Spatial working memory among middle-aged and older patients with schizophrenia and volunteers using fMRI
Introduction
Goldman-Rakic and Selemon (1997) have argued that many of the symptoms of schizophrenia can be explained by deficits of working memory (WM), associated with a distributed network including prefrontal and parietal cortices, basal ganglia, and thalamus (e.g., Levy et al., 1997). Among prefrontal brain regions involved in WM, neuropsychological and functional brain imaging data suggest that disordered behavior in schizophrenia is associated with aberrant functioning in dorsolateral prefrontal cortex (DLPFC; Goldberg et al., 1988, Kolb and Whishaw, 1983, Weinberger et al., 1986) Because intact visuospatial WM function relies more on the DLPFC than other prefrontal regions Funahashi et al., 1989, Fuster and Alexander, 1971, studies of spatial WM deficits in schizophrenia are especially relevant tests of the DLPFC hypothesis.
Many behavioral studies have demonstrated that patients with schizophrenia are impaired on visuospatial WM tasks since the seminal study by Park and Holzman (1992) Fleming et al., 1997, Harvey et al., 1995a, Harvey et al., 1995b, Keefe et al., 1995, Keefe et al., 1997, Park, 1999, Park and Holzman, 1993, Park et al., 1999, Salame et al., 1998. Performance on spatial WM is inversely related to increased genetic risk for schizophrenia (Cannon et al., 2000). Poorer performance on WM tests also predicts poorer community outcome and impaired skill learning (Liddle, 2000). Thus, visuospatial WM appears to be a domain of neuropsychological function especially suited to bridge basic and clinical neuroscience research in schizophrenia.
Although WM is consistently found to be impaired in schizophrenia patients, aberrant activation in DLPFC has been inconsistently found in functional imaging studies among younger patients with schizophrenia Artiges et al., 2000, Barch et al., 2001, Berman et al., 1986, Berman et al., 1992, Callicott et al., 1998, Manoach et al., 2000, Manoach et al., 2001, Pochon et al., 2001. Some studies have found diminished activation of the DLPFC during WM performance among individuals with schizophrenia Berman et al., 1986, Callicott et al., 1998, Stevens et al., 1998, whereas other studies have reported normal activation Honey et al., 2002, Manoach et al., 1999 or hyperactivation (Manoach et al., 2000). The interpretation of these findings is complicated by the wide variety of tasks used to probe WM. WM probes studied include a verbal n-back task, the Sternberg Memory Scanning Paradigm, a graded random number generation task, and the Wisconsin Card Sorting Task (WCST; Berg, 1948, Grant and Berg, 1948, Heaton et al., 1993). These tasks vary by difficulty; the WM subprocess they measure; the representational systems needed to store the presented material (e.g., verbal vs. spatial); the sensory modality stimulated; the extent to which responses were self-generated; and the type and number of non-WM cognitive processes involved. For example, the pioneering positron emission tomography (PET) studies from Berman and colleagues employed the WCST as their WM measure. This task is sensitive to frontal system dysfunction Milner, 1963, Milner, 1964; its principal neuropsychological use is to assess a subject's capacity for strategic cognitive flexibility. The WCST is complex, requiring integration of visual feedback from the cards themselves and auditory feedback from the examiner in addition to the WM component of keeping both of these continuously updated pieces of information in mind for several seconds. Although the WCST relies on WM for successful performance, it also tests other cognitive operations (see also Konishi et al., 1999).
Task difficulty has also been shown to be an important variable in accounting for differences in functional imaging studies of WM among individuals with schizophrenia. Patients with schizophrenia are not impaired on every WM task. Many perform as well as controls on easy tasks (Callicott et al., 2000, Goldberg et al., 1993; see also Kolb and Whishaw, 1983, McKenna et al., 1994). Imaging studies find that patients' DLPFC function on easy tasks can be in the normal range (Honey et al., 2002), whereas on more difficult tasks, as performance suffers, DLPFC activity may increase Callicott et al., 2000, Manoach et al., 1999 or decrease Callicott et al., 2000, Carter et al., 1998, Kuperbergo and Heckers, 2000. Using a graded memory task, Fletcher et al. (1998) found that patients with schizophrenia show robust activation in the DLPFC up to the point where the task becomes so difficult that their performance becomes impaired.
Even when functional imaging studies of schizophrenia patients performing WM tasks report aberrant results, the regions involved vary. For example, Callicott et al. (1998) found impaired activation of the DLPFC during WM when schizophrenia and control groups were matched for signal variance. Others, however, observed disordered activation of schizophrenic patients in the basal ganglia and thalamus (Manoach et al., 2000). Furthermore, some studies find reduced activation in one region but abnormally increased activation in others (Callicott et al., 1998).
