OUP user menu

Differential contributions of prefrontal and temporolimbic pathology to mechanisms of psychosis

Michio Suzuki, Shi-Yu Zhou, Tsutomu Takahashi, Hirofumi Hagino, Yasuhiro Kawasaki, Lisha Niu, Mie Matsui, Hikaru Seto, Masayoshi Kurachi
DOI: http://dx.doi.org/10.1093/brain/awh554 2109-2122 First published online: 1 June 2005

Summary

Common abnormalities within the schizophrenia spectrum may be essential for the pathogenesis of schizophrenia, but additional pathological changes may be required for the development of full-blown schizophrenia. Clarifying the neurobiological similarities and differences between established schizophrenia and a milder form of schizophrenia spectrum disorder would potentially discriminate the pathophysiological mechanisms underlying the core features of the schizophrenia spectrum from those associated with overt psychosis. High-resolution MRIs were acquired from 25 patients with schizotypal disorder, 53 patients with schizophrenia and 59 healthy volunteers matched for age, gender, handedness and parental education. Volumetric measurements of the medial temporal structures and the prefrontal cortex subcomponents were performed using consecutive 1-mm thick coronal slices. Parcellation of the prefrontal cortex into subcomponents was performed according to the intrinsic anatomical landmarks of the frontal sulci/gyri. Compared with the controls, the bilateral volumes of the amygdala and the hippocampus were reduced comparably in the schizotypal and schizophrenia patients. The parahippocampal gyrus volume did not differ significantly between diagnostic groups. Total prefrontal grey matter volumes were smaller bilaterally in the schizophrenia patients than in the controls and the schizotypal patients, whereas the schizotypal patients had larger prefrontal grey matter than the controls in the right hemisphere. In the schizophrenia patients, grey matter volumes of the bilateral superior frontal gyrus, left middle frontal gyrus, bilateral inferior frontal gyrus and bilateral straight gyrus were smaller than those in the controls. The schizophrenia patients also had reduced grey matter volumes in the right superior frontal gyrus, bilateral middle frontal gyrus and right inferior frontal gyrus relative to the schizotypal patients. Compared with the controls, the schizotypal patients had larger volumes of the bilateral middle frontal gyrus and smaller volumes of the right straight gyrus. There were no significant between-group differences in volumes of the ventral medial prefrontal cortex or the orbitofrontal cortex. These findings suggest that volume reductions in the amygdala and hippocampus are the common morphological substrates for the schizophrenia spectrum, which presumably represent the vulnerability. Additional widespread involvement of the prefrontal cortex in schizophrenia may lead to the loss of inhibitory control in other brain regions and suggests (although it is not specifically be related to) its critical role in the manifestation of overt psychosis.

  • schizotypal disorder
  • schizophrenia
  • MRI
  • medial temporal lobe
  • prefrontal cortex
  • BA = Brodmann area
  • DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, fourth edition
  • ICD-10 = International Classification of Diseases, 10th edition
  • ICV = intracranial volume
  • MANCOVA = multivariate analysis of covariance
  • ROI = region of interest
  • SANS = Scale for the Assessment of Negative Symptoms
  • SAPS = Scale for the Assessment of Positive Symptoms
  • VBM = voxel-based morphometry

Introduction

Pathological deviations genetically and phenomenologically related to schizophrenia are grouped under the schizophrenia spectrum. This concept reflects the assumption that schizophrenia has a multifactorial aetiology in which multiple susceptibility genes interact with environmental insults to yield a range of phenotypes (Siever and Davis, 2004). Common neurobiological abnormalities in the schizophrenia spectrum may be essential for the pathogenesis of schizophrenia. However, some additional pathological changes may also be required for the development of full-blown schizophrenia. Schizotypal (personality) disorder is thought to be a prototypic disorder within the schizophrenia spectrum (Siever et al., 2002). It is genetically related to schizophrenia (Siever et al., 1990; Kendler et al., 1993) and characterized by odd behaviour and attenuated forms of the features seen in schizophrenia without manifestation of overt and sustained psychosis (World Health Organization, 1993; American Psychiatric Association, 1994). Clarifying the neurobiological similarities and differences between established schizophrenia and schizotypal (personality) disorder would potentially discriminate the pathophysiological mechanisms underlying the core features of the schizophrenia spectrum from those associated with overt psychosis. Thus, this strategy may provide a clue to the mechanisms underlying the development of schizophrenic psychosis.

