Functional Neuroimaging Studies of Schizophrenia

21 Functional Neuroimaging Studies of Schizophrenia Sarah-J. Blakemore and Chris D. Frith Wellcome Department of Cognitive Neurology, University College London, London WCIN 3BG, United Kingdom

I. Introduction II. Neuropathology III. Structural Studies IV. Neuroreceptor Imaging of Antipsychotics Using PET V. Cognitive Activation Studies VI. Imaging Symptoms VII. Obstacles to Functional Neuroimaging and Schizophrenia Studies VIII. Conclusion References •

^

I. Introduction Schizophrenia is a complex mental illness characterized by acute phases of delusions, hallucinations, and thought disorder and, chronically, by apathy, flat affect, and social withdrawal. Schizophrenia affects 1% of the world's population, independent of country or culture and constitutes a severe public health issue (WHO, 1975). Schizophrenia was originally named "dementia Brain Mapping: The Disorders

praecox" by Emil Kraepelin in 1896 (Kraepelin, 1919), emphasizing its chronic and deteriorating course and separating it from psychoses with a good prognosis (which he called "maniacal-depressive insanity") and dementia in the elderly. Later, in 1911, Bleuler noted the fragmentation of personality and cognition that is fundamental to the illness and gave it its current label (schizo = split; phrenia = mind; Bleuler, 1987). This label encouraged the common misperception that schizophrenia is characterized by a split personality. In fact, schizophrenia is a complex illness comprising a combination of different symptoms, for example, auditory hallucinations, delusions, and incoherent speech. Schizophrenic people normally start to display symptoms in their late teens to early twenties, but the time of onset and the course of the illness are very variable (American Psychiatry Association, 1987). Approximately a third of patients experience one chronic episode after which they make a more or less full recovery; another third are affected by the illness throughout their lives but their symptoms are to some extent alleviated by anti-psychotic drugs; and the remaining third are so chronically ill that they show little or no improvement even with medication (Johnstone, 1991). 523

Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.

IV Psychiatric Disorders

524 A. General Intellectual Decline There is a striking deterioration in intellectual and cognitive function in schizophrenia, a feature emphasized by Kraepelin in his original description of the disorder as dementia praecox. Several recent studies have confirmed Kraepelin's observations showing that the IQ of schizophrenic patients is significantly lower than that of controls (e.g., Johnstone, 1991). In most studies, premorbid IQ of the patients and the controls was matched, suggesting that the difference in IQ was due to a rapid postmorbid intellectual decline in the patients.

B. Specific Psychological Impairments In addition to general intellectual decline, schizophrenic patients experience specific cognitive impairments. Findings demonstrating that intellectual impairment is more marked in some domains than others support the notion that circumscribed brain abnormalities may be associated with schizophrenia. Schizophrenic patients have difficulty in initiating and completing everyday tasks, being easily distracted and tending to give up when confronted by any obstacles. These deficits are similar to the problems in initiation, planning, and modification of behavior associated with frontal-lobe lesions (Shallice and Burgess, 1991). This has led many researchers to suggest that the core deficit of schizophrenia is a failure to activate frontal cortex appropriately during cognitive tasks involving planning and decision making ("task-related hypofrontality"), a notion supported in part by many functional imaging studies. In addition to planning deficits, schizophrenic patients also show deficits in attention and memory tasks that engage prefrontal, hippocampal, and medial temporal systems (Green, 1996), even in drug-free populations (Saykin et al, 1994). However, these observations do not directly address the causes of the disorder.

C. Etiology It is currently widely accepted that schizophrenia has a biological etiology. However, the shift toward this agreement is recent, and the etiology of schizophrenia has been the subject of lengthy and intense debate. The debate has been split between those who propose a biological etiology and those who postulate a psychodynamic origin of schizophrenia. In the latter camp, nonbiological factors such as family interaction and stressful life events (Kuipers and Bebbington, 1988), for example, have been proposed to play a causal role in the acquisition of schizophrenia. However, these theories have received httle empirical support. In addition.

in the past 50 years, two main lines of evidence supporting a biological role in its etiology have become apparent. First, there was the discovery of antipsychotic drugs in the 1950s, and second, a significant hereditary contribution to the disorder has been demonstrated. These observations strongly suggest that schizophrenia has a biological basis.

D. Dopamine Hypothesis In France in 1952, Delay, Deniker, and Harl found that dopamine-blocking drugs have a positive therapeutic effect on psychotic symptoms. This led in the 1960s to the dopamine (DA) hypothesis of schizophrenia, which posits that schizophrenia is caused by an overactivity of dopamine receptors (Van Rossum, 1966). This theory has received strong support since its conception: neuroleptic drugs produce some degree of antipsychotic or sedative action, or both, in most schizophrenic patients, and many studies have demonstrated a striking correlation between dosage of neuroleptic drug and its in vivo concentration for blocking DA receptors (Seeman, 1986). There are, however, some complications to the theory that have become apparent. First, there is a therapeutic lag. Whereas neuroleptics can alter the levels of DA activity on the first day of treatment, any distinct clinical improvement may take 10-50 days to appear (Straube and Oades, 1992). Second, the DA story is apparently more complicated, with some neuroleptics showing a similar but less marked blocking effect on the noradrenergic system (e.g., Iversen, 1986). In addition, many results on which the DA hypothesis is based have come from methodologies that are now beUeved to be problematic or confounded. In the 1960s and 1970s, techniques allowing the examination of cerebrospinal fluid (CSF) became available. However, CSF studies have produced weak or negative findings. Receptor binding studies have been performed on post-mortem brains. However, it has become clear that such studies will always be confounded by the effects of death, hypoxia, and neuroleptic treatment in life (Seeman, 1987). The most objective way to investigate the dopamine hypothesis is the in vivo visualization of radioactive ligand binding to quantify dopamine receptor densities in drugnaive patients using positron emission tomography (PET). These studies will be reviewed later in the chapter.

E. Genetic Studies That there is a significant genetic factor in schizophrenia is shown by studies demonstrating that monozygotic (MZ) twins (who share identical genes) have a 50% concordance rate for the illness (Gottesman and Shields, 1982). This finding supports the numerous stud-

21 Functional Neuroimaging Studies of Schizophrenia ies demonstrating that a schizophrenic person is more Hkely than a non-schizophrenic person to have a blood relation who is also schizophrenic (e.g., McGue et al, 1986). However, more recent studies have shown that the MZ twin concordance rate is lower than previously believed (e.g., 31%; Kendler and Robinette, 1983), and the lack of complete concordance between MZ twins indicates that non-genetic factors also play a part. In addition, the exact mode of inheritance is unclear. It is unknown whether what is inherited is a genetic predisposition to develop schizophrenia that is expressed in a direct biological way, or a vulnerabihty to express schizophrenic symptoms that depends on environmental or psychological factors. Thus, current methodologies present a challenge for research into the genetic and dopamine theories of schizophrenia. As a consequence, research has become increasingly focused on attempts to elucidate some structural or functional brain abnormality since it is widely held that schizophrenia is a disease of the brain (e.g., Ron and Harvey, 1990). The theory that some gross brain lesion characterizes schizophrenia seems unlikely. Instead, it is generally accepted that schizophrenia is characterized not by structural damage, but by functional abnormalities. This is supported by the relapsing and remitting course of the illness, fluctuations in symptoms, and response to pharmacological interventions. Therefore the advent of functional neuroimaging has been important in the study of schizophrenia because it enables brain function and its abnormahties to be investigated. As stated by Weinberger et al (1996): "Functional neuroimaging in psychiatry has had its broadest application and greatest impact in the study of schizophrenia." A major problem is that images of brain function reflect the current mental state of the patient (i.e., symptoms) and these are very variable. Symptoms include disorders of inferential thinking (delusions), perception (hallucinations), goal-directed behavior (avolition), and emotional expression. Current thinking generally distinguishes symptoms that comprise the presence of something that should be absent (positive symptoms) and the absence of something that should be present (negative symptoms) (Crow, 1980). Factor-analytic studies of symptoms suggest that positive symptoms should be subdivided further into a psychotic group (comprising delusions and hallucinations) and a disorganized group (comprising disorganized speech, formal thought disorder, disorganized behavior, and inappropriate affect) (Liddle, 1987). The original two-syndrome model proposed by Crow suggested that positive and negative symptoms might be due to a different pathophysiology. Positive symptoms. Crow suggested, were due to an overactivity of

525

the dopamine system, whereas negative symptoms were due to diffuse neuronal loss. Even though Crow's delineation of symptoms is rarely used by clinicians today, pathophysiology and abnormal psychology specific to the two symptom types have been supported by numerous studies, which will be addressed later in the chapter. The diversity of symptoms in schizophrenia makes it unlikely that the pathophysiology can be accounted for by a single localized brain dysfunction. Instead, the strategy of attempting to localize specific symptoms to specific brain regions or connections between regions is encouraging, and this chapter will evaluate the results of such studies.

II. Neuropathology Studies investigating structure in postmortem brains have shown evidence for a decrease in overall brain size (e.g.. Brown et al, 1986) and have linked schizophrenia to structural abnormalities in the prefrontal cortex and temporal lobe, especially the hippocampus and amygdala (Brown et al, 1986; Benes et al, 1991a). Several studies have found that schizophrenic brains tend to have enlarged lateral ventricles compared to nonschizophrenic brains (e.g.. Brown et al, 1986). The findings suggest that ventricular enlargement is associated with tissue loss in or abnormal development of the temporal lobes. Histological studies have shown evidence for abnormal synaptic function in the cingulate (Benes et al, 1991b; Honer et al, 1997) and hippocampal pyramidal cells (Benes etal, 1991a) in schizophrenia. The most reproducible positive anatomical finding in postmortem hippocampal formation has been the reduced size of neuronal cell bodies in schizophrenia. However, many studies have not revealed a clear change in hippocampal formation size, and recent studies of neuronal number and of cytoarchitecture have been largely negative. These inconsistent results might be due to the complexities surrounding this type of research, such as the effects of death, hypoxia, and neuroleptic treatment in life.

III. Structural Studies A. Computerized Tomography (CT) Studies The main finding from CT scan studies is that the lateral ventricles are enlarged in schizophrenic patients compared to normal controls. In the first study using this technique, Johnstone et al (1976) found that 17 chronically hospitalized schizophrenic patients had sig-

IV Psychiatric Disorders

526 nificantly enlarged ventricles compared to normal controls. In a review of the CT literature, Andreasen et al (1990) noted that 36 out of 49 subsequent studies have replicated this finding to some extent. There have also been two meta-analyses of the CT scan data (Raz and Raz, 1990; Van Horn and McManus, 1992), which supported the finding of enlarged ventricles in schizophrenic patients. However, the extent of ventricular enlargement in schizophrenia is often small and within the control range (e.g., Weinberger et al, 1979). Many further studies have found no evidence of ventricular enlargement. One problem contributing to the inconsistent results is the selection of controls. Smith and lacano (1986) compared 21 studies with equivocal results. They plotted the mean ventricle: brain ratio (VBR) for the schizophrenic patients and the normal controls in each study. Normal control brains in the studies with positive findings had, on average, smaller VBRs than control brains in the studies with negative findings. These results suggest that positive findings might be due to the control group having smaller ventricles rather than the schizophrenic patients having larger ventricles. In support of this. Van Horn and McManus (1992) found in their meta-analysis that choice of control group was a contributing factor to the variabihty of control VBR. Correlation between CT findings and symptoms has been investigated. Lewis et al (1990) reviewed 41 CT scan studies that addressed the issue of heterogeneity of schizophrenic symptoms. Only 1 of 18 studies found any association between increased ventricular area and chronicity of illness. Five out of 18 found evidence of a relationship with negative symptoms. Poor treatment response was found to be associated with increased ventricular area in about half the studies. The presence of tardive dyskinesia (in 4 out of 6 studies) and neuropsychological impairment (11 out of 14 studies) were also found to be associated with large ventricles. Chua and McKenna (1995) reviewed the CT literature and concluded that although the finding of increased ventricular area in schizophrenic patients is widespread and replicated, the difference between schizophrenic patients and normal controls is clearly small and depends significantly on the control group chosen. Overlap with the normal population is appreciable and might be complete if suitable controls were used (Smith and lacano, 1986).

