Abstract

Functional measures have consistently shown prefrontal abnormalities in schizophrenia. However, structural magnetic resonance imaging (MRI) findings of prefrontal volume reduction have been less consistent. In this study, we evaluated prefrontal gray matter volume in first episode (first hospitalized) patients diagnosed with schizophrenia, compared with first episode patients diagnosed with affective psychosis and normal comparison subjects, to determine the presence in and specificity of prefrontal abnormalities to schizophrenia. Prefrontal gray and white matter volumes were measured from first episode patients with schizophrenia (n = 17), and from genderand parental socio-economic status-matched subjects with affective (mainly manic) psychosis (n = 17) and normal comparison subjects (n = 17), age-matched within a narrow age range (18–29 years). Total (left and right) prefrontal gray matter volume was significantly reduced in first episode schizophrenia compared with first episode affective psychosis and comparison subjects. Follow-up analyses indicated significant left prefrontal gray matter volume reduction and trend level reduction on the right. Schizophrenia patients showed 9.2% reduction on the left and 7.7% reduction on the right compared with comparison subjects. White matter volumes did not differ among groups. These data suggest that prefrontal cortical gray matter volume reduction is selectively present at first hospitalization in schizophrenia but not affective psychosis.

Introduction

Patients with schizophrenia show cognitive and behavioral problems such as deficits in performance on the Wisconsin Card Sorting Test (Weinberger et al., 1986), smooth pursuit eye movement (Holzman et al., 1977) and performance on spatial working memory tasks (Park and Holzman, 1992) that are frequently associated with frontal lobe lesions in humans — see recent review of effects of prefrontal lesions (Knight et al., 1999). Additionally, both lesion and recording data in nonhuman primates implicate the prefrontal cortex in performance on analogs of these tasks, as well as more complex ‘executive’ behavior [reviewed by Miller (Miller, 1999)]. Moreover, both positron emission tomography (PET) and functional magnetic resonance imaging (fMRI) measures in schizophrenia of prefrontal cortex activity related to these tasks are reported to be abnormal (Weinberger et al., 1986), as reviewed by Wible et al. (Wible et al., 1995) and Selemon and Goldman-Rakic (Selemon and Goldman-Rakic, 1999).

It is thus not surprising that there has been considerable interest in structural studies of prefrontal cortex in schizophrenia. Our published review of the MRI structural studies showed that only slightly more than half (55%) of the 33 structural MRI studies of the frontal lobe reported abnormalities (Shenton et al., 1997, 2001; McCarley et al., 1999). Prefrontal gray matter abnormalities were more evident, however, in studies that separated gray matter from white matter and that examined separate regions of cortex (Schlaepfer et al., 1994; Sullivan et al., 1998; Gur et al., 2000).

Inconsistent findings for prefrontal volume change would also suggest the importance of adequate control of potential confounds such as age or factors associated with chronic illness, including chronic medication. MRI prefrontal gray matter volume is known to decline with age, and data suggests it does so in an accelerated manner compared with other regions (Pfefferbaum et al., 1998). Chronic medication is another potential confound. Selemon and co-workers have recently reported glial proliferation and hypertrophy of the prefrontal cortex (area 46) in rhesus monkeys given typical or atypical antipsychotic drugs for 6 months (Selemon et al., 1999). If this hypertrophy were present in humans on neuroleptics, then chronic neuroleptic dosage might reduce the opportunity to measure a schizophrenia-related decline in gray matter.

Use of a first psychotic episode population of schizophrenic subjects may thus be useful in controlling for both potential confounds. First, age distribution in first episode subjects can be more easily chosen to span a more limited and younger range than in chronic patients and, second, chronic neuroleptic exposure in this population is rare.

MRI frontal lobe abnormalities have been reported in first episode schizophrenia, but frontal gray and white matter have not generally been separately evaluated. Zipursky et al. (Zipursky et al., 1998) found whole brain gray matter reduced in first episode schizophrenia. Gur et al. (Gur et al., 1998) found whole frontal lobe volume significantly decreased in both first episode and chronic schizophrenics compared with controls. Nopoulos et al. (Nopoulos et al., 1995) showed a significant decrement in whole frontal lobe tissue in first episode patients compared with controls. However, to our knowledge, no MRI studies have separately evaluated whole prefrontal gray and white matter volumes in first episode schizophrenia — although Gur et al. (Gur et al., 2000) looked at dorsal and orbital prefrontal subdivisions.

A further advantage of a first episode sample is the ease with which comparisons to first episode psychotic affective disorder can be made. This appears important because it remains uncertain, and even controversial, whether the psychosis associated with affective disorder and that associated with schizophrenia represent manifestations of different disorders or are variants of a single underlying disorder with somewhat different clinical manifestations. MRI studies evaluating both first episode affective and schizophrenic patients may assist in answering this question. We note that MRI studies of affective disorder have reported volume reductions in several regions, including frontal lobe, temporal lobe and putamen, although these findings have been described as somewhat inconsistent (Dougherty and Rauch, 1997; Shenton et al., 1997, 2001; McCarley et al., 1999; Pearlson and Marsh, 1999). We have previously reported (Hirayasu et al., 1998) that first episode schizophrenic patients were different from first episode affective patients and normal comparison subjects in having a smaller gray matter volume in a posterior superior temporal gyrus (STG), and a significant left smaller than right asymmetry. We also reported that first episode affective patients, but not first episode schizophrenia patients, showed a reduction in the subgenual cingulate cortex (Hirayasu et al., 1999). Schlaepfer and co-workers reported gray matter volume reduction in heteromodal association cortex including prefrontal cortex in schizophrenia but not in bipolar disorder (Schlaepfer et al., 1994). However, to our knowledge, no studies have evaluated similarities or differences of prefrontal gray matter and white matter in first episode schizophrenia and affective psychosis.

The present study used high spatial resolution MRI to evaluate prefrontal gray and white matter in young (<30 years old) first episode schizophrenia patients, first episode affective psychosis patients (mainly manic) and normal comparison subjects. We predicted that first episode patients with schizophrenia, but not first episode patients with affective psychosis, would show reduced prefrontal volumes. Consistent with the hypothesis of reduced neuropil but not cell count (Selemon and Goldman-Rakic, 1999), we predicted reductions in gray but not white matter.

