Discriminating imagined from perceived information engages brain areas implicated in schizophrenia

Abstract

Some of the symptoms of schizophrenia may reflect a difficulty discriminating between information that was perceived from the outside world and information that was imagined. This study used fMRI to examine the brain regions associated with this reality monitoring ability in healthy volunteers, who recollected whether information had previously been perceived or imagined, or whether information had been presented on the left or right of a monitor screen. Recent studies have suggested that schizophrenia may be associated particularly with dysfunction in medial anterior prefrontal cortex, thalamus, and cerebellum. In our data, activation in all three of these regions of interest was significantly greater during recollection of whether stimuli had been perceived or imagined versus recollection of stimulus position. In addition, reduced prefrontal activation was associated with the same misattribution error that has been observed in schizophrenia. These results indicate a possible link between the brain areas implicated in schizophrenia and the regions supporting the ability to discriminate between perceived and imagined information.

Introduction

How do we distinguish events we imagined from those that we actually witnessed? A rich seam of theoretical and empirical work from cognitive psychology has led to a number of cognitive models of functions that have been termed reality monitoring and source monitoring, which may contribute to discriminating between imagined and perceived information (Johnson et al., 1993, Burgess and Shallice, 1996a, Schacter et al., 1998). In terms of the neural mechanisms that might support these processes, theorists have made great strides in understanding how the brain represents in memory information that was perceived from the outside world (Aggleton and Brown, 1999, Fletcher and Henson, 2001, Simons and Spiers, 2003), but less is known about the brain regions involved in remembering previously imagined information. An unresolved issue concerns which brain regions might be involved in discriminating between perceived and imagined information—a control function that is critical to feeling that we can trust our memories to supply us with true recollections and shield us from falsely remembering events that never occurred.

An indication of what can happen when such an ability breaks down may come from schizophrenia. Although the disorder can vary in its presentation, among the positive symptoms often reported in schizophrenia are hallucinations, whereby patients report, for example, hearing voices when none are present (Frith, 1992). It has been suggested that these symptoms may result from a difficulty in discriminating between information that is perceived in the external world and information that is imagined (Johnson and Raye, 2000), because of a deficit in monitoring the self-generation of thoughts (Frith and Done, 1989, Frith, 1992) or a bias towards misattributing internal thoughts to external sources (Bentall et al., 1991). One possibility is that this represents an intrusion of previous thoughts and experiences into current consciousness. In support of this view, schizophrenic patients with delusions and hallucinations have been shown to be impaired at judging the origin of previously encoded stimuli, often exhibiting particular difficulty in identifying the source of self-generated information (Bentall et al., 1991, Vinogradov et al., 1997, Danion et al., 1999, Keefe et al., 2002). For example, both Vinogradov et al. (1997) and Keefe et al. (2002) have reported that patients with schizophrenia were impaired at remembering the source of words they had generated themselves but performed similarly to controls at remembering the source of words they had seen or heard.

It is unclear which brain regions might be implicated in the discrimination difficulties observed in schizophrenia. Evidence from neuroimaging studies suggests that schizophrenia may be associated with changes in a number of brain regions, including lateral and medial prefrontal and temporal cortices, the thalamus, and cerebellum (Friston and Frith, 1995, Fletcher et al., 1996, Lawrie et al., 2002, Winterer et al., 2003). However, it has proved difficult to isolate the brain regions that might be responsible specifically for an impairment in discriminating between perceived and imagined information because patients with schizophrenia often exhibit additional cognitive symptoms impacting on language, perception, attention, and problem solving (Frith, 1992, Cornblatt and Keilp, 1994, Hoff and Kremen, 2003), and, in addition, many are on anti-psychotic medication at the time of testing, known to alter the functioning of multiple brain systems (Goff and Coyle, 2001). One possible solution has been to also study unmedicated first degree relatives of schizophrenic patients, who exhibit none of the cognitive impairments associated with schizophrenia but have an enhanced genetic risk of developing the disorder (Risch, 1990), and who can, to varying extents, experience delusions or hallucinations (Callicott et al., 2003, Whalley et al., 2004).

Inspection of a number of functional neuroimaging studies involving these patient groups reveals consistent activation reductions relative to controls in three main regions: medial anterior prefrontal cortex, thalamus, and cerebellum (Andreasen et al., 1996, Callicott et al., 2003, Whalley et al., 2004, MacDonald et al., 2005, Fox et al., 2005). These activation reductions occurred reliably across groups of patients with schizophrenia as well as symptomatic and asymptomatic first degree relatives, suggesting that they are not specific to one or other participant group. In addition, dysfunction in these three regions was observed during performance of a range of cognitive tasks (story recall (Andreasen et al., 1996), n-back working memory (Callicott et al., 2003, Fox et al., 2005), sentence completion (Whalley et al., 2004), and context processing (MacDonald et al., 2005)), suggesting that the effects are independent of the task being administered. Moreover, in a number of studies (Callicott et al., 2003, Whalley et al., 2004, Fox et al., 2005), the functional disruptions occurred in the context of normal behavioral performance, suggesting that the effects were not merely secondary to poor performance on the task. A correspondence, therefore, between dysfunction in these brain regions and those involved in discriminating between perceived and imagined information may suggest an exploratory thread which links evidence from three necessary levels: evidence of dysfunction at a neuroanatomical level, cognitive substrate (or “marker”), and behavioral symptoms.

In this event-related functional magnetic resonance imaging (fMRI) study, healthy participants were scanned while recollecting whether information had previously been perceived or imagined, or recollecting which position on a monitor screen information had been presented. Our analysis examined activation changes in each of the regions identified as functionally disrupted in schizophrenics and first degree relatives (Andreasen et al., 1996, Callicott et al., 2003, Whalley et al., 2004, MacDonald et al., 2005, Fox et al., 2005), with regions of interest defined on the basis of coordinates reported in the study by Whalley et al. (2004). The coordinates reported by Whalley et al. were chosen because that experiment involved the most extensive target group of any of the studies (69 first degree relatives of patients with schizophrenia, both with and without psychotic symptoms), none of whom were on anti-psychotic medication at the time of testing and who, as a group, showed no behavioral differences relative to controls on the administered task (the Hayling sentence completion test (Burgess and Shallice, 1996b)).

If hallucinations can be attributed to difficulty in discriminating information present in the outside world from information that is imagined, and medial anterior prefrontal cortex, thalamus, and cerebellum are the neural substrates of this ability, then we would predict that activation in all three of these regions of interest will be significantly greater in healthy volunteers during recollection of whether items had been perceived or imagined versus recollection of the items' position. Moreover, reduced activation may be associated specifically with misattributing imagined items as having been perceived, the type of error that has been observed in schizophrenia (Vinogradov et al., 1997, Danion et al., 1999, Keefe et al., 2002).