B. MRI Studies MRI improves on CT and other scanning techniques due to its better spatial resolution and ability to differentiate white and gray matter. This allows the size of cortical and subcortical regions to be measured. There

have been several replicated findings using structural MRI scans to investigate the structure of schizophrenics' brains. Studies vary significantly in the MRI methodology used, earlier studies tending to use a single or small number of slices, whereas more recent studies have been able to scan multiple slices, or the whole brain. In most studies, regions of interest have had to be identified by drawing round them manually (Coffman and Nasrallah, 1986; Kelsoe et al, 1988). A number of studies have shown evidence for reductions in overall brain size, but most of these have used poor control groups and/or small numbers of schizophrenic patients (e.g., Andreasen et al, 1986; Harvey et al, 1993). In addition, this finding has not been replicated by other studies (DeMyer et al, 1988; Andreasen et al, 1990). Many MRI studies provide evidence that schizophrenic patients have larger lateral ventricles than normal controls (e.g., Coffman and Nasrallah, 1986; Kelsoe et al, 1988; Andreasen et al, 1990; Suddath et al, 1990; Woodruff et al, 1997a). However, the results are conflicting, and there are several negative findings (e.g.. Smith et al, 1987; Johnstone et al, 1989; Rossi et al, 1994). Older studies were limited by poor control groups and small number of schizophrenic patients. Recent studies have used larger subject groups and multiple control groups. Sharma et al (1998) compared MRI scans from 31 people with schizophrenia, 57 relatives, and 39 unrelated control subjects. Subjects with schizophrenia had larger lateral ventricles than both control groups. Relatives who were "presumed obligate carriers'' had larger left lateral ventricles than other relatives and the control subjects. Temporal lobe reductions in schizophrenia have been reported and are especially prominent in the hippocampus (Fukuzako et al, 1997), parahippocampal gyrus, and the amygdala (Yurgelun-Todd et al, 1996a). Suddath et al (1990) using the identical twins of the schizophrenic patients as controls found evidence for reductions of the left temporal lobe, including the hippocampus. Kwon et al (1999) recently replicated the finding of a reduction in volume of the left temporal lobe. Cannon et al (1998) suggested that frontal and temporal structural changes might reflect genetic (or shared environmental) effects. In a study using a large group of patients and well-matched controls (their nonpsychotic siblings and a group of normal controls), they found that volume reductions of the frontal and temporal lobes were present in patients with schizophrenia and in some of their siblings without schizophrenia. However, temporal lobe reductions are equivocal, with some negative findings in the literature (e.g.. Young et al, 1991). There have been many conflicting and negative results in studies attempting to locate clinical correlates with temporal abnormalities. Many studies have

527

21 Functional Neuroimaging Studies of Schizophrenia found no association between chronicity and temporal lobe size (e.g., Kelsoe et al, 1988; Young et al, 1991), although one study found an inverse relationship between these two factors (DeLisi et al, 1991). Size of the superior temporal gyrus has been associated with hallucinations (Barta et al, 1990) and formal thought disorder (Shenton et al, 1992). Bilateral reduction in volume of the hippocampal formation has been associated with the severity of disorganization syndrome (Fukuzako et al, 1997). Some studies have found frontal lobe reductions in schizophrenic patients (Harvey et al, 1993; Cannon et al, 1998). However, the results are inconsistent, especially with relation to the precise localization of the frontal abnormalities (Raine et al, 1992). Buchanan et al (1998) improved on this by subdividing the prefrontal cortex (PFC) into superior, middle, inferior, and orbital regions and found that patients with schizophrenia exhibited selective gray matter volume reductions in the inferior PFC bilaterally. There was no difference between the schizophrenic and control groups in any other region of the frontal lobes. There are some MRI data that provide support for a hypothesis of disconnection between brain areas in schizophrenia. Breier et al (1992) found that schizophrenic patients, compared with matched healthy controls, had reductions in the right and left amygdala, the left hippocampus, and prefrontal white matter. Moreover, the right prefrontal white matter volume in schizophrenic patients was significantly related to right amygdala/hippocampal volume, suggesting there might be abnormal connections between these areas. Buchsbaum et al (1997) found evidence for a decreased left hemispheric volume in frontal and temporal regions in schizophrenic patients. This result was supported by Woodruff et al (1997a), who found that inter-regional correlations were significantly reduced in schizophrenics between prefrontal and superior temporal gyrus volumes. The authors propose that these results support the existence of a relative "fronto-temporal dissociation" in schizophrenia. Reversal or reduction of normal structural cerebral asymmetries may be related to the pathogenesis of schizophrenia (Crow, 1995). Lack of normal asymmetry has been especially associated with early onset of schizophrenia. For example, DeLisi et al (1997) investigated MRI structural scans of 87 patients with a first episode of schizophrenia, 52 normal controls, and an independent group of 14 pairs of siblings with schizophrenia to evaluate evidence of heritability to cerebral asymmetries. Width asymmetries were reduced in patients compared with controls in the posterior and occipital regions. Brain horizontal length, on the other hand, was significantly more asymmetrical in patients (left >

right). In the 14 pairs of psychotic siblings, within-pair correlations for the horizontal sylvian fissure asymmetry were significantly greater than between-pair correlations. These findings are consistent with the early presence (possibly genetic) of anomalous cerebral asymmetry. Fukuzako et al (1997) found a significant asymmetry in the hippocampal formation volume in control subjects, but not in patients, and a significant positive correlation between the asymmetry index and the patient's age at the onset of schizophrenia. Similarly, Maher et al (1998) found that low levels of hemispheric asymmetry in the frontal and temporal areas were associated with early onset of schizophrenia, the association with frontal volume being more marked than with temporal volume. These findings are consistent with the hypothesis that failure to develop asymmetry is an important component of the pathology underlying some forms of schizophrenia. 1. Voxel-Based Morphometry Voxel-based morphometry is a new approach of looking at structural brain abnormalities using MRI. It is a data-led technique in which the brain images are normalized and then differences between groups anywhere in brain are identified (Wright et al, 1995). Andreasen et al (1994) created normal and schizophrenic average brains, compared the latter with the former, and found decreased thalamus size in schizophrenic patients. Wolkin et al (1998) used linear intersubject averaging of structural MR images to create a single averaged brain for the schizophrenic group {n = 25) and for the control group {n = 25). The signal intensity differences between these average images were consistent with cortical thinning/sulcal widening and ventricular enlargement.

C. Functional Imaging 1. PET Resting Studies There are many PET metabolism studies in the literature, and this review is not exhaustive. Many of the earlier studies investigated small sample sizes, and we will concentrate on studies that have used large sample sizes. The main finding from PET metabolism studies is that there is lower metabolism in the frontal lobes of schizophrenic patients compared to normal controls. This became known as hypofrontality. The first study using isotope imaging was by Ingvar and Franzen (1974), who compared 15 normal controls with 11 patients with dementia, and two groups of schizophrenics (one group comprised nine chronic, elderly patients; the other comprised 11 younger patients). Whereas the demented patients showed a lower level of overall metabolism, both

528 schizophrenic groups showed some evidence for reduced blood flow in anterior, relative to posterior, regions (Fig. 1). However, disagreement about the definition of hypofrontahty has caused problems and inconsistencies in the results. Studies vary on the frontal areas in which activity was measured. Some included all anterior regions, whereas others looked at frontal or prefrontal subdivisions only. The earlier studies tended to use the frontalioccipital ratio as a measure of hypofrontality, whereas recently most studies have used absolute frontal flow values with or without correction for mean total brain blood flow rates. The majority of studies, using any of these definitions, have found little evidence for statistically significant hypofrontality, and in several studies, hypofrontahty was due to an elevated flow in posterior regions (Mathew et al, 1988; Siegel et al, 1993). In other studies, the differences between control and schizophrenic frontal cortex metabolism are small. It has been claimed that this is not due to the drug status of the patients at scanning (Waddington, 1990). However, recently anti-psychotic medication has been found to affect brain metabolism (Miller et al, 1997). This latter study showed that subjects treated with antipsychotic medication had significantly higher rCBF in the left basal ganglia and left fusiform gyrus and lower rCBF in the anterior cingulate, left dorsolateral and inferior frontal cortices, and both cerebellar hemispheres, compared to when they were off medication. Several studies have looked at clinical correlates associated with hypofrontality, but the results are inconsistent. Among those showing a positive relationship with hypofrontality are chronicity (Mathew et al, 1988), negative symptoms (e.g., Ebmeier et al, 1993), and neuropsychological task impairment (e.g., Paulman et al,

IV Psychiatric Disorders 1990). However, the interpretation of these findings is ambiguous, since it is impossible to attribute cause or effect. Liddle et al (1992) correlated rCBF levels with symptom severity scores for his three clinical subdivisions in 30 schizophrenic patients. They demonstrated that the psychomotor poverty syndrome, which has been shown to involve a diminished ability to generate words, was associated with decreased perfusion of the dorsolateral prefrontal cortex (DLPFC) at a locus that is activated in normal subjects during the internal generation of words (see Fig. 1). The disorganization syndrome, which has been shown to involve impaired suppression of inappropriate responses (e.g., in the Stroop test), was associated with increased perfusion of the right anterior cingulate gyrus at a location activated in normal subjects performing the Stroop test. The reality distortion syndrome, which might arise from disordered internal monitoring of activity, was associated with increased perfusion in the medial temporal lobe at a locus activated in normal subjects during the internal monitoring of eye movements. Therefore the abnormalities of brain metabolism underlying each of the three syndromes might be widely distributed over the brain. Using data from the same patients, Friston et al (1992) examined correlations between rCBF and a measure of psychopathology receiving equal contributions from each of Liddle's three subsystems. The degree of psychopathology correlated highly with rCBF in the left medial temporal region, and mesencephalic, thalamic, and left striatal structures. The highest correlation was in the left parahippocampal region, and the authors proposed that this might be a central deficit in schizophrenia. A canonical analysis of the same data highlighted the left parahippocampal region and left striatum (globus pallidus), in which rCBF increased with increasing severity of psychopathology. Friston and colleagues suggested that disinhibition of left medial temporal lobe activity mediated by fronto-limbic connections might explain these findings.

IV. Neuroreceptor Imaging of Antipsychotics Using PET A. Dopamine Receptors

F i g u r e 1 Approximate average localization of hypofrontality found by Liddle et al. (1992), Weinberger et al. (1986), Daniel et al (1991), Volz et al. (1997), Ragland et al. (1998), Spence et al (1998), Yurgelun-Todd et al (1996), and Fletcher et al (1998).

Neuroreceptor imaging of antipsychotics using PET has been carried out for over a decade. Many studies have shown evidence that DA (in particular D2) receptors are increased in schizophrenic brains and that several different antipsychotics bind to these receptors (Wong et al, 1986; Breier et al, 1992). However, other studies using PET have produced contradictory results

21 Functional Neuroimaging Studies of Schizophrenia or only very weak evidence of an abnormality in DA receptor number (e.g., Crawley et al, 1986; Farde et al, 1987; Pilowsky et al, 1994). These discrepant results might, in part, be due to the diversity of PET methodology used by these groups. Meta-analytic methods were applied to post-mortem and PET and single photon emission computerized tomography (SPECT) neuroimaging data by Zakzanis and Hansen (1998).They found increases in D2 receptor density to be approximately 70% in schizophrenia. That is, about 30% of patients with schizophrenia could not be discriminated from normal controls. Based on the findings, it was argued that D2 receptor density increases in patients with schizophrenia, although a reliable finding in many patients, are not a specific or consistent marker of schizophrenia. Laruelle's (1998) meta-analysis of studies measuring D2 receptor density parameters revealed that, compared to healthy controls, patients with schizophrenia showed a significant but mild increase and a larger variability of D2 receptor density. D2 receptors are expressed most highly in the striatum, but many PET studies have failed to show any change in D2 receptor density in the striatum of schizophrenics, raising the possibihty that other receptors may also be involved. Okubo et al (1997) used PET to examine the distribution of Di and D2 receptors in the brains of drug-naive or drug-free schizophrenic patients. Although no differences were observed in the striatum relative to control subjects, binding of radioligand to Di receptors was reduced in the PFC of schizophrenics. V. Cognitive Activation Studies Functional neuroimaging is most frequently used to evaluate the regional cerebral responses (rCBF in PET or BOLD in fMRI) to a particular cognitive or sensorimotor process. Typically, subjects are scanned while performing an activation task, which engages the cognitive/sensorimotor process of interest, and a baseline task, which engages all components of the activation task except the cognitive/sensorimotor process of interest. Regions that show significantly more activity in the experimental state than in the baseline state are considered to be involved in the cognitive/sensorimotor process of interest. These task-specific activity patterns can then be compared between patients and control subjects to determine how the condition affects brain function. Most of the cognitive processes on which schizophrenic patients' performance is impaired involve the frontal lobes, since patients with frontal lesions per-