Materials and Methods

Subjects

Patients were recruited from the inpatient population at McLean Hospital, a private psychiatric hospital in Belmont, Massachusetts. The total subject population from which the study sample was drawn consisted of 23 first episode patients with schizophrenia and 30 first episode patients with affective psychosis diagnosed based on The Structured Clinical Interview for Diagnosis–DSM IIIR (SCID) (Spitzer et al., 1990a), review of hospital course and medical records. Thirty normal comparison subjects (SCID NP, SCID II) (Spitzer et al., 1990b,c) were recruited through newspaper advertisements. Patients and comparison subjects met criteria for age (18–29), IQ above 75, right handedness (Oldfield, 1971), and negative history of seizures, head trauma with loss of consciousness, neurological disorder, no lifetime history of alcohol or drug dependence or current abuse (SCID criteria) (Spitzer et al., 1990a). The comparison subjects had no Axis I mental disorder in themselves (SCID NP), nor in their firstdegree relatives (by self-report).

Seventeen triads of subjects were selected with the primary criterion of as close as possible age match. All subjects were younger than 30 years old, so as to constitute a restricted age range. Where multiple affective or comparison subjects were tied in age to a schizophrenic subject, closest match on parental socio-economic status (SES) was used as a secondary criterion. Selection was performed blinded to MRI results. The resulting match showed the schizophrenic subjects were matched within 1 year of age to 14 (82%) of the pairs of affective patients and comparison subjects. The maximum age difference between members of the subject triads was 2 years. Each subject group included two female subjects. Twenty-four patients (11 affective and 13 schizophrenic subjects) were tested within 4 weeks of their first hospitalization, nine (six affective and three schizophrenic subjects) within 4 months of first hospitalization, and one schizophrenic subject within 11 months. Demographic data for subjects are summarized in Table 1.

The affective psychosis patient group included 13 bipolar (manic) disorder and four major depressive disorder patients. (The statistically significant results reported below remained the same with the exclusion of the four major depressive patients.) Diagnoses were confirmed via a 1-year follow-up interview. Table 1 shows the short median duration of any psychotropic medication prior to MRI acquisition (seven schizophrenics and 10 affective psychotic patients were unmedicated prior to admission). In terms of the time of onset of psychosis, we selected the date of first psychiatric hospitalization as a reliable marker. Although the time of onset of prodromal symptoms may be a better indicator of the actual onset of the disease, in practice this retrospective measure is difficult to verify and may be unreliable. Thus, for comparative purposes, we have provided age at time of first medication, which we believe to be a more objective measure (most medication onset dates were taken from hospital records), less equivocal than time of prodrome onset, and thus to be a better index of the severity and timing of symptom onset. The present study included 11 individuals in each subject group who were also included in our previous temporal lobe MRI study (Hirayasu et al., 1998).

The Mini-Mental State Examination was used to rule out any dementia or delirium. Additionally, the Information subscale of the WAIS-R (Wechsler Adult Intelligence Scale – Revised) (Wechsler, 1981) was used as a gross estimate of general fund of information, and the digits-forward and digits-backward subscales of the WAIS-R were used to evaluate immediate/short-term memory, attention and concentration. SES and parental SES (Hollingshead, 1965) were also assessed. Evaluation of psychotic features was made using the Brief Psychiatric Rating Scales (BPRS) (Overall and Gorham, 1962) for most patients. For six of the first episode schizophrenia patients and four of the affective psychosis patients, BPRS scores were calculated from the Positive and Negative Syndrome Scale (PANSS) (Kay et al., 1986), which includes all of the BPRS items. Overall functioning was evaluated using the Global Assessment Scale (GAS) (Endicott et al., 1976). After complete description of the study, written informed consent was obtained from all subjects. Subjects were paid for participating.

MRI Methodology

Magnetic resonance images were acquired with a 1.5 T General Electric scanner (GE Medical Systems, Milwaukee, WI). The acquisition protocol had two MRI sequences. The first was a coronal series of contiguous spoiled gradient (SPGR) images (124 slices of 1.5 mm thickness, voxel dimensions, 0.9375 × 0.9375 × 1.5 mm). The second acquisition resulted in an axial series of contiguous double-echo images (proton density and T2-weighted, with 3.0 mm slice thickness, voxel dimensions, 0.9375 × 0.9375 × 3.0 mm). This latter protocol was also used to assess wholebrain volume. Details are described elsewhere (Dickey et al., 2000). T2 information from the double-echo spin echo axial slices was registered with data from SPGR images by reformatting the axial voxels to the voxel dimensions corresponding to those in the coronal SPGR images. The intensity information from both the SPGR and T2 images was then used in a fully automated segmentation program to classify tissue into gray matter, white matter and CSF. An iterative expectation-maximization (EM) algorithm initially estimated image intensity inhomogeneities, applied intensity corrections based these estimates, and then classified tissue based on the same set of signal intensity parameters for all subjects (Wells et al., 1996). An anisotropic diffusion filter (k = 14 for SPGR and 90 for proton/T2 images, iteration = 3) was applied to reduce noise prior to processing each set of scans (Kikinis et al., 1990). The results of the segmentation were then superimposed on the coronal SPGR image and edited on the computer workstation to assign the left and right prefrontal gray/white matter to separate tissue classes.

Prefrontal Definition

Definition of the prefrontal cortex has been described in detail elsewhere (Wible et al., 1995). Briefly, the gray matter of the prefrontal cortex was measured starting anteriorly from the first slice that contained brain tissue. The posterior landmark was determined by first locating the most anterior slice that contained the temporal stem (the white matter tract connecting the temporal and frontal lobes), then moving anteriorly three slices. Anteriorly, the white matter was measured beginning with the first slice that contained white matter and extended posteriorly to the slice immediately anterior to the slice that contained the lateral ventricles. Prefrontal segmented images on a coronal slice are shown in Fig. 1A, and Fig. 1BD illustrate the anterior–posterior boundaries of the gray and white matter regions of interest (ROI). Inter-rater reliability was computed for the ROIs by three independent raters (S.T., J.J.L. and M.D.) who were blind to group membership. Ten cases were randomly selected for inter-rater reliability. Intraclass correlation coefficients for inter-rater reliability for the three raters were: 0.996 for left prefrontal gray, 0.998 for right prefrontal gray, 0.999 for left prefrontal white, and 0.999 for right prefrontal white.