Convergent evidence suggests that the pathological process in schizophrenia predominantly affects the fronto- temporolimbic-paralimbic regions (Shenton et al., 2001; Suzuki et al., 2002). The hippocampal formation and the prefrontal cortex are two of the major structures that have received the most attention in the search for the neural substrate of schizophrenia. Slight but significant volume reductions in the hippocampus, amygdala and frontal lobe have been reported in a number of volumetric MRI studies of schizophrenia (see reviews: Lawrie and Abukmeil, 1998; Harrison, 1999; Shenton et al., 2001). Dysfunction of these regions has been implicated in the cardinal characteristics of schizophrenia. Involvement of the hippocampal formation has been suggested to play a role in manifesting psychotic symptoms and verbal memory deficits in schizophrenia patients (Friston et al., 1992; Liddle et al., 1992; Goldberg et al., 1994), while prefrontal abnormalities have been related to negative symptoms and cognitive impairments, such as deficits in working memory, executive and problem solving functions (Goldman-Rakic and Selemon, 1997).

There is increasing evidence of alterations in the brain structures of schizotypal subjects (see reviews: Dickey et al., 2002a; Siever and Davis, 2004). Our previous study using voxel-based morphometry (VBM) demonstrated that grey matter reduction in the medial temporal region was common to patients with schizophrenia and schizotypal disorder, but schizophrenia patients showed more widespread involvement of the frontal lobe than schizotypal subjects (Kawasaki et al., 2004). These findings need to be confirmed by detailed volumetric region of interest (ROI) analyses. However, only a single volumetric study, by Dickey and colleagues (Dickey et al., 1999), has examined the medial temporal lobe structures in schizotypal subjects and found no abnormality in the amygdala or hippocampus volume. Previous MRI studies have provided evidence of preserved volume of the brain structures densely interconnected with the prefrontal cortex in schizotypal subjects relative to schizophrenia (Byne et al., 2001; Takahashi et al., 2002b, 2004; Suzuki et al., 2004). These findings suggest that the prefrontal cortex may be structurally spared in schizotypal subjects. As to the prefrontal cortex per se, however, only preliminary data referring to preserved frontal lobe volume in schizotypal patients have been reported (Siever and Davis, 2004). Siever and Davis (2004) have made an extensive review of neurobiological findings in subjects with schizotypal personality disorder and proposed a model regarding the pathophysiology of the schizophrenia spectrum disorders. Their model also predicted that temporal volume reductions would be common across the schizophrenia spectrum disorders, whereas frontal volumes would be more preserved in schizotypal subjects than in schizophrenia patients. More data on the volume changes of both the medial temporal lobe and the prefrontal cortex in schizotypal subjects are needed for comparison with those in schizophrenia patients. Detailed volumetric analyses of both structures in the same subjects would allow more compelling conclusions to be drawn. In addition, the great multiplicity of structural and functional organization within the prefrontal cortex necessitates examination of the structural alterations in each subcomponent of the prefrontal cortex. This has been conducted in several studies of schizophrenia patients (Wible et al., 1997; Buchanan et al., 1998, 2004; Goldstein et al., 1999; Crespo-Facorro et al., 2000; Convit et al., 2001; Yamasue et al., 2004) but has never been reported for schizotypal subjects.

The present study aimed to elucidate the implications of structural abnormalities of the medial temporal structures and the prefrontal cortex in the manifestation of psychosis in schizophrenia. We employed high-resolution MRI and performed volumetric assessments of the amygdala, hippocampus, parahippocampal gyrus and prefrontal cortex in patients with schizotypal disorder, comparable patients with established schizophrenia and healthy control subjects. The prefrontal cortex was subdivided into subcomponents according to the intrinsic anatomical landmarks. We hypothesized, from our previous VBM findings (Kurachi, 2003a, b; Kawasaki et al., 2004) and the model by Siever and Davis (2004) that patients with schizotypal disorder would have volume deficits in the medial temporal lobe but limited abnormalities in the prefrontal cortex, whereas patients with schizophrenia would show volume reductions in the medial temporal lobe as well as in widespread regions of the prefrontal cortex.