529

form poorly on these tasks and these tasks activate various regions of the frontal lobes in normal controls. There has been considerable evidence for hypofrontality in schizophrenia from cognitive activation studies. In addition, there is increasing evidence that schizophrenics show abnormal functioning of the temporal lobes, and, less consistently, the parietal lobes and hippocampus, during cognitive tasks. Early studies used less sensitive imaging techniques, for example, ^^^Xe inhalation or SPECT. In this review, we shall discuss some SPECT studies, but we concentrate on studies using more sensitive methods, specifically PET and functional magnetic resonance imaging (fMRI). There have been many negativefindings,due to a number of possible factors, which will be discussed. Finally, in studies finding abnormal activity in the frontal and temporal lobes in schizophrenia, there have been inconsistencies in terms of precise localization. However, the evidence is now starting to point toward anomalous frontotemporal integration during cognitive tasks. In this review, we shall discuss several types of task that have been used to evaluate brain function in schizophrenic patients. A. Task-Based Studies of Executive Function In many behaviorial studies the most striking impairments are seen when schizophrenic patients perform the various complex executive tasks associated with the frontal lobes. Brain imaging makes it possible to identify the abnormal pattern of brain activity associated with this abnormal performance. 1. Wisconsin Card Sorting Task (WCST) Schizophrenia is largely characterized by impairments in planning and execution and therefore tasks that involve this kind of planning and modification of behavior have been exploited in the scanner. Several researchers have investigated brain activity in schizophrenic patients while they perform a version of the Wisconsin Card Sorting Task (WCST). This task is known to activate the DLPFC in normal controls and is particularly sensitive to damage to DLPFC (Berman et al, 1995; Nagahama et al, 1996). In the typical computerized version of the WCST, subjects view a computer screen that displays a number of stimuli. These stimuli differ along three dimensions: color, shape, and number. On each of a series of trials, the subjects have to match a target stimulus with one of the four standard stimuU. However, the match is not exact, but has to made in terms of either color, shape, or form. Subjects are not informed of how to make the match, but are informed after each choice whether they are right or wrong. They have to determine from trial and error which dimension

IV Psychiatric Disorders

530 is correct. After subjects have made a series of correct responses, the rule is changed and subjects must determine a new rule for matching. Weinberger et al (1986) measured rCBF using ^^^^Xe inhalation SPECT in 20 medication-free patients with chronic schizophrenia and 25 normal controls during the WCST and a number-matching control task. During the WCST, but not during the control task, normal subjects showed increased DLPFC rCBF, whereas patients did not (patient performance was worse than that of the controls). Furthermore, in patients, DLPFC rCBF correlated positively with WCST performance. The authors suggest that this result shows that the better DLPFC was able to function, the better patients could perform the task. However, this conclusion is based on an illfounded assumption. It is impossible to determine whether task-related underactivation causes or reflects poor task performance. This is a crucial issue in cognitive activation studies and will be discussed in more detail throughout the chapter. Daniel et al (1991) used SPECT to study the effect of amphetamine (a DA agonist) and a placebo on rCBF in 10 chronic schizophrenic patients while they performed a version of the WCST. They compared this to rCBF during a sensorimotor control task matched to the WCST in terms of visual stimulation and motor responses, but without the abstract reasoning and working memory components of the WCST. On placebo no significant activation was seen during the WCST compared with the control task. In contrast, significant activation of the left DLPFC occurred on the amphetamine trials. Daniel and colleagues point out that patients' performance improved with amphetamine relative to placebo and that with amphetamine, but not with placebo, a significant correlation was found between activation of DLPFC and performance on the WCST task. Again, this is a finding that is impossible to interpret: did amphetamine facilitate task performance, which then caused an increased rCBF in the PFC? Or did amphetamine cause PFC activity to increase, which facilitated task performance? Volz et al (1997) used fMRI to investigate activity during the WCST in 13 chronic schizophrenics on stable neuroleptic medication. They also showed evidence for lack of activation in the right PFC and a trend toward increased left temporal activity during the WCST compared to normal controls. However, again the task performance was different between the two groups and therefore the results remain ambiguous. In addition, this study was limited because a one-slice imaging technique was used so no information about the activation pattern in adjacent brain regions was obtained. Ragland et al (1998) found that better WCST performance correlated with rCBF increase in prefrontal

regions for controls and in the parahippocampal gyrus for patients. The results suggest that schizophrenia may involve a breakdown in the integration of a frontotemporal network that is responsive to executive and declarative memory demands in healthy individuals. 2. Tower of London Task The Tower of London task involves high-level strategic planning among a number of other processes (Shallice, 1982). In this task, subjects have to rearrange a set of three balls presented on a computer screen so that their positions match a goal arrangement also presented on the screen (see Fig. 2). The complexity of the task, in terms of the number of moves necessary to complete the task, can be varied. Andreasen et al (1992) used SPECT during the Tower of London task in three different groups: 13 neuroleptic-naive schizophrenic patients; 23 non-naive schizophrenic patients who had been chronically ill but were medication free for at least 3 weeks; and 15 healthy normal volunteers. The Tower of London task activated the left mesial frontal cortex (probably including parts of the cingulate gyrus) in normal controls, but not in either patient group. Both patient groups also lacked activation of the right parietal cortex, representing the circuitry specifically activated by the Tower of London task in normal controls. Importantly, decreased activation occurred only in the patients with high scores

F i g u r e 2 The Tower of London task involves high-level strategic planning among a number of other processes (Shallice, 1982). In this task, subjects have to rearrange a set of three balls presented on a computer screen so that their positions match a goal arrangement also presented on the screen. The complexity of the task, in terms of the number of moves necessary to complete the task, can be varied.

21 Functional Neuroimaging Studies of Schizophrenia for negative symptoms. The authors therefore suggested that hypofrontahty is related to negative symptoms and is not a long-term effect of neuroleptic treatment or of chronicity of illness. This is an important study because of the large number of neuroleptic-naive patients investigated, but it is limited in several ways, in particular the low-resolution imaging technique that was used. There was also the methodological problem that schizophrenic patients performed poorly on the tasks involved, so whether less activation of the PFC was due to poorer performance or vice versa cannot be resolved.

B. The Component Processes Underlying Executive Function One problem with studies that used complex executive tasks is that these tasks involve many processes. For example, the WCST involves choosing a strategy, remembering the previous responses in order to learn by trial and error, attending to one dimension rather than another, and so on. In the absence of a series of carefully constructed comparison tasks, it is not possible to relate the various brain regions activated with each of the component processes. In the following section, we review studies in which simpler tasks were used with far fewer component processes. 1. Motor Tasks Even tasks that require no more than the production of a simple sequence of movements can be associated with abnormal patterns of brain activity in schizophrenia. Mattay et al (1997) studied seven patients with schizophrenia and seven normal subjects while they performed a finger movement task of increasing complexity. Patients showed greater ipsilateral activation in the primary sensorimotor and lateral premotor regions and had a significantly lower laterality quotient than normal subjects. These functional abnormahties increased with the complexity of the task. The authors proposed that these results demonstrate a functional disturbance in the cortical motor circuitry of schizophrenic patients. Schroder et al (1999) asked 12 patients and 12 healthy controls to produce sequences of pronation and supination movements at three different speeds during fMRI scanning. Both groups showed increasing activity with increasing speed in sensorimotor cortex and supplementary motor area (SMA). However, the patients showed less overall activation than the controls. While the patients did not differ in the amplitude and frequency of their movements, they did show significantly greater variabihty in their movements than the controls. The differences were most marked in a subgroup of pa-

531

tients who were drug free at the time of testing. Both these studies raise the possibility of a fundamental, but subtle problem of motor control associated with schizophrenia. 2. Willed Action Willed action involves a somewhat "higher" stage in the control of action. There is a fundamental distinction between actions elicited by external stimuU and actions elicited by internal goals (acts of will). Routine actions are specified by a stimulus. In contrast, in willed (or selfgenerated) acts, the response is open-ended and involves making a deliberate choice. Willed actions are a fundamental component of executive tasks. In normal subjects, willed acts in two response modaUties (speaking a word or lifting a finger), relative to routine actions, were associated with increased blood flow in the DLPFC (Brodmann area 46; Frith et al, 1991). Schizophrenic patients typically show abnormalities of willed behavior. In chronic patients, intentions of will are no longer properly formed and so actions are rarely elicited via this route. This gives rise to behavioral negative signs (e.g., poverty of speech and action). Spence et al (1998) performed a PET study in which subjects had to make voluntary joystick movements in the experimental condition, and stereotyped (routine) movements in the basehne condition, and do nothing in a control rest condition. They analyzed data from 13 schizophrenic patients, comparing two occasions when symptoms were severe and when they had subsided, and included data from a normal control group to clarify the role of the left DLPFC in volition. The DLPFC was activated by normal controls for the free-choice task only. However, it was not activated by schizophrenics with symptoms but became activated when their symptoms decreased. The authors noted that the DLPFC was also activated during the stereotyped joystick movement task in schizophrenic patients in remission, in contrast to a control group. Spence and colleagues concluded that, since hypofrontahty was evident in schizophrenics who can perform the experimental task, the DLPFC is not necessary for that task. In addition, hypofrontahty seems to depend on current symptoms. They suggested the reason for previous equivocal hypofrontahty results is that schizophrenics with a varying amount and combination of symptoms were being compared with normal controls. This study raises the puzzle of the precise role of the DLPFC in freely chosen movements. As the authors point out, DLPFC was activated by both the freely chosen and the stereotyped movement tasks in schizophrenics without symptoms. If patients do behave in a more random manner than controls, it follows that the

532 patients might find making stereotyped movements harder than controls. In other words, it is possible that stereotyped movements are not truly stereotyped for schizophrenic patients. On the other hand, the production of random movements may be easier. TTius, although this study represents an advance on those in which patients show impaired task performance, as the authors acknowledge, caution is necessary when interpreting the results. What is the functional significance of DLPFC activation in normal controls and patients without symptoms if schizophrenic patients with symptoms can perform the task without recruiting the DLPFC? 3. Verbal Fluency Verbal fluency tasks involve subjects having to generate words to a given cue. For example, subjects might have to produce a word beginning with a certain letter, a different letter being presented every 5 s. This can be seen as a task that involves willed action since subjects have to choose for themselves precisely which word to say. Verbal fluency tasks engage a distributed brain system similar to that engaged by motor response selection tasks associated with willed action (Frith et al., 1991). Yurgelun-Todd et al. (1996b) investigated verbal fluency using fMRI in 12 schizophrenic patients and 11 normal control subjects. They showed evidence for reduced left PFC activation and increased left temporal activation relative to control subjects during the verbal fluency task. However, the lack of frontal activation by cognitive tasks in schizophrenic patients has not consistently been located in the PFC. Dolan et al. (1995) and Fletcher et al. (1996) used a factorial design to test the effect of apomorphine, a nonselective DA agonist, which when given in very low doses as in this experiment acts primarily on autoreceptors, thus decreasing the release of endogenous dopamine. Brain systems engaged by a paced verbal fluency task in unmedicated schizophrenic patients and normal controls were studied using PET. Activation of the DLPFC was normal, but they found a failure of task-related activation in anterior cingulate cortex and deactivation of the left superior temporal gyrus in the schizophrenic subjects (see Fig. 3). Fletcher and colleagues therefore suggested that schizophrenia is associated with both segregated (anterior cingulate) and integrative (fronto-temporal) functional abnormalities. Cingulate activation was restored by low-dose apomorphine in schizophrenics. Additionally, the abnormal fronto-temporal pattern of activation in schizophrenic subjects was normalized by this neuropharmalogical intervention. OveraU, in schizophrenic subjects