Statistical Analyses

Three-factor ANOVA was performed with GROUP (schizophrenia, affective psychosis, control) as a between-subjects factor, and TISSUE (gray, white) and SIDE (left, right) as within-subjects factors. Statistics were performed based on relative volumes [ROI volume/volume of intra-cranial contents (ICC)] × 100. (Testing absolute volumes with ICC as a covariate did not change the results reported below.) Follow-up analyses included a two-factor (GROUP and SIDE) ANOVA for each tissue (gray and white), one-way ANOVA, and post hoc Tukey's Honestly Significant Difference (HSD) tests for individual regions with left and right compared separately. Results were considered significant at P<0.05. In addition, an asymmetry coefficient (AC) was calculated by using the formula (RL)/0.5(R + L), where R and L were the prefrontal volumes on the right and left sides in subjects, respectively.

Correlations of prefrontal structures using absolute and relative volumes were performed with the demographic data of Table 1. Correlations were also performed for schizophrenic and affective patients between the syndrome factors of the BPRS and prefrontal left and right gray and white matter to assess the degree of association between symptomatology and prefrontal volume measurements with absolute volumes and relative volumes.

Results

There were no significant group differences in age or parental SES. First episode schizophrenia patients showed significantly lower socio-economic status than comparison subjects, consistent with reduced premorbid functioning. There were no significant differences between schizophrenia and affective patients on any of the clinical scale measures, age of first medication, medication dosage, or duration of illness. The age, SES, parental SES, age of first medication, duration of medication, and dose (chlorpromazine equivalent) of medication did not correlate with the volume of any prefrontal lobe ROI in the patients. The WAIS-R Information score was correlated with right white matter volume in first episode schizophrenia (rho = 0.755, d.f. = 16, P = 0.001 for absolute volume; rho = 0.723, d.f. = 16, P = 0.002 for relative volume).

Although there were no significant differences on BPRS scores between schizophrenia and affective psychosis, the schizophrenic patients showed significantly higher scores for items of hallucination and suspiciousness compared with affective patients. Total BPRS score significantly correlated with left white matter volume in first episode schizophrenia (rho = 0.542, d.f. = 16, P = 0.030 for absolute; rho = 0.620, d.f. = 16, P = 0.010 for relative volume). We emphasize that correlations are reported only if values were significant using both absolute and relative volume measures.

There was no significant difference in ICC volume among the three groups. The results of the main three factor ANOVA are presented in Table 2, and all absolute and relative gray and white matter volumes for each group and follow-up analyses are presented in Table 3. The main ANOVA showed a significant difference in total cortical gray and white matter among the three groups [F(2,48) = 3.60, P = 0.035]. Post hoc tests revealed that the schizophrenia patients showed smaller total relative volume (13.85) compared with affective patients (14.95) and control subjects (14.95). There was more gray matter than white matter volume for all three groups [F(1,48) = 2220.09, P<0.001]. However, groups differed in the relative amounts of gray to white matter at the trend level, evinced by a GROUP by TISSUE interaction [F(2,48) = 3.00, P = 0.059]. Analysis of total gray matter (left plus right) revealed significant differences between groups [F(2,48) = 3.78, P = 0.030; see Fig. 2], with schizophrenia patients showing smaller gray matter volume than affective patients or comparison subjects (see Table 3). By contrast, analysis of total white matter (left plus right) revealed only fractionally smaller white matter volume in the schizophrenia group that was non-significant (see Table 3).

The main ANOVA revealed significantly more tissue in the right hemisphere for all three groups [F(1,48) = 8.76, P = 0.005]. The ratio of gray to white matter differed in the hemispheres, with relatively more gray on the left and more white on the right [F(1,48) = 38.87, P<0.001]. In addition, in the main ANOVA, there was a significant GROUP × TISSUE × SIDE interaction [F(2,48) = 3.39, P = 0.042], suggesting that the differences between groups in the ratio of gray to white matter described above (GROUP by TISSUE and TISSUE by SIDE) was different in the two hemispheres. In the left hemisphere, groups were significantly different in total tissue [F(2,48) = 4.42, P = 0.017]. All three groups showed greater gray matter than white volume in the left hemisphere [F(1,48) = 2524.39, P<0.001], but again this ratio differed among groups [F(2,48) = 3.76, P = 0.030]. One-way ANOVA showed significant volume reduction in left prefrontal gray matter in first episode schizophrenia patients compared with first episode affective patients and normal comparison subjects [F(2,48) = 4.64, P = 0.014], but not in left white matter (see Table 3). In the right hemisphere, there was a trend for differences in total gray plus white matter volume [F(2,48) = 2.76, P = 0.074]. All three groups showed more gray matter than white matter volume in the right hemisphere [F(1,48) = 1819.11, P < 0.001]. There was no GROUP by TISSUE interaction on the right. Despite the lack of statistical significance on the right, follow-up analyses are presented in Table 3 for comparative purposes. There was no significant difference of AC among three groups for gray or white matter volumes.

Discussion

In this study, first episode patients with schizophrenia showed a significant volume reduction in total (left + right) prefrontal gray matter volume compared with first episode patients with affective psychosis and normal comparison subjects. Compared with comparison subjects, the percent reduction for total gray matter was 8.4% with an effect size of 0.69 [(schizophrenic – control subjects' volume)/SD of controls]. The percent reduction on the left (9.2% with an effect size of 0.81) was slightly greater than on the right (7.7% with an effect size of 0.59). The left difference showed statistical significance while the right difference showed only trend level statistical significance. We regard these data as providing suggestive evidence of a slight difference between left and right volume reductions, although the absence of a GROUP × SIDE interaction for gray matter indicates this must be a tentative conclusion.

