Schizophrenia is a chronic mental illness characterized by psychotic symptoms of delusions and hallucinations against a background of neurocognitive impairment, poor motivation and poor psychosocial functioning (Heinrichs&Zakzanis, 1998; Wong & Van Tol, 2003). Cognitive deficits appear before psychotic symptom onset (Cornblatt et al., 1999; Fuller et al., 2002), are not substantially improved by antipsychotic treatment (Keefe et al., 2007; Woodward et al., 2005), and are a significant predictor of functional outcome (Green, 1996; Harvey et al., 1998). Moderate to severe deficits are reported in a number of cognitive domains including memory, attention, and executive function (Heinrichs&Zakzanis, 1998). These deficits are a major contributor to the globally high loss-of-productivity costs of schizophrenia (Knapp et al., 2004) and estimated employment rate of only 10% among working-age individuals (Evensen et al., 2016).

Building on prior work that focused on more discrete forms of cognition in schizophrenia spectrum disorders (SSDs), we harnessed computer simulations to develop a naturalistic city environment to assess complex, goal-directed navigation (Zawadzki et al., 2013). Our single-trial paradigm is analogous to performing daily activities such as going to a medical centre and later trying to relocate a shop you had spotted on your way there. On this task, individuals with SSDs had difficulties in wayfinding, measured as distance travelled to find targets previously encountered within the virtual city. They were also more likely to miss targets presented during a passive viewing phase, less likely to generate novel shortcuts to targets, and more likely to become lost. Moreover, navigation performance on this ecologically valid task was at least as sensitive as standardized cognitive measures and strongly correlated with psychosocial functioning (Zawadzki et al., 2013).

Here we apply the same task in combination with functional magnetic resonance imaging (fMRI) to assess the activation of key brain areas supporting successful navigation. Goal-directed navigation offers a powerful paradigm for studying neural system interactions during complex human behaviours (Spiers& Maguire, 2006; Wolbers&Hegarty, 2010). Several core brain regions are involved in successful goal-directed navigation, including the hippocampus, caudate, and prefrontal cortex (Bohbot&Corkin, 2007; Bohbot et al., 2004; Doelleret al., 2008; Maguire et al., 1998; Moser et al., 2008; O'Keefe &Dostrovsky, 1971; Packard &McGaugh, 1996), are also involved in the pathobiology of SSDs (Csernansky et al., 2002; Ebdrup et al., 2011; Folley et al., 2010; Glahn et al., 2005; Kegeles et al., 2010; Ledoux et al., 2013; Nelson et al., 1998; McClure et al., 2013; Ragland et al., 2009). Optimum navigation is associated with the ability to flexibly switch between hippocampal-dependent brain systems supporting allocentric spatial maps and striatal networks supporting egocentric and stimulus-response based learning (e.g., "route-following") (Hartley et al., 2003; Iaria et al., 2003), and to integrate wayfinding networks with prefrontal working memory (Hanlon et al., 2012; Wolbers et al., 2007) and executive functions (Hartley et al., 2003; Ledoux et al., 2014; Maguire et al., 1998).

Based on the literature and our previous finding that individuals with SSDs tend to take an egocentric route-following approach typically associated with the caudate (Zawadzki et al., 2013), we expected to find hyperactivity in the striatum and hypoactivation of the hippocampus and dorsolateral prefrontal cortex (DLPFC) during our navigation task. In addition to replicating the behavioural results, we further investigated strategies that may underlie performance deficits in SSD.