Reality monitoring impairment in schizophrenia reflects specific prefrontal cortex dysfunction

Abstract

Reality monitoring impairment is often reported in schizophrenia but the neural basis of this deficit is poorly understood. Difficulties with reality monitoring could be attributable to the same pattern of neural dysfunction as other cognitive deficits that characterize schizophrenia, or might instead represent a separable and dissociable impairment. This question was addressed through direct comparison of behavioral performance and neural activity associated with reality monitoring and working memory in patients with schizophrenia and matched healthy controls. Participants performed a word-pair reality monitoring task and a Sternberg working memory task while undergoing fMRI scanning. Distinct behavioral deficits were observed in the patients during performance of each task, which were associated with separable task- and region-specific dysfunction in the medial anterior prefrontal cortex for reality monitoring and dorsolateral prefrontal cortex for working memory. The results suggest that reality monitoring impairment is a distinct neurocognitive deficit in schizophrenia. The findings are consistent with the presence of a range of dissociable cognitive deficits in schizophrenia which may be associated with variable functional and structural dysconnectivity in underlying processing networks.

Keywords: Cognitive dysfunction; Prefrontal cortex; Reality monitoring; Schizophrenia; Working memory; fMRI.

Fig. 1

Stimuli used in the reality monitoring and working memory tasks. The top panel shows stimuli used in the study and test phases of the reality monitoring task which employed a 2 × 2 design involving either the subject of researcher speaking aloud the stimuli, which were presented either complete (perceived) or incomplete (the second word to be imagined). The bottom panel shows stimuli used for load-three of the working memory task, the respond phase shows only three out of fourteen trials undertaken, one with a target and two with foils.

Fig. 2

Task order and functional run structure. The order of presentation of the three experimental tasks is shown for each of the six run orders, through which participants cycled. Task order was counterbalanced between participants by starting each participant with a different run order. RM = reality monitoring task, WM = working memory task, BASELINE = perceptual motor baseline task, SCANNER = scanner switch on. Scanned versions of the tasks are shown in bold text. Functional runs lasted from 482 to 700 s but scanning time was fixed at 370 s.

Fig. 3

Reality monitoring and working memory task performance. Error bars for all charts represent standard error.

Fig. 4

Neuroimaging results – group, region and task level dissociations. Percentage signal change for patients and controls in the reality monitoring and working memory contrasts, within the a priori left medial anterior PFC voxel: x = −16, y = 56, z = 14; right medial anterior PFC voxel: x = 20, y = 56, z = 9; left dorsolateral PFC voxel: x = −45, y = 25, z = 25; and right dorsolateral PFC voxel: x = 43, y = 38, z = 18, derived from previous studies.

Fig. 5

Patients with schizophrenia show reduced medial anterior PFC activity during reality monitoring, and reduced dorsolateral PFC activity during working memory. Left panel: coronal sections taken at z = 8 showing areas of greater activity associated with reality monitoring than perceptual motor baseline. Controls but not patients exhibited significant activity in the left medial anterior PFC (circled, peak: x = −15, y = 59, z = 8). A trend towards significant left hemisphere medial anterior PFC activity was observed in the group contrast of Controls > Patients (circled, peak: x = −15, y = 56, z = 8; p = 0.057). Right panel: coronal sections taken at z = 16 showing areas of greater activity associated with working memory load 5 than load 3. Controls but not patients exhibited significant bilateral activity in the dorsolateral PFC (peak: x = −39, y = 23, z = 28 and x = 39, y = 32, z = 16). Significantly greater dorsolateral PFC activity was observed in the group contrast of Controls > Patients (circled, peak: x = 39, y = 32, z = 14). Activity in all contrasts was significant at voxel-wise height threshold of p < 0.05, small-volume corrected, and is illustrated here for display purposes at p < 0.01, uncorrected.