IV Psychiatric Disorders the effect of apomorphine was to modify the pattern of brain activity, making it more similar to that seen in control subjects. The interpretation of the apomorphine-induced reversal of the deactivation in the left temporal lobe in schizophrenic subjects is unclear. It might reflect a direct influence of apomorphine on the temporal lobe; alternatively, the reversal could be due to a ''downstream" effect of the change in anterior cingulate function. The authors interpret the absence of the normal reciprocal interaction between the frontal and the superior temporal cortex in schizophrenia (the failure of task-related deactivation of the superior temporal gyrus in the schizophrenic group) to suggest the presence of impaired functional integration. This is an important concept, especially given the relatively large amount of evidence showing a lack of temporal deactivation in the presence of a lack of frontal activation in schizophrenia. The finding of normal prefrontal activation found in this study is at variance with some previous functional neuroimaging studies investigating schizophrenia, as outlined above. It is in agreement, however, with a study in which task performance was optimized by pacing the task (Frith et al., 1995). Using PET, these researchers investigated rCBF of 18 chronic schizophrenic patients and 6 normal controls matched for age, sex, and premorbid IQ while they performed (a) paced verbal fluency, (b) paced word categorization, and (c) paced word repetition. The schizophrenic patients were split into three groups according to their verbal fluency task performance level. All patient groups showed the same pattern of left PFC activation as control subjects, independent of their level of performance. However, in the left superior temporal cortex, all patient groups failed to show a normal deactivation when verbal fluency was compared with word repetition. Again, this result was interpreted to reflect abnormal functional connectivity between frontal and temporal cortex. Friston and Frith (1995) performed an additional analysis of their data from the same three groups of schizophrenic patients according to the level of task performance: poverty (no words), odd (wrong words), and unimpaired. They used analytic techniques specifically to assess cortico-cortical interactions. Normal controls showed negative frontotemporal interactions whereas all the schizophrenic patients showed positive interactions, mostly between the left PFC and inferotemporal cortex. Friston and Frith suggested that this might represent a failure of the PFC to inhibit the temporal lobes. They postulated that the temporal lobes may be required to recognize the consequences of actions initiated by the frontal lobes in order to integrate action and perception.

533

21 Functional Neuroimaging Studies of Schizophrenia sagittal

-104 C-

Oi

coronal

transverse • ^

h'N

V,. 64

F i g u r e 3 Reproduced with permission from Fletcher et al. (1996). (a) Statistical parametric maps (SPMs) showing brain regions where there was a significant {p < 0.005) difference in drug (apomorphine)-task (verbal fluency) interaction between the schizophrenic group and the control group. The area in which there was an augmenting effect of the drug on the task-related activity occurring in schizophrenics compared to controls was the anterior cingulate gyrus. In other words, the impaired cingulate activation seen during the verbal fluency task in schizophrenic patients was significantly reversed by apomorphine. (b) Transverse section showing the verbal fluency task-related deactivation of the superior temporal gyrus that was absent in schizophrenic relative to control subjects. The authors interpret the failure of taskrelated deactivation of the superior temporal gyrus in the schizophrenic group to suggest the presence of impaired functional temporo-frontal integration.

IV Psychiatric Disorders

534 C. Memory Tasks Memory impairments are especially enduring symptoms in schizophrenia (Green, 1996), with memory storage particularly affected (Feinstein et al, 1998). Working memory (WM) depends upon executive processes instantiated in frontal cortex and is involved in complex tasks such as the WCST and the Tower of London task. The neuroanatomy of memory in normal subjects has been mapped to a network of cortical and subcortical structures in the human brain (Ungerleider, 1995). This provides a model of memory with which to compare the functioning of memory systems in schizophrenia in order to investigate disturbances of specific memory components. In particular, the DLPFC and the hippocampal formation have been the subject of investigation in schizophrenia, as these are involved in various aspects of memory (Arnold, 1997; Goldman-Rakic and Selemon, 1997). The DLPFC is activated by semantic processing during encoding and retrieval, whereas hippocampal activation is associated with the detection of novelty and the creation of associations during encoding (Schacter et al, 1996; Dolan and Fletcher, 1997). As for other executive and cognitive functions, impairment of WM in schizophrenia has been associated with decreased blood flow in the DLPFC (Goldman-Rakic and Selemon, 1997). Recent studies have used memory tasks on which the patients' memory is at normal levels and have shown evidence for hypofrontality. Several functional neuroimaging studies have failed to find evidence for abnormal activation of temporal or frontal cortex in schizophrenia during memory tasks (Busatto et al, 1994, using a verbal memory task with SPECT; Ragland et al, 1998, using a paired associate recognition test with PET). Other studies have shown rCBF changes that overlapped in the schizophrenic and control groups, with a trend toward patients showing smaller activations than controls in frontal and superior temporal cortical regions (e.g., Ganguh et al, 1997, using a verbal free-recall supraspan memory task). These differences may be due to the different type of memory tasks used by each group. Other groups have found evidence for hypofrontality during memory tasks, but these have often been confounded by performance levels. For example. Carter et al (1998) used PET to evaluate rCBF associated with the "N-back" WM task, which activates the PFC as a function of WM load in normal subjects. Under low-working-memory-load conditions, the accuracy of both groups in the N-back task was equal, but when the memory load increased, the patients' performance deteriorated more than did that of control subjects. The rCBF response to increased WM load was significantly reduced in the patients' right DLPFC. Callicott et al (1998) investigated BOLD sig-

nal changes in 10 patients with schizophrenia and 10 controls performing a novel N-back WM task, using fMRL After confounds were removed and subjects were matched for signal variance (voxel stabihty), decreased DLPFC activity and a tendency for overactivation of parietal cortex were seen. However, these findings are difficult to interpret in the context of abnormal task performance in patients. There may be nothing inherently abnormal about the physiology of the frontal cortex in schizophrenia, but patients may be failing to select frontally mediated cognitive strategies because of abnormal connectivity between otherwise normal regions. Wiser et al (1998) used PET to measure rCBF during a long-term recognition memory task for words in schizophrenic patients and in healthy subjects. The task was designed so that performance scores were similar in the patient and control subjects. This memory-retrieval task did not activate PFC, precuneus, and cerebeUum in patients as much as it did in the control group. This finding suggests that there is a dysfunctional corticocerebellar circuit in schizophrenia. Hypofrontality has not always been found in studies using modified tasks to optimize the performance of schizophrenic subjects. Hypofrontality was not found in a study by Heckers and colleagues (1998) in which they used PET to evaluate 13 schizophrenic patients and a group of normal control subjects during memoryretrieval tasks. Prior to each memory-retrieval scan, subjects learned a list of written words. In half the trials, a ''shallow" learning task was used, in which subjects were asked to count the number of right angles (T-junctions) in the letters of each word. This is known to result in poor subsequent memory for those words ('iow recall"). In the other half of the trials, subjects performed a deep-encoding task in which words were learned by counting the number of meanings for each word, which is known to produce good subsequent memory for those words ("high recall"). Subjects were then scanned during high and low levels of memory recall. During the scans subjects were asked to retrieve studied words on the basis of three-letter (stem) cues, allowing a comparison between the effort of retrieval and the process of successful retrieval. In this study, activation of the PFC correlated with effort of retrieval and hippocampal and parahippocampal activation occurred during successful retrieval in normal control subjects, a finding consistent with previous studies of memory in normals (Schacter et al, 1996; Dolan and Fletcher, 1997). The schizophrenic patients recruited the PFC during the effort of retrieval but did not recruit the hippocampus during conscious recollection. This pattern of activation was associated with higher accuracy during low recall and lower accuracy

535

21 Functional Neuroimaging Studies of Schizophrenia during high recall in schizophrenics than in control subjects. In addition to the task-specific hippocampal underactivation, the authors observed a generally higher overall level of nonspecific hippocampal activity, supporting previous metabolic studies (e.g., Liddle et ai, 1992). The authors suggested that high baseline hippocampal activity together with an absence of task-specific activation demonstrates abnormal cortico-hippocampal functional integration in schizophrenia. The schizophrenic patients showed a more widespread activation of prefrontal areas and parietal cortex during recollection than controls, and the authors propose that this overactivation represents an "effort to compensate for the failed recruitment of the hippocampus." This interpretation again moves away from the simple notion of dysfunction in isolated brain regions explaining the cognitive deficits in schizophrenia and toward the idea that neural abnormahty in schizophrenia reflects a disruption of integration between brain areas. Other groups have found evidence for abnormal activation of different frontal regions. For example, Stevens et al (1998) used fMRI to study the neural basis of the dissociation of auditory verbal and nonverbal WM in schizophrenia using the Word Serial Position Task and Tone Serial Position Task. Activation in the left inferior frontal gyrus (Brodmann areas 6, 44, and 45) was much reduced during the Word Serial Position Task in the patients and failed to show the same taskspecific activation as in controls. Reduced activation in patients also extended to a medial area during the Tone Serial Position Task and to premotor and anterior temporal lobes during both tasks. Fletcher et al (1998) used PET to compare rCBF in memory-impaired and nonimpaired schizophrenic patients with normal controls during a parametrically graded memory task. They found that DLPFC activity correlates with memory task difficulty and performance in the control group. In contrast, for both schizophrenic groups, DLPFC activity levels plateaued as task difficulty increased despite a significant difference in performance between the two schizophrenic groups. The authors therefore suggested that hypofrontality in schizophrenics correlates with task difficulty, rather than task performance, since the memory-impaired schizophrenic group performed worse than the nonimpaired group even though both groups showed no increase in PFC activity as task difficulty increased. Unlike the control group, there was no inferior temporal/parietal deactivation in either schizophrenic group. The authors suggest that the lack of deactivation of these areas might represent a temporofrontal disconnection in schizophrenics. Indeed they suggested that because temporal/parietal activations were not corre-

lated with performance, they therefore might represent a core pathology of schizophrenia. This study improves on previous cognitive activation studies in which the confound of non-matched task performance occurs. However, the function of the PFC in memory is difficult to interpret since the unimpaired group could still perform the task when the level of PFC rCBF had plateaued as task difficulty increased. Further evidence for abnormal integration between brain areas in schizophrenia comes from a study that specifically investigated the functional integration between brain areas (Fletcher et al, 1999). Functional integration considers complex cognitive processes as emergent properties of interconnected brain area, building on the idea that simple cognitive processes can be localized in discrete anatomical modules (referred to as ''functional segregation"). Brain areas A and B may be functionally connected if it can be shown that an increase (or decrease) of activity in area A is associated with an increase (or decrease) in area B, which can be shown empirically by analysis of covariance. In this case, activity in A might cause activity in B, or activations in A and B might be caused by changes in another area (C) that projects to A and B. Alternatively, areas A and B may be effectively connected if their relationship can be shown to be causal. This requires a more complex approach in which the anatomical components of a cognitive system are defined. Connections between these regions are designated on the basis of empirical neuroanatomy and the connections are allocated weights or path strengths by an iterative leastsquares approach in such a way that the resultant functional model of inter-regional influences best accounts for the observed variance-covariance structure generated by the functional neuroimaging observations (Frision etal, 1993). A simplified version of effective connectivity (Friston et al, 1997) was employed by Fletcher et al (1999) to evaluate effective connectivity between regions in the data from their PET study of a graded memory task in schizophrenia. They demonstrated that in control subjects, but not in the schizophrenic patients, the product of PFC and ACG activity predicted a bilateral temporal and medial PFC deactivation. They interpreted these results as showing that in schizophrenia there is an abnormality in the way in which left PFC influences left superior temporal cortex, and this abnormality is due to a failure of the ACG to modulate the prefrontotemporal relationship (see Fig. 4). D. Conclusion There is a body of evidence suggesting that schizophrenic patients show abnormal interactions and influ-

536

IV Psychiatric Disorders

transverse

0

b

-T

PFC*ACC

r-

•

-JI

5 I

1

!