Methods

Subjects

Twenty-five patients (15 males, 10 females) with schizotypal disorder, 53 patients with schizophrenia (32 males, 21 females) and 59 control subjects (35 males, 24 females) were included in this study. All subjects were right-handed. Demographic and clinical data of the subjects are presented in Table 1.

View this table:
Table 1

Demographic and clinical characteristics of patients with schizotypal disorder, patients with schizophrenia and healthy comparison subjects

Schizotypal disorder patients (n = 25)Schizophrenia patients (n = 53)Healthy comparison subjects (n = 59)
Male/female15/1032/2135/24
Handedness25 right53 right59 right
Age (years)25.5 ± 5.725.3 ± 5.024.3 ± 5.3
Height (cm)164.6 ± 8.7166.1 ± 7.3167.0 ± 7.3
Weight (kg)60.3 ± 9.761.7 ± 12.758.1 ± 9.4
Education (years)13.5 ± 1.813.2 ± 1.916.0 ± 2.5
Parental education (years)12.1 ± 1.912.2 ± 2.112.8 ± 2.4
Age at onset (years)21.7 ± 4.5
Duration of illness (years)3.7 ± 3.8
Total SAPS score16.0 ± 8.524.1 ± 20.5
Total SANS score46.8 ± 24.545.7 ± 22.5
Drug dose (mg/day, haloperidol equivalent)*3.9 ± 4.711.6 ± 9.4
Duration of medication (years)0.3 ± 0.42.7 ± 3.1
  • Values represent mean ± SD.

  • * Neuroleptic dosages of different classes of antipsychotic drugs were converted into haloperidol equivalents using the guideline by Toru (2001). Post hoc comparisons following analysis of variance (ANOVA) revealed:

  • P < 0.01, smaller than in controls;

  • P < 0.01, larger than in schizotypal disorder patients. SAPS = Scale for the Assessment of Positive Symptoms; SANS = Scale for the Assessment of Negative Symptoms.

The patients with schizotypal disorder were recruited from among the subjects who visited the clinics of the Department of Neuropsychiatry, Toyama Medical and Pharmaceutical University Hospital manifesting schizotypal features with distress or associated problems in their lives. Structured clinical interviews were performed using the Comprehensive Assessment of Symptoms and History (CASH) (Andreasen et al., 1992) and Structured Clinical Interview for DSM-IV axis II disorders (SCID-II) (First et al., 1997). They all met the criteria for schizotypal disorder in the International Classification of Diseases, 10th edition (ICD-10) (World Health Organization, 1993) as well as the criteria for schizotypal personality disorder in the Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV) (American Psychiatric Association, 1994). Based on the data from the CASH and SCID-II, subjects were diagnosed by a consensus of at least two experienced psychiatrists, and when necessary the propriety of including cases in the study was discussed among clinical staff members involved. None of the subjects was judged to meet the criteria for schizophrenia of ICD-10 or of DSM-IV currently or previously. At the time of MRI scanning, six patients were neuroleptic-naive and 19 patients were being treated with low doses of antipsychotics; six patients were being treated with typical neuroleptics and 13 patients were receiving atypical neuroleptics. All subjects have received consistent clinical follow-up and none of them has developed overt schizophrenia to date (mean follow-up period after MRI scanning = 2.5 years, SD = 1.9). Four of the 25 patients with schizotypal disorder were relatives of individuals with schizophrenia. Since schizotypal subjects rarely present themselves for clinical care, our clinic-based sample was considered to be somewhat more severely ill than may be expected of schizotypal individuals among the general population.