It is also of interest to compare first episode schizophrenia with first episode affective psychosis (mainly manic) on gray matter volumes. Compared with this group, schizophrenic subjects on the left showed a 8.7% decrease and an effect size of 0.77 (the SD of control subjects was used in this computation). On the right there was a 8.2% decrease with an effect size of 0.63. No difference was found in prefrontal white matter volume between first episode schizophrenic patients, first episode affective patients or normal comparison subjects.

The subjects in this study were selected from larger samples in order to provide closeness of age matching and a narrow age range in a relatively young (< age 30) sample. These findings suggest that prefrontal cortical gray matter abnormalities may be present at first hospitalization in schizophrenic but not in affective psychosis patients.

Overall, in reviewing the MRI literature, MRI abnormalities in the frontal lobe in schizophrenia have been somewhat inconsistent (Shenton et al., 1997, 2001; McCarley et al., 1999). In fact, our own previous study of chronic schizophrenia (Wible et al., 1995) found no difference. Several factors may be involved in this inconsistency of findings.

First, reductions in frontal volume in schizophrenia may be just at the threshold for detection. For example, the quantitative post-mortem study of Selemon et al. (Selemon et al., 1998) found but a small reduction in prefrontal cortical thickness in schizophrenia, a reduction that was not statistically significant (and thus comparable to the MRI negative results), although significant abnormalities in density of various cell types were present. Thus, these post-mortem data would be consistent with the observed findings in MRI studies, since random variation in subjects and their prior history, as well as variation in imaging technology could easily account for the split in findings for the prefrontal structures. Our previous reviews have extensively discussed MRI technical factors that may be responsible for such variation (Shenton et al., 1997, 2001; McCarley et al., 1999).

Second, it is likely that control of age factors may be useful, since prefrontal gray matter decreases with age (Pfefferbaum et al., 1994) and there is the possibility that this decrease may vary in a non-linear manner, with an increased rate of decline with increasing age. Our choice of a young subject group with a narrow age range (18–29) and with a 1:1 age-matching strategy from a larger sample may have decreased the variance associated with age variations. It can be seen that our coefficient of variation (SD/mean) for relative gray matter is small for schizophrenic subjects, 0.90 for whole, 0.92 for left and 0.87 for right gray matter. While statistical correction methods can be applied for age differences, it is a statistical truism that such correction methods necessitate a model for age correction, and both the parameters of the model and/or the nature of the model itself (usually linear) may not be precise and thus contribute additional variance. Thus, we think that the advantage of age matching and the consequent increase in effect size more than outweighs the reduction in subject number it entailed.

Third, there is the possibility that prefrontal structural MRI volumes may be influenced by chronic medication. The most direct animal evidence comes from Selemon et al. (Selemon et al., 1999) who gave typical or atypical antipsychotic drugs orally to monkeys for 6 months, and then used a stereologic method to assess neuronal and glial density and cortical thickness in prefrontal area 46. While neuronal density in drug-treated monkeys and their controls did not differ, glial density was elevated in monkeys that received antipsychotic medications. This was as much as 33% in layers that receive dense excitatory afferents (layers I in typicaland IV in atypical-treated monkeys). In addition, layer V was wider in all drug-treated monkeys. There are no comparable prefrontal cortex studies in humans, but data from basal ganglia have indicated typical neuroleptics are associated with a reversible increase in MRI volume (Chakos et al., 1995). Thus, a possible confound in chronic studies is drug-associated structural changes. While some subjects in the present study had a brief exposure to neuroleptics, neither duration nor dosage was correlated with ROI volumes. Moreover, while schizophrenics had somewhat more exposure to neuroleptics than the other groups, the volume reduction in schizophrenia vs affective disorder and controls was in the opposite direction from that expected as a drug effect, based on the animal studies.

The present study is one of a few MRI structural studies of the prefrontal region in first episode schizophrenia. A recent study by Szeszko et al. (Szeszko et al., 1999) reported that male patients with first episode schizophrenia (sample had 10 male and 9 female subjects) had significantly larger whole (gray + white) right orbital frontal volume compared with left orbital frontal volume and compared with healthy men (Szeszko et al., 1999). In a study of chronic schizophrenia, Buchanan et al. (Buchanan et al., 1998) showed selective gray matter volume reductions in the right and left inferior prefrontal cortex (12 male and 6 female subjects). Additionally, these patients showed decreased prefrontal total white matter and total volumes although there was no significant difference in prefrontal total gray matter volumes (reduction was 5% for left and 6.7% for right). Gur et al. studied 20 first episode and 20 chronic patients and found whole (gray + white) frontal lobe volume significantly decreased in both first episode and chronic schizophrenics compared with controls(Gur et al., 1998). Nopoulos et al. studied 12 male and 12 female first episode patients with schizophrenia and controls; their automated analysis revealed that the patient group had a significant regionally specific decrement in whole (gray + white) frontal lobe tissue compared with controls (Nopoulos et al., 1995).

Functionally, the frontal lobe has been intensely studied and associations with functions disturbed in schizophrenia have been reported, as described in the Introduction. In contrast, structural MRI–clinical correlations have been relatively sparse, perhaps due to the small extent of the frontal volume change and the failure to segment into gyri, which might have stronger correlations. Several reports have suggested associations between frontal abnormalities and negative symptoms. Supporting this is our laboratory's finding (Wible et al., 1995) of Scale for the Assessment of Negative Symptoms scores having a negative correlation with left prefrontal cortical white matter volume in chronic schizophrenia. However, findings have been inconsistent, since many studies have failed to find a correlation between negative symptoms and frontal volumes, and some results have suggested the opposite, e.g. Buchanan et al.'s finding that non-deficit chronic patients had significantly smaller right and left total prefrontal volumes than did deficit patients (Buchanan et al., 1993).

The present first episode study did not find a significant correlation between prefrontal volumes and clinical symptoms except between total BPRS and left white matter volume. However, this was one of multiple correlations, and should be replicated before considering it a firm finding. In addition to the factors just discussed, we hypothesize that an important additional factor might be the unstable nature of symptoms and signs in the early stages of schizophrenia, compared with a more stable clinical picture in chronic patients. Obviously, a test of this hypothesis would require a longer duration follow-up of the present sample.