1

1

i

1-^ 1

X

0 -2 -4 -10

' l„

-5

1.1

0

1

L

10

1

15

20

F i g u r e 4 Reproduced with permission from Fletcher et al (1999). Left panel: Region of left superior temporal cortex showing a significant difference between control and schizophrenic subjects (p < 0.05). A SPM rendered onto sagittal and transverse section of a stereotactically normalized structural MRI is shown. Right panel: Graphic representation of the relationship between activity in left superior temporal cortex (STG; y axis) and the combination of cingulo-prefrontal activity (PFC x ACG; x axis) in (a) controls and (b) schizophrenic subjects. The linear best fit is shown for both groups. The regression coefficient of each of the two groups is considered to provide a measure of Contribution of PFC x ACC to superior temporal cortical activity. Thus, in the controls an increase in PFC x ACC produces an inhibition of superior temporal activity. In the schizophrenic subjects, the line slopes upward, indicating the opposite effect.

ences among brain regions (or functional integration) during cognitive tasks. There is currently little direct evidence in favor of this hypothesis, and several regions have been found to function abnormally, with no unequivocal evidence for any particular region being involved. However, the majority of positive findings suggest that a disruption of fronto-temporal integration is a core feature of schizophrenia. However,findingshave

been confounded by several factors, especially use of poor control tasks such as rest and nonmatched task performance in the schizophrenic and control groups. Future cognitive activation studies using improved methodologies should resolve issues such as whether abnormal frontal function causes or reflects poor task performance in schizophrenia. A question of clinical importance is whether different patterns of cortical inter-

21 Functional Neuroimaging Studies of Schizophrenia action correlate with or predict schizophrenic symptoms or outcome. VI. Imaging Symptoms Functional neuroimaging is also useful for evaluating neural activity in patients experiencing psychotic symptoms. A. Hallucinations Hallucinations, perceptions in the absence of external stimuH, are prominent among the core positive symptoms of schizophrenia. Although auditory hallucinations are more common than any other type, hallucinations in other sensory modalities occur in a proportion of patients. Hallucinations in all modalities tend to be associated with activity in the neural substrate associated with that particular sensory modality. However, there is evidence that activation of the sensory cortex particular to the false perception is not in itself sufficient for the perception. Instead, research suggests that the interaction of a distributed cortico-subcortical neural network might provide a biological basis for schizophrenic hallucinations. 1. Auditory Hallucinations The most common type of hallucination in schizophrenia is in the auditory domain and normally consists of spoken speech or voices (Hoffman, 1986). Functional neuroimaging studies of auditory hallucinations suggest that they involve neural systems dedicated to auditory speech perception as well as a distributed network of other cortical and subcortical areas. There are two distinct approaches to the study of the physiological basis of auditory hallucinations. The first, which we will call the state approach, asks what changes in brain activity can be observed at the time hallucinations are occurring. The second, which we call the trait approach, asks whether there is permanent abnormality of brain function present in patients who are prone to experience auditory hallucinations when they are ill. This abnormality will be observable even in the absence of current symptoms. a. State Studies Silbersweig et al (1995) used PET to study brain activity associated with the occurrence of hallucinations in six schizophrenic patients. Five patients with classic auditory verbal hallucinations demonstrated activation in subcortical (thalamic and striatal) nuclei, limbic structures (especially hippocampus), and paralimbic regions (parahippocampal and

537

cingulate gyri and orbitofrontal cortex). Temporoparietal auditory-hnguistic association cortex activation was present in each subject. One drug-naive patient had visual as well as auditory verbal hallucinations and showed activations in visual and auditory-linguistic association cortices. The authors propose that activity in deep-brain structures seen in all subjects may generate or modulate hallucinations, and the particular sensory cortical regions activated in individual patients may affect their specific perceptual content. Importantly, this study pointed to the possibility that hallucinations coincide with activation of the sensory and association cortex specific to the modahty of the experience, a notion that has received support from several further studies. David et al. (1996) used fMRI to scan a schizophrenic patient while he was experiencing auditory hallucinations and again when hallucination-free. The subject was scanned during presentation of exogenous auditory and visual stimuli and while he was on and off antipsychotic drugs. The BOLD signal in the temporal cortex to exogenous auditory stimulation (speech) was significantly reduced when the patient was experiencing hallucinating voices, regardless of medication. Visual cortical activation to flashing lights remained the same over all four scans, whether the subject was experiencing auditory hallucinations or not. A similar result was obtained by Woodruff et al (1997b), who used fMRI to study seven schizophrenic patients while they were experiencing severe auditory verbal hallucinations and again after their hallucinations had subsided. On the former occasion, these patients had reduced responses in temporal cortex, especially the right middle temporal gyrus, to external speech, compared to when their hallucinations were mild. The authors thus proposed that auditory hallucinations are associated with reduced responsivity in temporal cortical regions that overlap with those that normally process external speech, possibly due to competition for common neurophysiological resources. Recently, Dierks et al (1999) used event-related fMRI to investigate three paranoid schizophrenics who were able to indicate the on-set and off-set of their hallucinations as in the study by Silbersweig et al. Using this design, they found that primary auditory cortex, including Heschl's gyrus, was associated with the presence of auditory hallucinations. Secondary auditory cortex, temporal lobe, and frontal operculum (Broca's area) were also activated during auditory hallucinations, supporting the notion that auditory hallucinations are related to inner speech. Finally, hallucinations were also associated with increased activity in the hippocampus

538

IV Psychiatric Disorders

and amygdala. The authors suggested that these activations could be due to retrieval from memory of the hallucinated material and emotional reaction to the voices, respectively. Studies are divided on the issue of laterality of activity in the temporal lobe associated with schizophrenic hallucinations. Most studies report activity in the left temporal lobe (e.g., Cleghorn et ai, 1990; Suzuki et al, 1993; McGuire et al, 1996), consistent with the lateralization of language areas. Very few studies report activity in the right temporal lobe, and most that do have used a very small number of subjects, in some cases a single subject (e.g.. Woodruff era/., 1997b). b. Trait Studies The finding that auditory hallucinations are associated with activation of auditory and language association areas is consistent with the proposal that auditory verbal hallucinations arise from a disorder in the experience of inner speech (Frith, 1992). This was investigated by McGuire et al (1996). They used PET to evaluate the neural correlates of tasks that engaged inner speech and auditory verbal imagery in schizophrenic patients with a strong predisposition to auditory verbal hallucinations (hallucinators), schizophrenic patients with no history of hallucinations (nonhallucinators), and

normal controls. There were no differences between hallucinators and controls in regional cerebral blood flow during thinking in sentences. However, when imagining sentences spoken in another person's voice, which entails both the generation and monitoring of inner speech, hallucinators showed reduced activation of the left middle temporal gyrus and the rostral supplementary motor area, regions activated by both normal subjects and nonhallucinators. Conversely, when nonhallucinators imagined speech, they differed from both hallucinators and controls in showing reduced activation in the right parietal operculum (see Fig. 5). McGuire and his colleagues suggest that the presence of verbal hallucinations is associated with a failure to activate areas concerned with the monitoring of inner speech. 2. Visual Hallucinations The neural correlates of visual hallucinations also seem to be located in the neural substrate of visual perception. At least part of the activity in the brain associated with the experience of visual hallucinations is located in the visual cortex. For example, using SPECT, Hoksbergen et al (1996) found that visual hallucinations were associated with hypoperfusion in the right occipito-temporal region, which showed partial normal-

F i g u r e 5 PET data in a group of patients with schizophrenia. Subjects were shown ambiguous pictures that provoked incoherent (thought-disordered) verbal responses when they were asked to describe what the pictures showed. Each subject was shown 12 different stimuH, which ehcited a range of responses that varied with respect to how thought disordered they were. The severity of thought disorder was correlated with rCBF across the 12 scans, controlling for differences in the total number of words articulated. Positive correlations (areas that were more active the more disordered the speech) were evident in the left parahippocampal/anterior fusiform and right anterior fusiform gyri, whereas negative correlations were seen in the inferior frontal gyri, the cingulate gyrus, and the left superior temporal gyrus. Modified from McGuire et al. (1998).

21 Functional Neuroimaging Studies of Schizophrenia ization after the visual hallucinations had subsided. Howard et al (1997) used fMRI to investigate the visual cortical response to photic stimulation during and in the absence of continuous visual hallucinations. When visual hallucinations were absent, photic stimulation produced a normal bilateral activation in striate cortex. During hallucinations, very limited activation in striate cortex could be induced by exogenous visual stimulation. Similarly, Ffytche et al (1998) found activity in ventral extrastriate visual cortex in patients with the Charles Bonnet syndrome when they experienced visual hallucinations. Moreover, the content of the hallucinations reflected the functional specialization of the region; for example, hallucinations of color activated V4 (the human color center) (Zeki et al, 1991). In conclusion, functional neuroimaging studies suggest that hallucinations involve an interaction between the neural systems dedicated to the particular sensory modality in which the false perception occurs and a widely distributed cortico-subcortical system, including limbic, paralimbic, and frontal areas. Intersubject variability in the specific location of the sensory activation associated with the hallucination could arise from differences between the patients in the sensory content and experience of their hallucinations.

B. Thought Disorder McGuire et al (1998) scanned six schizophrenic subjects with PET while they described a series of ambiguous pictures, which provoked different degrees of thought-disordered speech in each patient. The severity of thought disorder was correlated with rCBF across the 12 scans, controlling for differences in the total number of words articulated. Verbal disorganization (positive thought disorder) was inversely correlated with activity in the inferior frontal, cingulate, and left superior temporal cortex, areas implicated in the regulation and monitoring of speech production (see Fig. 5). The authors propose that this reduced activity might contribute to the articulation of the linguistic anomalies that characterize positive thought disorder. Verbal disorganization was positively correlated with activity in the parahippocampal/anterior fusiform region bilaterally, which may reflect this region's role in the processing of linguistic anomalies.

C. Passivity Spence et al (1997) performed a PET study in which subjects had to make voluntary joystick movements in the experimental condition, and stereotyped (routine) movements in the baseline condition, and do nothing in

539

the rest condition. They investigated a group of schizophrenic patients with passivity (delusions of control) and without (the same group in remission) and a group of normal controls. Schizophrenic patients with passivity showed hyperactivation of the inferior parietal lobe (Brodmann area 40), the cerebellum, and the cingulate cortex relative to schizophrenic patients without passivity (see Fig. 6). Similar results were found when schizophrenic patients with passivity were compared to normal controls. A comparison of all schizophrenics with normal controls revealed hypofrontahty in the patients. When patients were in remission and no longer experienced passivity symptoms, a reversal of the hyperactivation of the parietal lobe and cingulate was seen. Hyperactivity in the parietal cortex may reflect the "unexpected" nature of the experienced movement in patients, as though it were being caused by an external force.

VII. Obstacles to Functional Neuroimaging and Schizophrenia Studies A. Subject Matching It is important that patients are matched to the control group on as many factors as possible. However, the best way to match IQ and education level is uncertain. It is not clear whether the control group's education level should be matched to the patient's parental or premorbid education level. However, both of these are su-

F i g u r e 6 Diagram of overactivity in right parietal lobe (Brodmann area 40) in patients with passivity while making voluntary joystick movements (in the study by Spence et al.. 1997). When patients were in remission and no longer experienced passivity symptoms, a reversal of this hyperactivation of parietal lobe was seen. Hyperactivity in parietal cortex might reflect the patient's experiencing the movement as "unexpected," as if it were being caused by some external force.

