent effects of ongoing illness, as opposed to normal time-related changes. In all, our findings suggest connectome structure to have predictive value for the course of illness in schizophrenia.
Schizophrenia's etiology has long since been related to alterations in the wiring architecture of the brain's network (Stephan et al., 2009, Rubinov and Bassett, 2011, Van den Heuvel and Kahn, 2011, Fornito et al., 2012, Van den Heuvel and Fornito, 2014 and Wheeler and Voineskos, 2014). A comprehensive map of the white matter pathways connecting disparate areas of the human brain is referred to as the macroscale connectome (Hagmann, 2005 and Sporns et al., 2005). Emerging evidence on connectome structure in schizophrenia suggests disease-related changes to include affected neural communication, aberrant local organization and modular structure and a less central position of brain hubs (Bassett et al., 2008, Lynall et al., 2010, Skudlarski et al., 2010 and Van den Heuvel et al., 2010). These putative brain hubs have been suggested to reside in multimodal association areas of the cortex, to participate in complex and diverse neuronal communication (Rubinov and Bullmore, 2013, Van den Heuvel and Sporns, 2013, De Reus and Van den Heuvel, 2014 and Senden et al., 2014) and to be mutually connected into a core collective referred to as a ‘rich club’ (Van den Heuvel and Sporns, 2011 and Van den Heuvel et al., 2012). The white matter pathways comprising this central communication system have been suggested to be disproportionally affected in schizophrenia (Van den Heuvel et al., 2013). Moreover, unaffected siblings of patients to show similar, though attenuated, effects (Collin et al., 2014). Such findings of connectome alterations in first-degree relatives (Repovs et al., 2011, Fornito et al., 2013 and Collin et al., 2014), who are at increased genetic risk for schizophrenia but lack the potential impact of (untreated) psychosis (Cahn et al., 2009) and psychotropic medication (Nejad et al., 2012 and Vita et al., 2012), have led to the hypothesis that affected connectome organization might be reflective of an inherited neurodevelopmental vulnerability to the disorder (Collin and Van den Heuvel, 2013, Skudlarski et al., 2013 and Van den Heuvel and Fornito, 2014).
Cross-sectional investigations of brain network organization in relation to illness severity in schizophrenia have suggested global and local network efficiency to be related to severity of both positive (Wang et al., 2012) and negative (Yu et al., 2011 and Wang et al., 2012) symptoms. In addition, reduced levels of functional network cost-efficiency have been associated with poorer working memory performance (Bassett et al., 2009). An open question regarding connectome abnormalities in schizophrenia (Dauvermann et al., 2014)—altered hub connectivity in particular (Van den Heuvel and Kahn, 2011)—is whether, and if so how, alterations in macroscale connectome wiring relate to illness progression and outcome. Persistent symptoms (Lieberman, 1999) and real-world deficits in areas such as employment (Harvey and Velligan, 2011) and everyday living (Harvey et al., 2009 and Leifker et al., 2009) are common in patients, but prognosis at the individual level is heterogeneous (Schultz and Andreasen, 1999). Relating connectome architecture to progression of illness and functional deficits might inform prognostic estimations. In this longitudinal study, a group of schizophrenia patients, investigated previously in two cross-sectional connectome studies (Van den Heuvel et al., 2013 and Collin et al., 2014), was reassessed after 3 years follow-up. Changes over time in general and intellectual functioning and clinical symptoms were evaluated and related to connectome structure at baseline. Particular emphasis was placed on examining the predictive value of measures of connectome topology (e.g., clustering, global efficiency and rich club organization) in terms of illness progression in the 3 years following MRI assessment.