Parcellation of the prefrontal region is potentially important because this area is both functionally and structurally heterogeneous. It is also worth noting that few studies have been reported for evaluation of individual frontal gyral gray matter in schizophrenia, and it may be that this approach, for which the technical methodology has been developed (Wible et al., 1997), may show changes in volumes in some gyri but not in others. Gur et al. have recently investigated 70 patients with schizophrenia including 29 neuroleptic-naive cases (Gur et al., 2000). They found reduced prefrontal gray matter volume for the dorsolateral area in men (9%) and women (11%), for the dorsomedial area only in men (9%), and for orbital regions only in women (23% and 10% for lateral and medial, respectively). Medication also did not affect the reduction since the differences were evident in neuroleptic-naive patients. The specificity of MR abnormalities to schizophrenic vs affective psychosis is a long-standing conundrum, as is the more general question of whether these psychoses are manifestations of a single basic disorder or are distinct. Few first episode studies have compared prefrontal gray matter volumes between schizophrenia and affective psychosis although several studies suggested global and/or regional cortical volume reduction may be specific to schizophrenia relative to affective disorder. Schlaepfer et al. reported that chronic bipolar patients did not show heteromodal cerebral cortex volume reduction while gray matter volume was reduced in chronic schizophrenic patients (Schlaepfer et al., 1994). Zipursky et al. similarly found regional volume reduction in gray matter in schizophrenic patients but not in non-first episode bipolar patients, as well as a decrease in global gray matter volume in schizophrenia (Zipursky et al., 1997). Harvey et al. also reported a decrease in cortical volume in chronic schizophrenic patients, but not in bipolar patients (Harvey et al., 1994).

In conclusion, our findings of volume reduction in prefrontal gray matter in first episode schizophrenia suggest the presence of structural abnormalities in regions associated with cognitive and behavioral impairment in this disorder. Our sample gender distribution did not allow us to contrast findings based on this important variable. Finally, we suggest that first episode studies provide a major advantage in having fewer confounds from chronicity variables and, additionally, offer a baseline for longitudinal studies evaluating the presence or absence of progression of MRI changes over time.

Notes

This work was supported, in part, by grants from the National Institute of Mental Health (NIMH 01110 and 50740 to M.E.S.; NIMH 40977 to R.W.M.); by VA MERIT Awards (M.E.S., R.W.M.), by a Schizophrenia Center Award from the Department of Veterans Affairs (R.W.M.); by a NARSAD Young Investigator Award (D.F.S.); and by an NAC Biomedical Technology Resource Grant (P41 RR13218-01) and Core Segmentation Tools for Computer Assisted Surgery grant (1R01 RR11747-01A to R.K. and F.J.).

Address correspondence to Drs Robert W. McCarley/Martha E. Shenton, Department of Psychiatry (116A), VA Boston Healthcare System, Brockton Division, Harvard Medical School, 940 Belmont Street, Brockton, MA 02301, USA. Email: robert_mccarley@hms.harvard.edu; martha_shenton@hms.harvard.edu .

Table 1

Subject characteristics and test scoresa

 Schizophrenic patients (n=17) Affective patients (n=17) Normal subjects (n=17) F d.f. P 
Handedness was assessed by a modified Oldfield Inventory (Oldfield, 1971). 
aStatistical significance was determined by a one-factor ANOVA. 
bHigher scores indicate lower SES. 
cPost hoc Tukey's Honestly Significant Difference (HSD) test showed significant difference (P < 0.05) compared with normal subjects. 
Age (range)  22.8 ± 3.6 (18–29)  22.6 ± 3.8 (18–29) 22.2 ± 3.8 (18–27) 0.119 2,48 0.888 
Male/female  15/2  15/2 15/2 – – – 
Handedness 0.8 ± 0.2 0.7 ± 0.2  0.8 ± 0.2 0.634 2,48 0.535 
SESb 3.5 ± 1.2c 2.9 ± 1.4  2.1 ± 0.9 6.97 2,48 0.002 
Parental SESb 2.1 ± 0.9 1.5 ± 0.8  1.4 ± 0.8 0.843 2,48 0.437 
Mini-Mental  27.9 ± 2.9  28.6 ± 1.8 28.9 ± 1.3 1.008 2,48 0.373 
WAIS-R Information  19.1 ± 5.8  20.3 ± 5.4 22.6 ± 4.2 1.945 2,48 0.155 
WAIS-R digits-forward 8.6 ± 2.3 8.4 ± 2.4  9.4 ± 2.4 0.932 2,48 0.401 
WAIS-R digits-backward 7.1 ± 2.6 6.7 ± 2.3  8.6 ± 2.4 2.687 2,48 0.079 
GAS  34.7 ± 10.0  37.6 ± 11.7 – 0.608 1,32 0.442 
Age first medicated  22.6 ± 3.7  22.3 ± 3.8 – 0.036 1,32 0.965 
Median duration of medication (months) 1.3 0.0 – – – – 
Medication (CPZ equiv.) 200 ± 228 177 ± 43 – 0.138 1,32 0.713 
BPRS total  38.5 ± 11.4  37.5 ± 11.8 – 0.058 1,31 0.812 
 Schizophrenic patients (n=17) Affective patients (n=17) Normal subjects (n=17) F d.f. P 
Handedness was assessed by a modified Oldfield Inventory (Oldfield, 1971). 
aStatistical significance was determined by a one-factor ANOVA. 
bHigher scores indicate lower SES. 
cPost hoc Tukey's Honestly Significant Difference (HSD) test showed significant difference (P < 0.05) compared with normal subjects. 
Age (range)  22.8 ± 3.6 (18–29)  22.6 ± 3.8 (18–29) 22.2 ± 3.8 (18–27) 0.119 2,48 0.888 
Male/female  15/2  15/2 15/2 – – – 
Handedness 0.8 ± 0.2 0.7 ± 0.2  0.8 ± 0.2 0.634 2,48 0.535 
SESb 3.5 ± 1.2c 2.9 ± 1.4  2.1 ± 0.9 6.97 2,48 0.002 
Parental SESb 2.1 ± 0.9 1.5 ± 0.8  1.4 ± 0.8 0.843 2,48 0.437 
Mini-Mental  27.9 ± 2.9  28.6 ± 1.8 28.9 ± 1.3 1.008 2,48 0.373 
WAIS-R Information  19.1 ± 5.8  20.3 ± 5.4 22.6 ± 4.2 1.945 2,48 0.155 
WAIS-R digits-forward 8.6 ± 2.3 8.4 ± 2.4  9.4 ± 2.4 0.932 2,48 0.401 
WAIS-R digits-backward 7.1 ± 2.6 6.7 ± 2.3  8.6 ± 2.4 2.687 2,48 0.079 
GAS  34.7 ± 10.0  37.6 ± 11.7 – 0.608 1,32 0.442 
Age first medicated  22.6 ± 3.7  22.3 ± 3.8 – 0.036 1,32 0.965 
Median duration of medication (months) 1.3 0.0 – – – – 
Medication (CPZ equiv.) 200 ± 228 177 ± 43 – 0.138 1,32 0.713 
BPRS total  38.5 ± 11.4  37.5 ± 11.8 – 0.058 1,31 0.812 
Table 2