540

IV Psychiatric Disorders

perior to matching controls to patients' current IQ level, which may be considerably impaired by illness. Another question is whether the control group should be normal or psychiatric. Normal controls are important in order to establish a baseline model of the neural circuitry involved in an experimental task. However, there are clearly several problems with using nonpsychiatric controls. These include the fact that normal controls are not taking medication, hospitalized, or affect-flattened, factors that might have a direct effect on rCBF or that might cause the patients to be more or less motivated or to attend or think more or less. Since psychiatric patients will be more matched on these factors, they may constitute a preferable control group. However, there are also problems with using psychiatric patients as controls. There is the question of which psychiatric population should be used. Should they be taking the same medication, or is hospitalization the most important factor? They might not be able to perform the experimental task for some reason that is different from that causing impairment in schizophrenic patients. Therefore, comparing schizophrenic to depressed patients, for example, may reveal activity specific to depression or to schizophrenia, rather than the task in question. Some studies have used non-affected siblings, or even MZ twins, as controls for the schizophrenic patients. However, these are not matched for medication or hospitalization, even though they are usually matched for background. Furthermore, many studies investigate the non-affected MZ twins of schizophrenic patients as a high-risk group, since there is evidence and theory that this group may be affected in some way by predisposing factors (Crow, 1995). It is questionable whether male and female schizophrenic patients should be evaluated as a single group since gender differences have been documented in schizophrenia (e.g., Andreasen et al, 1994). B. Experimental Tasks As has been discussed at length throughout this chapter, there is a conceptual problem with applying the approach of cognitive activation studies to patient groups. It is impossible to interpret patterns of brain activity that differ between control and patient groups when the performance of the two groups on tasks differs in terms of degrees of efficiency and success. Many functional neuroimaging studies have used unmodified frontal tasks, in which the performance of schizophrenics falls below that of controls. It is unclear how to interpret increases or decreases of regionally specific activation in these studies. Any difference in brain activity between the two tasks could represent a critical

abnormality in schizophrenia and might cause poor task performance, or alternatively it might reflect poor performance. It is difficult to distinguish these two alternatives. Therefore, recent studies have employed tasks on which performance of the patient and control group is matched. However, there is also a problem with interpreting the results of activation studies in which the task performance of the patient and control group is matched. What do differences in brain activity mean in the context of normal task performance? If an area is activated more in the controls than in the patient group during such a task, the functional significance of that activation is difficult to understand—it is clearly not necessary for performing the task. The most obvious interpretation is that patients and controls are using different strategies to achieve similar task performance. Therefore, interpretational difficulties remain: What is the nature of the relationship between differences in brain activity and behavior? How do these two variables relate to the schizophrenic state? These problems have not been resolved and remain when interpreting data from studies in which task performance of patient and control groups is matched. Comparing an experimental condition with a very low-level baseUne, or rest, is problematic because schizophrenics might show different brain activation during baseline than normal controls. In other words, schizophrenics' brains might not activate in the "normal" way during rest. Indeed there is much evidence from PET metabolism studies to indicate that blood flow in the brains of schizophrenic patients at rest is different from that in normal controls. Different rCBF at rest could be interpreted as activation (or deactivation) during the experimental task if rest is used as the control task. C. Symptom-Specific Groups Schizophrenia is a heterogeneous illness, comprising a variety of different symptoms. Using groups of patients defined by diagnosis (schizophrenia) may explain the inconsistent and equivocal results of functional imaging studies, since each symptom can be associated with a different brain pattern or functional abnormality. Attempts have been made to correlate cognitive brain activity with specific symptoms or clinical signs. However, as reviewed in this chapter, the results are inconsistent. Correlation studies should be hypothesisdriven, since correlations will always occur by chance if enough variables are evaluated. It would be an improvement if a hypothesis were made based on past data, for example that temporal lobe abnormality might contribute to auditory hallucinations because temporal lobe epileptics experience such symptoms. If, on the

21 Functional Neuroimaging Studies of Schizophrenia other hand, a trawhng exercise is performed to identify symptom-brain activation pattern correlations, proper statistical thresholds with corrections for multiple comparisons need to be made to obtain any informative results. A clear advantage of using symptom-specific schizophrenic groups is that the control group can comprise people with a diagnosis of schizophrenia, who are thus matched in terms of medication and hospitahzation, but who do not have a particular symptom. Better still, the same group of schizophrenic patients can be used as their own control group if and when the symptom evaluated remits (see the study by Spence et al, 1998). However, a clear shortcoming of using symptom-specific groups is that little can be discovered about schizophrenia as a syndrome.

VIII. Conclusion Although there has been no one specific, unequivocal finding pecuhar either to the syndrome of schizophrenia or to a particular symptom, there have been repeated findings of frontal and temporal abnormalities in schizophrenia, in both resting metabolism and cognitive activation studies. It is becoming increasingly evident that schizophrenic patients show abnormal interactions between brain regions during cognitive tasks. There is currently httle direct evidence in favor of the hypothesis of functional disconnection, and several regions have been found to function abnormally, but with no unequivocal evidence for the consistent involvement of any particular region. However, the majority of positive findings suggest that a disruption of cortico-cortical (or cortico-subcortical) integration is a core feature of the schizophrenic syndrome. Future studies would do well to investigate this possibility, using methods of analyzing functional or effective connectivity to evaluate the influence of one brain region over another.

References American Psychiatry Association. (1987). ''Diagnostic and Statistical Manual of Mental Disorders," DSM-III-R. 3rd ed. American Psychiatry Association, Washington. DC. Andreasen, N., Nasrallah, H. A., Dunn, V., Olson, S. C . Grove. W. M., Ehrhardt, J. C , Coffman, J. A., and Crossett, J. H. (1986). Structural abnormalities in the frontal system in schizophrenia. A magnetic resonance imaging siudy.Arch. Gen. Psychiatry 43(2), 136-144. Andreasen, N. C , Swayze, V. W., 2nd, Flaum, M., Yates, W. R., Arndt, S., and McChesney, C. (1990). Ventricular enlargement in schizophrenia evaluated with computed tomographic scanning. Effects of gender, age, and stage of illness. Arch. Gen. Psvchiatrv 47(11). 1008-1015. Andreasen, N. C , Rezai, K., Alliger, R., Swayze, V. W., Flaum. M., Kirchner, P, Cohen, G., and O'Leary, D. S. (1992). Hypofrontality

541

in neuroleptic-naive patients and in patients with chronic schizophrenia: Assessment with xenon-133 single-photon emission computed tomography and the Tower of London. Arch. Gen. Psychiatry 49, 943-958. Andreasen, N. C, Arndt, S., Swayze, V., 2nd. Cizadlo, T., Flaum, M., O'Leary, D. S.. Ehrhardt, J. C , and Yuh, W. T. (1994). Thalamic abnormalities in schizophrenia visualised through magnetic resonance image averaging. Science 266, 294-298. Arnold, S. E. (1997). The medial temporal lobe in schizophrenia. J. Neuropsychiatry CHn. Neiirosci. 9, 460-470. Barta. P. E.. Pearlson. G. D., Powers, R. E., Richards. S. S.. and Tune, L. E. (1990). Auditory hallucinations and smaller superior temporal gyral volume in schizophrenia. Am. J. Psychiatry 147(11). 1457-1462. Benes, F. M. (1991). Evidence for neurodevelopment disturbances in anterior cingulate cortex of post-mortem schizophrenic brain. Schizophr Res. 5(3), 187-188. Benes, F. M.. Sorensen, L. and Bird, E. D. (1991). Reduced neuronal size in posterior hippocampus of schizophrenic patients. Schizophr Bull. 17(4), 597-608. Berman, K. F, Ostrem. J. L., Randolph, C . Gold. J., Goldberg, T E.. Coppola, R.. Carson, R. E., Herscovitch, P.. and Weinberger, D. R. (1995). Physiological activation of a cortical network during performance of the Wisconsin Card Sorting Test: A positron emission tomography study. Neuropsychologia 33(8), 1027-1046. Bleuler. E. (1987). Dementia praecox or the group of schizophrenias. In "'The Clinical Routes of the Schizophrenia Concept" (J. Cutting and M. Shepherd. Eds.). Cambridge Univ. Press. Cambridge, UK. Breier. A., Buchanan, R. W., Elkashef. A.. Munson. R. C . Kirkpatrick. B., and Gellad. F. (1992). Brain morphology and schizophrenia. A magnetic resonance imaging study of limbic, prefrontal cortex, and caudate structures. Arch. Gen. Psychiatry 49(12), 921-926. Brown. R., Colter. N., Corsellis, J. A., Crow, T J., Frith, C. D., Jagoe, R., Johnstone, E. C , and Marsh, L. (1986). Postmortem evidence of structural brain changes in schizophrenia. Differences in brain weight, temporal horn area, and parahippocampal gyrus compared with affective disorder. Arch. Gen. Psychiatry 43(1). 36-42. Buchanan, R. W.. Vladar, K., Barta, P E., and Pearlson, G. D. (1998). Structural evaluation of the prefrontal cortex in schizophrenia. Am. J. Psychiatry 155(8). 1049-1055. Buchsbaum, M. S.. Yang, S., Hazlett. E., Siegel, B. V.. Jr.. Germans, M., Haznedar, M., OTlaithbheartaigh, S., Wei, T, Silverman, J., and Siever, L. J. (1997). Ventricular volume and asymmetry in schizotypal personality disorder and schizophrenia assessed with magnetic resonance imaging. Schizophr Res. 27(1), 45-53. Busatto. G. F , Costa, D. C , Ell, P J., Pilowsky, L. S., David, A. S., and Kerwin, R. W (1994). Regional cerebral blood flow (rCBF) in schizophrenia during verbal memory activation: A '^"^"^Tc-HMPAO single photon emission tomography (SPET) study. Psychol. Med. 24(2), 463-472. Callicott, J. H., Ramsey. N. F, Tallent, K., Bertolino, A., Knable, M. B., Coppola. R.. Goldberg, T, van Gelderen, P.. Mattay, V. S., Frank, J. A.. Moonen. C. T . and Weinberger, D. R. (1998). Functional magnetic resonance imaging brain mapping in psychiatry: Methodological issues illustrated in a study of working memory in schizophrenia. Neuropsychopharmacology 18(3), 186-196. Cannon.T. D., van Erp,T. G., Huttunen, M., Lonnqvist, J., Salonen, O.. Valanne, L., Poutanen, V. P., Standertskjold-Nordenstam, C. G., Gur, R. E., and Yan, M. (1998). Regional gray matter, white matter, and cerebrospinal fluid distributions in schizophrenic patients, their siblings, and controls. Arch. Gen. Psychiatry 55(12). 1084-1091. ^ Carter. C. S., Perlstein, W., Ganguli, R.. Brar, J., Mintun. M., and Cohen. J. D. (1998). Functional hypofrontality and working memory dysfunction in schizophrenia. Am. J. Psychiatry 155(9), 1285-1287.

542 Chua, S. E., and McKenna, P. J. (1995). Schizophrenia, a brain disease? A critical review of structural and functional cerebral abnormality in the disorder. Br. J. Psychiatry 166, 563-582. Cleghorn, J. M., Garnett, E. S., Nahmias. C . Brown. G. M.. Kaplan. R. D., Szechtman, H., Szechtman, B., Franco. S., Dermer. S. W.. and Cook, P. (1990). Regional brain metabolism during auditory hallucinations in chronic schizophrenia. Br J. Psychiatry 157, 562-570. Coffman, J. A., and Nasrallah, H. A. (1986). Magnetic brain imaging in schizophrenia. In "The Neurology of Schizophrenia" (H. A. Nasrallah and D. R. Weinberger, Eds.), pp. 251-266. Elsevier. Amsterdam. Crawley, J. C , Owens, D. G., Crow,T. J., Poulter, M., Johnstone, E. C . Smith, T, Oldland, S. R., Veall, N., Owen, E, and Zanelli, G. D. (1986). Dopamine D2 receptors in schizophrenia studied in vivo. Lancet 2, 224-225. Crow,T. J. (1980). Positive and negative schizophrenic symptoms and the role of dopamine. Br J. Psychiatry 137, 383-386. Crow, T.J. (1995). Aetiology of schizophrenia: An evolutionary theory. Int. Clin. Psychopharmacol. 10(Suppl. 3), 49-56. Daniel, D. G., Weinberger, D. R., Jones, D. W. Zigun, J. R., Coppola, R., Handel, S., Bigelow, L. B., Goldberg, T. E.. Berman, K. E. and Kleinman, J. E. (1991). The effect of amphetamine on regional cerebral blood flow during cognitive activation in schizophrenia. /. Neurosci. 11,1907-1917. David, A. S., Woodruff, P W., Howard, R., Mellers, J. D., Brammer, M., Bullmore, E., Wright, I., Andrew, C , and Williams, S. C (1996). Auditory haUucinations inhibit exogenous activation of auditory association cortex. NeuroReport 7(4), 932-936. Delay, I , Deniker, P., and Harl, J. M. (1952). Traitement des etats d'exicitation et d'agitation par une methode medicamenteuse derives de rhibernotherapie. Ann. Med. Psychol. 110, 267-273. DeLisi, L. E., Hoff, A. L., Schwartz, J. E., Shields, G. W., Halthore, S. N . Gupta, S. M., Henn, F. A., and Anand, A. K. (1991). Brain morphology in first-episode schizophrenic-like psychotic patients: A quantitative magnetic resonance imaging study. Biol. Psychiatry 29(2), 159-175. DeLisi, L. E., Sakuma, M., Kushner, M., Finer, D. L.. Hoff, A. L., and Crow, T. J. (1997). Anomalous cerebral asymmetry and language processing in schizophrenia. Schizophr Bull. 23(2), 255-271. DeMyer, M. K., Gilmor, R. L., Hendrie, H. C , DeMyer, W. E.. Augustyn, G. T, and Jackson, R. K. (1988). Magnetic resonance brain images in schizophrenic and normal subjects: Influence of diagnosis and education. Schizophr Bull. 14(1), 21-37. Dierks, T, Linden, D. E. J., Jandl, M., Formisano, E., Goebel, R., Lanfermann, H., and Singer, W (1999). Activation of Heschl's gyrus during auditory haflucinations. Neuron 22(3), 615-621. Dolan, R. J., and Fletcher, P. C. (1997). Dissociating prefrontal and hippocampal function in episodic memory encoding. Nature 388, 582-585. Dolan, R. I , Fletcher, P, Frith, C. D., Friston, K. J., Frackowiak. R. S. J., and Grasby, P. J. (1995). Dopaminergic modulation of an impaired cognitive activation in the anterior cingulate cortex in schizophrenia. Nature 378, 180-183. Ebmeier, K. P., Blackwood, D. H., Murray, C , Souza, V., Walker, M., Dougall, N., Moffoot, A. P, O'CarroU, R. E., and Goodwin, G. M. (1993). Single-photon emission computed tomography with ^^""Tc-exametazime in unmedicated schizophrenic patients. Biol. Psychiatry 33(7), 487-495. Farde, L., Wiesel, F A., Hafl, H., Halldin, C , Stone-Elander. S.. and Sedvall, G. (1987). No D2 receptor increase in PET study of schizophrenia. Arch. Gen. Psychiatry 44, 671-672. Feinstein, A., Goldberg, T E., Nowlin, B.. and Weinberger, D. R. (1998). Types and characteristics of remote memory impairment in schizophrenia. Schizophr Res. 30(2), 155-163. ffytche, D. H., Howard, R. J., Brammer, M. J., David. A., Woodruff. P.