Although the human and nonhuman primate studies providing the neuroanatomical foundation for Goldman-Rakic and Selemon's (1997) focus on the DLPFC in schizophrenia involved spatial WM, none of the functional imaging studies described above investigated the brain activation of schizophrenia patients during spatial WM performance. Thus, none was a direct test of the prediction of dysfunction of the spatial WM network in schizophrenia. McCarthy et al. (1994) developed a spatial WM task to extend the nonhuman primate spatial WM studies of Goldman-Rakic and colleagues to healthy human volunteers. Using functional magnetic resonance imaging (fMRI) to investigate spatial WM in humans in regions of interest (ROIs) from a single slice, McCarthy found task-related activity in right and left middle frontal gyrus and anterior cingulate.
The aim of the present study was to localize differences in brain function between schizophrenia patients and volunteers using a spatial WM task adapted from McCarthy et al. (1994). We surveyed the entire brain because the task had not been used before as the basis of a voxel-wise analysis. We were interested in learning what other areas might display aberrant activity and what areas of differential activity might correlate with the patients' symptoms. Based on findings from nonhuman primate studies and from McCarthy's study, we predicted that the spatial WM task would activate left and right DLPFC (BA 46), the parietal cortex (BA 7/40), and anterior cingulate when performed by controls. Given the impairment of spatial WM found among individuals with schizophrenia, we expected aberrant activation among schizophrenia patients in the spatial WM network activated in healthy participants. We especially anticipated group differences in spatial WM activation in the DLPFC.
Section snippets
Subjects
From the Interventions Research Center (IRC) at the University of California, San Diego, we recruited middle-aged and older community-dwelling, right-handed patients with schizophrenia or schizoaffective disorder [n=8 and n=2, respectively; age (mean±S.D.)=58.0±7.73] and healthy volunteers (n=12, age=63.8±10.19), obtaining informed consent from each subject after explaining the study. The study was conducted in accordance with institutional review board guidelines. Demographic information is
Demographic and behavioral data
Although the groups were similar in age, the patients were significantly less well educated than the volunteers (Table 1A). Because one patient was also receiving a short-acting benzodiazepine, we performed all the two-group image analyses including and excluding this patient, with no differences in results. Those data were included in what follows.
Reaction time and d′ data for each group are shown in Table 1D. To ascertain whether there were significant differences in discriminability, we
Discussion
We found little support for abnormal brain activity in the DLPFC during spatial WM using the AV baseline among patients with schizophrenia when the task is substantially equally difficult for patients and controls. We further supported this result by comparing the spatial WM condition against the PV baseline (mitigating somewhat the ambiguity of interpretation of negative response). Thus, the absence of abnormal brain activity in the DLPFC when schizophrenia patients are performing a spatial WM
Acknowledgements
We wish to acknowledge the comments of Monte Buchsbaum, MD, on an earlier version of this manuscript. This research is supported in part by the Mental Illness Research, Education and Clinical Center of the Department of Veterans Affairs VISN 22; MH01768; MH49671; MH19934-07VA; and Merit Grant 304.
References (84)
Hypofrontality in schizophrenia: distributed dysfunctional circuits in neuroleptic-naive patients
Lancet
(1997)Functional magnetic resonance imaging brain mapping in psychiatry: methodological issues illustrated in a study of working memory in schizophrenia
Neuropsychopharmacology
(1998)The inheritance of neuropsychological dysfunction in twins discordant for schizophrenia
Am. J. Hum. Genet.
(2000)AFNI: software for analysis and visualization of functional magnetic resonance neuroimages
Comput. Biomed. Res.
(1996)Regional frontal abnormalities in schizophrenia: a quantitative gray matter volume and cortical surface size study
Biol. Psychiatry
(2000)Entorhinal cortex pre-alpha cell clusters in schizophrenia: quantitative evidence of a developmental abnormality
Biol. Psychiatry
(2000)Visuospatial working memory in patients with schizophrenia
Biol. Psychiatry
(1997)The neurocognitive effects of low-dose haloperidol: a two-year comparison with risperidone
Biol. Psychiatry
(2002)Cognitive functioning in chronically hospitalized schizophrenic patients: age-related changes and age disorientation as a predictor of impairment
Schizophr. Res.
(1995)De-coupling of cognitive performance and cerebral functional response during working memory in schizophrenia
Schizophr. Res.
(2002)