Effects and interactions of factors for three-factor ANOVA based on relative volumes

 F d.f. P 
Three-factor ANOVA was performed with GROUP (schizophrenia, affective psychosis, control) as a between factor, and TISSUE (gray, white) and SIDE (left, right) as within factors. Statistics were performed based on relative volumes [ROI volume/volume of intra-cranial contents (ICC) × 100]. 
GROUP 3.60 2,48 0.035 
TISSUE 2220.09 1,48 <0.001 
SIDE 8.76 1,48 0.005 
SIDE × GROUP 0.691 2,48 0.506 
TISSUE × GROUP 3.00 2,48 0.059 
SIDE × TISSUE 38.87 1,48 <0.001 
SIDE × TISSUE × GROUP 3.39 2,48 0.042 
 F d.f. P 
Three-factor ANOVA was performed with GROUP (schizophrenia, affective psychosis, control) as a between factor, and TISSUE (gray, white) and SIDE (left, right) as within factors. Statistics were performed based on relative volumes [ROI volume/volume of intra-cranial contents (ICC) × 100]. 
GROUP 3.60 2,48 0.035 
TISSUE 2220.09 1,48 <0.001 
SIDE 8.76 1,48 0.005 
SIDE × GROUP 0.691 2,48 0.506 
TISSUE × GROUP 3.00 2,48 0.059 
SIDE × TISSUE 38.87 1,48 <0.001 
SIDE × TISSUE × GROUP 3.39 2,48 0.042 
Table 3

Absolute (ml) and relative volumes (± SD) for prefrontal gray and white matter in first episode schizophrenia, first episode affective psychosis and normal controls

 Schizophrenic patients (n=17) Affective patients (n=17) Normal subjects (n=17) F d.f. P 
Relative volumes (% total ICC) are in parentheses. For absolute volumes, statistical significance levels are computed using one-factor ANCOVA on absolute volumes with ICC as a covariate. Note the similarity of P values compared with the use of relative volumes (% ICC) and one-factor ANOVA. 
aPost hoc Tukey's Honestly Significant Difference (HSD) tests indicated that schizophrenia patients were significantly different from both affectives and normal subjects. 
Total intracranial contents 1513 ± 117 1506 ±142 1469 ±154 0.50 2,48 0.612 
Total gray matter 155.0±19.4 164.3±19.1 168.8±25.7 3.63 2,47 0.034 
 (10.26±1.14) (11.21±0.94) (11.20±1.36) 3.78 2,48 0.030 
Left gray matter 77.5±9.6 82.4±9.3 85.1±12.7 4.57 2,47 0.015 
 (5.13±0.56a(5.62±0.45) (5.65±0.64) 4.37 2,48 0.010 
Right gray matter 77.5±10.0 81.9±9.9 83.7±13.2 2.75 2,47 0.074 
 (5.13±0.59) (5.59±0.51) (5.56±0.73) 2.97 2,48 0.061 
Total white matter 54.6±9.1 55.2±10.7 56.7±9.1 0.92 2,47 0.407 
 (3.59±0.44) (3.74±0.50) (3.75±0.41) 0.66 2,48 0.520 
Left white matter 26.1±4.4 26.9±5.1 26.7±4.2 1.43 2,47 0.249 
 (1.72±0.22) (1.83±0.25) (1.77±0.19) 1.11 2,48 0.336 
Right white matter 28.5±5.0 28.2±5.9 30.0±5.3 0.79 2,47 0.460 
 (1.88±0.25) (1.91±0.28) (1.98±0.26) 0.74 2,48 0.481 
 Schizophrenic patients (n=17) Affective patients (n=17) Normal subjects (n=17) F d.f. P 
Relative volumes (% total ICC) are in parentheses. For absolute volumes, statistical significance levels are computed using one-factor ANCOVA on absolute volumes with ICC as a covariate. Note the similarity of P values compared with the use of relative volumes (% ICC) and one-factor ANOVA. 
aPost hoc Tukey's Honestly Significant Difference (HSD) tests indicated that schizophrenia patients were significantly different from both affectives and normal subjects. 
Total intracranial contents 1513 ± 117 1506 ±142 1469 ±154 0.50 2,48 0.612 
Total gray matter 155.0±19.4 164.3±19.1 168.8±25.7 3.63 2,47 0.034 
 (10.26±1.14) (11.21±0.94) (11.20±1.36) 3.78 2,48 0.030 
Left gray matter 77.5±9.6 82.4±9.3 85.1±12.7 4.57 2,47 0.015 
 (5.13±0.56a(5.62±0.45) (5.65±0.64) 4.37 2,48 0.010 
Right gray matter 77.5±10.0 81.9±9.9 83.7±13.2 2.75 2,47 0.074 
 (5.13±0.59) (5.59±0.51) (5.56±0.73) 2.97 2,48 0.061 
Total white matter 54.6±9.1 55.2±10.7 56.7±9.1 0.92 2,47 0.407 
 (3.59±0.44) (3.74±0.50) (3.75±0.41) 0.66 2,48 0.520 
Left white matter 26.1±4.4 26.9±5.1 26.7±4.2 1.43 2,47 0.249 
 (1.72±0.22) (1.83±0.25) (1.77±0.19) 1.11 2,48 0.336 
Right white matter 28.5±5.0 28.2±5.9 30.0±5.3 0.79 2,47 0.460 
 (1.88±0.25) (1.91±0.28) (1.98±0.26) 0.74 2,48 0.481 
Figure 1.