IV Psychiatric Disorders and Williams. S. (1998). The anatomy of conscious vision: An fMRI study of visual hallucinations. Nat. Neurosci. 1(8), 738-742. Fletcher, P C , Frith. C. D . Grasby, P M.. Friston, K. J., and Dolan, R. J. (1996). Local and distributed effects of apomorphine on fronto-temporal function in acute unmedicated schizophrenia. / Neurosci. 16(21), 7055-7062. Fletcher. P C , McKenna, P J.. Frith. C. D., Grasby. P M., Friston, K. J., and Dolan. R. J. (1998). Brain activations in schizophrenia during a graded memory task studied with functional neuroimaging. Arch. Gen. Psychiatry 55(11), 1001-1008. Fletcher, P C , McKenna, P J., Friston, K. J., Frith, C. D., and Dolan, R. J. (1999). Abnormal cingulate modulation of fronto-temporal connectivity in schizophrenia. Neurolmage 9, 337-342. Friston, K. J., and Frith, C. D. (1995). Schizophrenia: A disconnection syndrome? Clin. Neurosci. 3, 89-97. Friston, K. J., Liddle, P F, Frith, C. D., Hirsch, S. R.. and Frackowiak, R. S. (1992). The left medial temporal region and schizophrenia. A PET study. Brain 115 (Part 2), 367-382. Friston. K. J., Frith, C D , Liddle, P F , and Frackowiak. R. S. (1993). Functional connectivity: The principal-component analysis of large (PET) data sets. / Cereb. Blood Flow Metab. 13(1), 5-14. Friston. K. L, Buechel, C , Fink, G. R., Morris, J., Rolls, E., and Dolan, R. J. (1997). Psychophysiological and modulatory interactions in neuroimaging. Neurolmage 6(3). 218-229. Frith. C. D. (1992). "The Cognitive Neuropsychology of Schizophrenia." Lawrence Erlbaum, Hove. Frith, C. D., Friston, K., Liddle. P F . and Frackowiak, R. S. (1991). Willed action and the prefrontal cortex in man: A study with PET. Proc. R. Soc. Lond. B. Biol. Sci. 244(1311). 241-246. Frith, C. D.. Friston. K. J., Herold, S., Silbersweig, D., Fletcher, P, Cahill. C , Dolan, R. J., Frackowiak. R. S. J., and Liddle, P F (1995). Regional brain activity in chronic schizophrenic patients during the performance of a verbal fluency task. Br J. Psychiatry 167, 343-349. Fukuzako, H., Yamada, K., Kodama, S.. Yonezawa, T, Fukuzako, T, Takenouchi, K.. Kajiya. Y. Nakajo, M.. and Takigawa, M. (1997). Hippocampal volume asymmetry and age at illness onset in males with schizophrenia. Eur Arch. Psychiatry Clin. Neurosci. 247(5), 248-251. Ganguli, R., Carter, C , Mintun, M.. Brar, J., Becker, J., Sarma, R., Nichols,T, and Bennington, E. (1997). PET brain mapping study of auditory verbal supraspan memory versus visual fixation in schizophrenia. Biol. Psychiatry 41(1). 33-42. Goldman-Rakic, P S., and Selemon. L. D. (1997). Functional and anatomical aspects of prefrontal pathology in schizophrenia. Schizophr Bull. 23, 437-458. Gottesman, L I., and Shields, J. A. (1982). "Schizophrenia: The Epigenetic Puzzle." Cambridge Univ. Press, Cambridge, UK. Green. M. F. (1996). What are the functional consequences of neurocognitive deficits in schizophrenia? Am. J. Psychiatry 153, 321-330. Harvey. I., Ron, M. A., Du Boulay, G., Wicks, D , Lewis, S. W., and Murray. R. M. (1993). Reduction of cortical volume in schizophrenia on magnetic resonance imaging. Psychol. Med. 23(3), 591-604. Heckers, S., Ranch, S. L., Goff, D., Savage, C. R., Schacter, D. L., Fischman, A. J., and Alpert, N. M. (1998). Impaired recruitment of the hippocampus during conscious recollection in schizophrenia. Nat. Neurosci. 1(4), 318-323. Hoffman. R. E. (1986). Verbal hallucinations and language production processes in schizophrenia. Behav. Brain Sci. 9(3), 503-517. Hoksbergen, I., Pickut, B. A.. Marien, P, Slabbynck, H., Kunnen, J., and De Deyn, P P (1996). SPECT findings in an unusual case of visual hallucinosis./ Neurol. 243 (8). 594-598. Honer, W. G., Falkai, P, Young, C , Wang,T, Xie, J., Bonner, J., Hu, L., Boulianne, G. L., Luo, Z.. and Trimble. W. S. (1997). Cingulate cor-

21 Functional Neuroimaging Studies of Schizophrenia tex synaptic terminal proteins and neural cell adhesion molecule in schizophrenia. Neuroscience 78(1), 99-110. Howard, R., David, A., Woodruff, P., Mellers, L. Wright, J., Brammer. M., BuUmore, E., and Williams. S. (1997). Seeing visual hallucinations with functional magnetic resonance imaging. Dement. Geriatr. Cogn. Disord. ^2).13-71. Ingvar, D. H., and Franzen, G. (1974). Distribution of cerebral activity in chronic schizophrenia. Lancet 2, 1484-1486. Iversen, S. D. (1986). Animal models of schizophrenia. In "The Psychopharmacology and Treatment of Schizophrenia" (P. B. Bradlev and S. R. Hirsch, Eds.). Oxford Univ. Press, Oxford, UK. Johnstone, E. C. (1991). Defining characteristics of schizophrenia. Br. J. Psychiatry 13 (SuppL), 5-6. Johnstone, E. C , Crow, T J., Frith, C. D., Husband. X. and Kreel. L. (1976). Cerebral ventricular size and cognitive impairment in chronic schizophrenia. Lancet 2, 924-926. Johnstone. E. C , Owens, D. G., Crow, T. J., Frith, C. D., AlexandropoHs, K., Bydder, G, and Colter, N. (1989). Temporal lobe structure as determined by nuclear magnetic resonance in schizophrenia and bipolar affective disorder. /. Neurol. Neurosiirg. Psvchiatrv 52(6). 736-741. Kelsoe, J. R., Jr., Cadet, J. L.. Pickar. D., and Weinberger. D. R. (1988). Quantitative neuroanatomy in schizophrenia. A controlled magnetic resonance imaging study. Arch. Gen. Psychiatry 45(6), 533-541. Kendler, K. S., and Robinette, C. D. (1983). Schziophrenia in National Academy of Sciences-National Research Council Twin Registry: A 16-year update, ylm. /. Psychiatry 140,1551-1563. Kraepelin, E. (1919). "Dementia Praecox and Paraphrenia.*" Livingstone, Edinburgh. Kuipers, L., and Bebbington, P. (1988). Expressed emotion research in schizophrenia: Theoretical and clinical implications. Psvchol. Med. 18(4), 893-909. Kwon, J. S., McCarley, R. W, Hirayasu. Y., Anderson, J. E.. Fischer. I. A., Kikinis, R., Jolesz, F A., and Shenton. M. E. (1999). Left planum temporale volume reduction in schizophrenia. Arch. Gen. Psychiatry 56(2). 142-148. Laruelle, M. (1998). Imaging dopamine transmission in schizophrenia. A review and meta-analysis. Q. J. Niicl. Med.42(3), 211-221. Lewis, S. W (1990). Computerised tomography in schizophrenia 15 years on. Br J. Psychiatry 9(Suppl.). 16-24. Liddle, P. F. (1987). The symptoms of chronic schizophrenia. A reexamination of the positive-negative dichotomy. Br J. Psychiatry 151,145-151. Liddle, P F , Friston, K. J., Frith, C. D., and Frackowiak. R. S. (1992). Cerebral blood flow and mental processes in schizophrenia. / R. Soc. Med. 85(4), 224-227. Maher, B. A., Manschreck, T C . Yurgelun-Todd, D. A., and Tsuang. M.T. (1998). Hemispheric asymmetry of frontal and temporal gray matter and age of onset in schizophrenia. Biol. Psychiatry 44(6). 413^17. Mathew, R. J., Wilson, W H.,Tant. S. R.. Robinson. L.. and Prakash. R. (1988). Abnormal resting regional cerebral blood flow patterns and their correlates in schizophrenia. Arch. Gen. Psychiatry 45(6), 542-549. Mattay, V. S., Callicott. J. H.. BertoHno. A.. Santha. A. K..Tallent. K. A.. Goldberg, T E.. Frank. J. A., and Weinberger. D. R. (1997). Abnormal functional lateralization of the sensorimotor cortex in patients with schizophrenia. NeiiroReport 8(13). 2977-2984. McGue, M.. Gottesman, L L. and Rao. D. C. (1986). The analysis of schizophrenia family data. Behav. Genet. 16, 75-87. McGuire, P. K., Silbersweig, D. A., Wright, L, Murray. R. M.. Frackowiak. R. S., and Frith, C. D. (1996). The neural correlates of inner speech and auditory verbal imagery in schizophrenia: relationship to auditory verbal hallucinations. Br J. Psychiatry 169(2). 148-159. McGuire. P. K.. Quested, D. J., Spence. S. A.. Murray. R. M.. Frith.

543

C. D.. and Liddle, P F (1998). Pathophysiology of ''positive" thought disorder in schizophrenia. Br J. Psychiatry 173, 231-235. Miller. D. D.. Rezai. K., Alliger. R.. and Andreasen. N. C (1997). The effect of antipsychotic medication on relative cerebral blood perfusion in schizophrenia: Assessment with technetium-99m hexamethyl-propyleneamine oxime single photon emission computed tomography. Biol. Psychiatry 41, 550-559. Nagahama. Y.. Fukuyama. H.. Yamauchi. H.. Matsuzaki. S.. Konishi. J.. Shibasaki. H.. and Kimura, J. (1996). Cerebral activation during performance of a card sorting test. Brain 119(5). 1667-1675. Okubo. Y. Suhara. T . Suzuki, K.. Kobayashi. K.. Inoue. O.. Terasaki. O.. Someya. Y. Sassa. T . Sudo, Y, Matsushima, E.. lyo, M.. Tateno. Y. and Toru. M. (1997). Decreased prefrontal dopamine Dj receptors in schizophrenia revealed by PET. Nature 385(6617). 634-636. Paulman. R. G . Devous. M. D . Sr.. Gregory. R. R., Herman, J. H., Jennings. L.. Bonte. F J.. Nasrallah. H. A., and Raese, J. D. (1990). Hypofrontality and cognitive impairment in schizophrenia: Dynamic single-photon tomography and neuropsychological assessment of schizophrenic brain function. Biol. Psychiatry 27(4). 377-399. Pilowsky. L. S.. Costa. D. C . Ell. P J.. Verhoeff. N. P. Murray. R. M.. and Kerwin. R. W. (1994). D2 dopamine receptor binding in the basal ganglia of antipsychotic-free schizophrenic patients. An •"^MBZM single photon emission computerised tomography study. Br J. Psyc/?/V7f/T 164(1). 16-26. Ragland. J. D . Gur. R. C . Glahn. D. C . Censits. D. M.. Smith. R. J., Lazarev. M. G . Alavi. A., and Gur. R. E. (1998). Frontotemporal cerebral blood flow change during executive and declarative memory tasks in schizophrenia: A positron emission tomography study. Neuropsychology 12(3). 399-413. Raine. A.. Lencz. T . Reynolds. G P, Harrison, G . Sheard. C . Medley. L. Reynolds. L. M., and Cooper. J. E. (1992). An evaluation of structural and functional prefrontal deficits in schizophrenia: MRI and neuropsychological measures. Psychiatry Res. 45(2). 123-137. Raz. S.. and Raz, N. (1990). Structural brain abnormalities in the major psvchoses: A quantitative review of the evidence from computerized imaging. Psychol Bull. 108(1). 93-108. Ron. M. A., and Harvey. L (1990).The brain in schizophrenia./ Neurol. Neurosurg. Psychiatry 53(9). 725-726. Rossi. A.. Stratta. P., Mancini, F.. Gallucci. M.. Mattel. P., Core, L.. Di Michele. V.. and Casacchia. M. (1994). Magnetic resonance imaging findings of amygdala-anterior hippocampus shrinkage in male patients with schizophrenia. Psychiatry Res. 52(1). 43-53. Saykin, A. J.. Shtasel. D. L.. Gur. R. E.. Kester, D. B., Mozley, L. H., Stafiniak, P. and Gur. R. C. (1994). Neuropsychological deficits in neuroleptic naive patients with first-episode schizophrenia. Arch. Gen. Psychiatry 51(2). 124-131. Schacter. D L.. Alpert, N. M.. Savage. C. R., Rauch, S. L., and Albert, M. S. (1996). Conscious recollection and the human hippocampal formation: Evidence from positron emission tomography. Proc. Natl. Acad. Sci. U.S.A. 93, 321-325. Schroder. I . Essig. M.. Baudendistel, K.. Jahn. T. Gerdsen, I., Stockert, A.. Schad. L. R.. and Knopp. M. V. (1999). Motor dysfunction and sensorimotor cortex activation changes in schizophrenia: A study with functional magnetic resonance imaging. Neurolmage 9(1), 81-87. Seeman, P. (1986). Dopamine/neuroleptic receptors in schizophrenia. In "Handbook on Studies of Schizophrenia" ( G D. Burrows, T R. Norman, and G Rubenstein. Eds.). Part 2. Elsevier, Amsterdam. Seeman. P. (1987). Dopamine receptors and the dopamine hypothesis of schizophrenia. Synapse 1,133-152. Shallice.T (1982). Specific impairments of planning. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 25(298). 199-209. Shallice.T. and Burgess, P W. (1991). Deficits in strategy application following frontal lobe damage in man. Brain 114 (Part 2), 727-741. Sharma.T. Lancaster. E.. Lee. D.. Lewis. S.. Sigmundsson,T.,Takei, N.,

544 Gurling, H., Barta, P., Pearlson. G.. and Murray, R. (1998). Brain changes in schizophrenia. Volumetric MR! study of famihes muUiply affected with schizophrenia: The Maudsley Family Study 5. Br. J. Psychiatry 173,132-138. Shenton, M. E., Kikinis, R., Jolesz, F. A., Pollak, S. D.. LeMay. M., Wible, C. G., Hokama, H., Martin, J., Metcalf, D.. Coleman. M., et al. (1992). Abnormahties of the left temporal lobe and thought disorder in schizophrenia. A quantitative magnetic resonance imaging study. K Engl. J. Med. 327(9), 604-612. Siegel, B. V., Jr., Buchsbaum, M. S., Bunney. W. E., Jr.. Gottschalk. L. A., Haier, R. J., Lohr, J. B., Lottenberg, S., Najafi. A.. Nuechterlein, K. H., Potkin, S. G., et al. (1993). Cortical-striatal-thalamic circuits and brain glucose metabolic activity in 70 unmedicated male schizophrenic patients, ylm. /. Psychiatry 150(9), 1325-1336. Silbersweig, D. A., Stern, E., Frith. C . Cahill. C , Holmes. A.. Grootoonk, S., Seaward, I , McKenna, P., Chua. S. E.. Schnorr. L.. et al. (1995). A functional neuroanatomy of hallucinations in schizophrenia. Nature 378,176-179. Smith, G. N., and Iacono,W. G. (1986). Lateral ventricular size in schizophrenia and choice of control group. Lancet 1, 1450. Smith, R. C , Baumgartner, R., and Calderon, M. (1987). Magnetic resonance imaging studies of the brains of schizophrenic patients. P^vchiatry Res. 20(1), 33-46. Spence, S. A., Brooks, D. J., Hirsch, S. R., Liddle, P F . Meehan. J., and Grasby, P. M. (1997). A PET study of voluntary movement in schizophrenic patients experiencing passivity phenomena (delusions of alien control). Brain 120,1997-2011. Spence, S. A., Hirsch, S. R., Brooks, D. J., and Grasby. P M. (1998). Prefrontal cortex activity in people with schizophrenia and control subjects. Evidence from positron emission tomography for remission of "'hypofrontahty'" with recovery from acute schizophrenia. Br. J. Psychiatry 172, 316-323. Stevens, A. A., Goldman-Rakic, P S., Gore, J. C , Fulbright. R. K.. and Wexler, B. E. (1998). Cortical dysfunction in schizophrenia during auditory word and tone working memory demonstrated by functional magnetic resonance imaging. Arch. Gen. Psychiatry 55(12). 1097-1103. Straube, E. R., and Oades, R. D. (1992). In "Schizophrenia: Empirical Research and Findings," Chapter 13. Academic Press, San Diego. Suddath, R. L., Christison, G. W., Torrey, E. K. Casanova, M. F . and Weinberger, D. R. (1990). Anatomical abnormalities in the brains of monozygotic twins discordant for schizophrenia. N. Engl. J. Med. 322(12), 789-794. Suzuki, M., Yuasa, S., Minabe, Y., Murata, M., and Kurachi. M. (1993). Left superior temporal blood flow increases in schizophrenic and schizophreniform patients with auditory hallucination: A longitudinal case study using ^^^MMP SPECT Eur Arch. Psychiatry Clin. Neurosci. 242(5). 257-261. Ungerleider, L. G. (1995). Functional brain imaging studies of cortical mechanisms for memory. Science 270, 769-775. Van Horn, J. D., and McManus, L C. (1992). Ventricular enlargement in schizophrenia. A meta-analysis of studies of the ventricle : brain ratio. Br. J. Psychiatry 160, 687-697. Van Rossum, J. M. (1966). The significance of dopamine receptor blockade for the mechanism of action of neuroleptic drugs. Arch. Int. Pharmacodyn. Then 160, 492-494. Volz, H. P, Gaser, C , Hager, E, Rzanny, R.. Mentzel. H. L. Kreitschmann-Andermahr, I., Kaiser, W. A., and Sauer. H. (1997). Brain activation during cognitive stimulation with the Wisconsin Card Sorting Test: A functional MR! study on healthy volunteers and schizophrenics. Psychiatry Res. 75(3). 145-157.

IV Psychiatric Disorders Waddington. J. L. (1990). Sight and insight: Regional cerebral metabolic activity in schizophrenia visualised by positron emission tomography, and competing neurodevelopmental perspectives. Br J. Psychiatry 156, 615-619. Weinberger. D. R.. Torrey, E. F . Neophytides, A. N., and Wyatt, R. X (1979). Lateral cerebral ventricular enlargement in chronic schizophrenia. Arch. Gen. Psychiatry 36(7), 735-739. Weinberger. D. R.. Berman. K. F . and Zee. R. F (1986). Physiologic dysfunction of dorsolateral prefrontal cortex in schizophrenia. I. Regional cerebral blood flow evidence. Arch. Gen. Psychiatry 43, 114-124. Weinberger. D. R.. Mattay. V. Callicott. J.. Kotrla. K.. Santha. A., van Gelderen. P. Duyn. J.. Moonen. C . and Frank. J. (1996). fMRI applications in schizophrenia research. Neiirolmage 4(3). S118-S126. Wiser. A. K.. Andreasen. N. C . O'Leary. D. S.. Watkins. G. L., Boles Ponto, L. L.. and Hichwa. R. D. (1998). Dysfunctional corticocerebellar circuits cause "cognitive dysmetria" in schizophrenia. NeiiroReport 9(S), 1895-1899. Wolkin. A.. Rusinek. H.. Vaid. G.. Arena. L.. Lafargue.T, Sanfilipo, M., Loneragan, C , Lautin, A., and Rotrosen. J. (1998). Structural magnetic resonance image averaging in schizophrenia. Am. J. Psychiatry 155(8), 1064-1073. Wong, D. F, Wagner, H. N.. Jr., Tune. L. E.. Dannals. R. F. Pearlson. G. D.. Links. J. M.. Tamminga. C. A.. Broussolle. E. P. Ravert, H. T . Wilson. A. A.. Toung. J. K. T . Malat. L, Williams. J. A.. OTuama, L. A.. Snyder. S. H.. Kuhar. M. I . and Gjedde. A. (1986). Positron emission tomography reveals elevated D2 dopamine receptors in drug-naive schizophrenics. Science 234, 1558-1563. Woodruff. P W.. Wright. I. C . Shuriquie. R . Russouw. H.. Rushe, T, Howard, R. J., Graves, M., Bullmore. E. T, and Murray, R. M. (1997a). Structural brain abnormalities in male schizophrenics reflect fronto-temporal dissociation. Psychol. Med. 27(6), 1257-1266. Woodruff. P W, Wright. I. C . Bullmore, E. T, Brammer, M.. Howard, R. L. Wilhams. S. C , Shapleske. J.. Rossell. S., David, A. S., McGuire, P. K.. and Murray. R. M. (1997b). Auditory hallucinations and the temporal cortical response to speech in schizophrenia: A functional magnetic resonance imaging study. Am. J. Psychiatry 154(12), 1676-1682. World Health Organization. (1975). "Schizophrenia: A Multinational Study." WHO. Geneva. Wright. I. C . McGuire. P K.. Poline. J. B..Travere, J. M., Murray, R. M., Frith. C. D.. Frackowiak. R. S.. and Friston. K. J. (1995). A voxelbased method for the statistical analysis of gray and white matter density applied to schizophrenia. Neiirolmage 2(4), 244-252. Young, A. H.. Blackwood. D. H.. Roxborough, H., McQueen, J. K., Martin. M. J., and Kean. D. (1991). A magnetic resonance imaging study of schizophrenia: Brain structure and clinical symptoms. Br J. Psychiatry 158, 158-164. Yurgelun-Todd. D. A.. Renshaw. P F . Gruber. S. A., Ed. M., Waternaux. C . and Cohen. B. M. (1996a). Proton magnetic resonance spectroscopy of the temporal lobes in schizophrenics and normal controls. Schizophn Res. 19(1). 55-59. Yurgelun-Todd. D. A.. Waternaux. C. M.. Cohen. B. M.. Gruber, S. A., English. C. D.. and Renshaw, P F (1996b). Functional magnetic resonance imaging of schizophrenic patients and comparison subjects during word production. Am. J. Psychiatry 153(2). 200-205. Zakzanis. K. K.. and Hansen. K.T (1998). Dopamine D2 densities and the schizophrenic brain. Schizophr Res. 32(3), 201-206. Zeki, S., Watson, J. D., Lueck, C. J., Friston. K. J., Kennard, C , and Frackowiak. R. S. (1991). A direct demonstration of functional specialization in human visual cortex./ Neiirosci. 11(3). 641-649.