(A) Coronal slice (1.5mm) through the prefrontal region of a normal control subject. Prefrontal gray matter is outlined in red. (B) Three-dimensional reconstruction of prefrontal white matter (yellow) with semi-transparent gray matter (beige) on a mid-sagittal MR slice viewed from the left side. This illustrates the anterior–posterior extent of gray and white matter ROI. (C) Top-down view of a three-dimensional reconstruction of prefrontal gray matter (beige) on an axial MR slice. (D) Top-down view of a three-dimensional reconstruction of prefrontal white matter (yellow) on an axial MR slice.

Figure 1.

(A) Coronal slice (1.5mm) through the prefrontal region of a normal control subject. Prefrontal gray matter is outlined in red. (B) Three-dimensional reconstruction of prefrontal white matter (yellow) with semi-transparent gray matter (beige) on a mid-sagittal MR slice viewed from the left side. This illustrates the anterior–posterior extent of gray and white matter ROI. (C) Top-down view of a three-dimensional reconstruction of prefrontal gray matter (beige) on an axial MR slice. (D) Top-down view of a three-dimensional reconstruction of prefrontal white matter (yellow) on an axial MR slice.

Figure 2.

Relative volume of the prefrontal gray matter and prefrontal white matter in the first episode schizophrenia patients, first episode affective psychosis patients, and normal control subjects. Horizontal lines represent means. Open triangles indicate unipolar affective psychosis patients, closed triangles represent bipolar (manic) patients.

Figure 2.

Relative volume of the prefrontal gray matter and prefrontal white matter in the first episode schizophrenia patients, first episode affective psychosis patients, and normal control subjects. Horizontal lines represent means. Open triangles indicate unipolar affective psychosis patients, closed triangles represent bipolar (manic) patients.

2
Present address: Department of Neuropsychiatry, Kyorin University School of Medicine, Tokyo, Japan

References

Buchanan RW, Breier A, Kirkpatrick B, Elkashef A, Munson RC, Gellad F, Carpenter WT, Jr (
1993
) Structural abnormalities in deficit and nondeficit schizophrenia.
Am J Psychiat
 
150
:
59
–65.
Buchanan RW, Vladar K, Barta PE, Pearlson GD (
1998
) Structural evaluation of the prefrontal cortex in schizophrenia.
Am J Psychiat
 
155
:
1049
–1055.
Chakos MH, Lieberman JA, Alvir J, Bilder R, Ashtari M (
1995
) Caudate nuclei volumes in schizophrenic patients treated with typical antipsychotics or clozapine.
Lancet
 
345
:
456
–457.
Dickey CC, Shenton ME, Hirayasu Y, Fischer I, Voglmaier MM, Niznikiewicz MA, Seidman LJ, Fraone S, McCarley RW (
2000
) Large CSF volume not attributable to ventricular volume in schizotypal personality disorder.
Am J Psychiat
 
157
:
48
–54.
Dougherty D, Rauch SL (
1997
) Neuroimaging and neurobiological models of depression.
Harvard Rev Psychiat
 
5
:
138
–159.
Endicott J, Spitzer RL, Fleiss JL, Cohen J (
1976
) The Global Assessment of Scale: a procedure for measuring overall severity of psychiatric disturbance.
Arch Gen Psychiat
 
33
:
766
–771.
Gur RE, Cowell P, Turetsky BI, Gallacher F, Cannon T, Bilker W, Gur RC (
1998
) A follow-up magnetic resonance imaging study of schizophrenia. Relationship of neuroanatomical changes to clinical and neurobehavioral measures.
Arch Gen Psychiat
 
55
:
145
–152.
Gur RE, Cowell PE, Latshaw A, Turetsky BI, Grossman RI, Arnold SE, Bilker WB, Gur RC (
2000
) Reduced dorsal and orbital prefrontal gray matter volumes in schizophrenia.
Arch Gen Psychiat
 
57
:
761
–768.
Harvey I, Persaud R, Ron MA, Baker G, Murray RM (
1994
) Volumetric MRI measurements in bipolars compared with schizophrenics and healthy controls.
Psychol Med
 
24
:
689
–699.
Hirayasu Y, Shenton ME, Salisbury DF, Dickey CC, Fischer IA, Mazzoni P, Kisler T, Arakaki H, Kwon JS, Anderson JE, Yurgelun T, Tohen M, McCarley RW (
1998
) Lower left temporal lobe MRI volumes in patients with first-episode schizophrenia compared with psychotic patients with first-episode affective disorder and normal subjects.
Am J Psychiat
 
155
:
1384
–1391.
Hirayasu Y, Shenton ME, Salisbury DF, Kwon JS, Wible CG, Fischer IA, Yurgelun-Todd D, Zarate C, Kikinis R, Jolesz FA, McCarley RW (
1999
) Subgenual cingulate cortex volume in first-episode psychosis.
Am J Psychiat
 
156
:
1091
–1093.
Hollingshead AB (1965) Two factor index of social position. New Haven, CT: Yale Station.
Holzman PS, Kringlen E, Levy DL, Proctor LR, Haberman SJ, Yasillo NJ (
1977
) Abnormal pursuit eye movements in schizophrenia: evidence for a genetic indicator.
Arch Gen Psychiat
 
34
:
802
–805.
Kay SR, Opler LA, Fiszbein A (1986) Positive and negative syndrome scale manual. North Tonawanda, NY: Multi-Health Systems.
Kikinis R, Jolesz FA, Greig G (1990) 3D morphometric and morphometric information derived from clinical brain MR images. In: 3D imaging in medicine (Hohne KH, Fuchs H, Pizer SM, eds), pp. 441–454. Berlin: Springer-Verlag.
Knight RT, Staines WR, Swick D, Chao LL (
1999
) Prefrontal cortex regulates inhibition and excitation in distributed neural networks.
Acta Psychol
 
101
:
159
–178.
McCarley RW, Wible CG, Frumin M, Hirayasu Y, Levitt JJ, Fischer IA, Shenton ME (
1999
) MRI anatomy of schizophrenia.
Biol Psychiat
 
45
:
1099
–1119.
Miller EK (
1999
) The prefrontal cortex: complex neural properties for complex behavior.
Neuron
 
22
:
15
–17.
Nopoulos P, Torres I, Flaum M, Andreasen NC, Ehrhardt JC, Yuh WT (
1995
) Brain morphology in first-episode schizophrenia.
Am J Psychiat
 
152
:
1721
–1723.
Oldfield RC (
1971
) The assessment and analysis of handedness: The Edinburgh Inventory.
Neuropsychologia
 
9
:
97
–113.
Overall JE, Gorham DR (
1962
) The brief psychiatric rating scale.
Psychol Rep
 
10
:
799
–812.
Park S, Holzman PS (
1992
) Schizophrenics show spatial working memory deficits.
Arch Gen Psychiat
 
49
:
975
–982.
Pearlson GD, Marsh L (
1999
) Structural brain imaging in schizophrenia: a selective review.
Biol Psychiat
 
46
:
627
–649.
Pfefferbaum A, Mathalon DH, Sullivan EV, Rawles JM, Zipursky RB, Lim KO (
1994
) A quantitative magnetic resonance imaging study of changes in brain morphology from infancy to late adulthood.
Arch Neurol
 
51
:
874
–887.
Pfefferbaum A, Sullivan EV, Rosenbloom MJ, Mathalon DH, Lim KO (
1998
) A controlled study of cortical gray matter and ventricular changes in alcoholic men over a 5-year interval.
Arch Gen Psychiat
 
55
:
905
–912.
Schlaepfer TE, Harris GJ, Tien AY, Peng LW, Lee S, Federman EB, Chase GA, Barta PE, Pearlson GD (
1994
) Decreased regional cortical gray matter volume in schizophrenia.
Am J Psychiat
 
151
:
842
–848.
Selemon LD, Goldman-Rakic PS (
1999
) The reduced neuropil hypothesis: a circuit based model of schizophrenia.
Biol Psychiat
 
45
:
17
–25.
Selemon LD, Rajkowska G, Goldman-Rakic PS (
1998
) Elevated neuronal density in prefrontal area 46 in brains from schizophrenic patients: application of a three-dimensional stereologic counting method.
J Comp Neurol
 
392
:
402
–412.
Selemon LD, Lidow MS, Goldman-Rakic PS (
1999
) Increased volume and glial density in primate prefrontal cortex associated with chronic antipsychotic drug exposure.
Biol Psychiat
 
46
:
161
–172.
Shenton ME, Frumin M, McCarley RW, Maier S, Westin CF, Fischer IA, Dickey CC, Kikinis R (2001) MR morphometric findings in schizophrenia. In: Psychiatric neuroimaging strategies: research and clinical applications (Dougherty DRS, Rosenbaum J, ed). Washington DC: American Psychiatric Association.
Shenton ME, Wible CG, McCarley RW (1997) A review of magnetic resonance imaging studies of brain abnormalities in schizophrenia. In: Brain imaging in clinical psychiatry (Krishnan KRR, Doraiswamy PM, ed.), pp. 297–380. New York: Marcel Dekker.
Spitzer RL, Williams JBW, Gibbson M, First M (1990a) The Structured Clinical Interview for DSM IIIR (SCID). Washington, DC: American Psychiatric Association.
Spitzer RL, Williams JBW, Gibbson M, First M (1990b) The Structured Clinical Interview for DSM IIIR (SCID-NP) — non-patient edition. Washington, DC: American Psychiatric Association.
Spitzer RL, Williams JBW, Gibbson M, First M (1990c) The Structured Clinical Interview for DSM IIIR (SCID-II) — personality disorders. Washington, DC: American Psychiatric Association.
Sullivan EV, Lim KO, Mathalon D, Marsh L, Beal DM, Harris D, Hoff AL, Faustman WO, Pfefferbaum A (
1998
) A profile of cortical gray matter volume deficits characteristic of schizophrenia.
Cereb Cortex
 
8
:
117
–124.
Szeszko PR, Bilder RM, Lencz T, Pollack S, Alvir JM, Ashtari M, Wu H, Lieberman JA (
1999
) Investigation of frontal lobe subregions in first-episode schizophrenia.
Psychiat Res
 
90
:
1
–15.
Wechsler D (1981) Wechsler Adult Intelligence Scale – Revised, 1981. New York: Harcourt Brace Jovanovich Inc.
Wells W, Kikinis R, Grimson W, Jolesz AF (
1996
) Adaptive segmentation of MRI data.
IEEE Trans Med Imag
 
15
:
429
–442.
Weinberger DR, Berman KF, Zec RF (
1986
) Physiologic dysfunction of prefrontal cortex in schizophrenia: I. Regional cerebral bloodflow evidence.
Arch Gen Psychiat
 
43
:
114
–125.
Wible CG, Shenton ME, Hokama H, Kikinis R, Jolesz FA, Metcalf D, McCarley RW (
1995
) Prefrontal cortex and schizophrenia. A quantitative magnetic resonance imaging study.
Arch Gen Psychiat
 
52
:
279
–288.
Wible CG, Shenton ME, Fischer IA, Allard JE, Kikinis R, Jolesz FA, Iosifescu DV, McCarley RW (
1997
) Parcellation of the human prefrontal cortex using MRI.
Psychiat Res
 
76
:
29
–40.
Zipursky RB, Seeman MV, Bury A, Langevin R, Wortzman G, Katz R (
1997
) Deficits in gray matter volume are present in schizophrenia but not bipolar disorder.
Schizophr Res
 
26
:
85
–92.
Zipursky RB, Lambe EK, Kapur S, Mikulis DJ (
1998
) Cerebral gray matter volume deficits in first episode psychosis.
Arch Gen Psychiat
 
55
:
540
–546.