The patients with schizophrenia were diagnosed based on the CASH and Structured Clinical Interview for DSM-IV axis I disorders (SCID-I) (First et al., 1996). They fulfilled both ICD-10 and DSM-IV criteria for schizophrenia. All schizophrenia patients apart from one female patient were receiving neuroleptic medication; 25 patients were being treated with typical neuroleptics and 27 patients were receiving atypical neuroleptics. The clinical status of the schizophrenia patients was variable; some of them were in an active psychotic episode and others were in partial remission or in a residual phase. All patients with schizotypal disorder and schizophrenia were physically healthy and none had a history of head trauma, neurological illness, serious medical or surgical illness, or substance abuse disorder. Clinical symptoms were rated by well-trained psychiatrists or psychologist within 1 month of scanning using the Scale for the Assessment of Negative Symptoms (SANS; Andreasen, 1983) and the Scale for the Assessment of Positive Symptoms (SAPS; Andreasen, 1984). Inter-rater intraclass correlation coefficients were over 0.92 for all the subscale scores and the total scores of the SANS and the SAPS.

The control subjects consisted of healthy volunteers recruited from among the community and hospital staff or were medical and pharmaceutical students. They were interviewed by psychiatrists using the questionnaire concerning their family and past histories, and present illness. Subjects were excluded if they had a history of psychiatric illness, head trauma, neurological illness, serious medical or surgical illness, or substance abuse disorder. They were also screened for a history of psychiatric disorders in their first-degree relatives. All control subjects were given the Minnesota Multiphasic Personality Inventory, and subjects were excluded if they had abnormal profiles with any T-score exceeding 70. The three groups were matched for age, sex, handedness, height and parental education (Table 1).

After complete description of the study to the subjects, written informed consent was obtained. This study was approved by the Committee on Medical Ethics of Toyama Medical and Pharmaceutical University.

There are considerable overlaps between the subjects in the present study and those in previous MRI studies from our group. Of the 25 patients with schizotypal disorder, 15 and 17 patients overlapped with those in the volumetric MRI studies of the anterior cingulate gyrus (Takahashi et al., 2002b, 2004) and of the internal capsule (Suzuki et al., 2004), respectively. Of the 53 patients with schizophrenia, 34 overlapped with those in the volumetric MRI studies (Takahashi et al., 2002a; Zhou et al., 2003; Niu et al., 2004). In the VBM study by Kawasaki et al. (2004), 17 schizotypal patients and 20 schizophrenia patients were the same as those in the present study. In these previous studies, 37–54 of the control subjects also overlapped with those in the present study according to the stages of our research.

MRI acquisition and processing

MRI scans were acquired with a 1.5 T scanner (Vision; Siemens Medical System, Erlangen, Germany). A three-dimensional T1-weighted gradient-echo sequence FLASH (fast low-angle shots) with 1 × 1 × 1 mm voxels was used. Imaging parameters were: TE (echo time) = 5 ms; TR (repetition time) = 24 ms; flip angle = 40°; field of view = 256 mm; matrix size = 256 × 256.

Image processing for volumetric ROI analysis has been described in detail previously (Takahashi et al., 2002a). Briefly, on a Unix workstation (Silicon Graphics, Mountain View, CA, USA), the image data were processed with the software package Dr View 5.0 (Asahi Kasei Joho System, Tokyo, Japan). Brain images were realigned in three dimensions and reconstructed into entire contiguous coronal slices of 1 mm thickness perpendicular to the anterior commissure–posterior commissure line. The whole cerebrum was separated from the brainstem and cerebellum. The signal intensity histogram distributions across the whole cerebrum were used to segment the voxels semiautomatically into grey matter, white matter and cerebrospinal fluid (CSF). Using the thresholds between the tissue compartments, volumes of whole hemispheric grey matter and white matter were calculated. These whole hemispheric grey matter and white matter volumes summed to the whole cerebral hemisphere volume, which did not include CSF or ventricles. Intracranial volume (ICV) was measured by manual tracing of the intracranial cavity on reformatted 5 mm thick sagittal slices as described previously (Zhou et al., 2003).

Volumetric analysis of ROIs

The ROIs for volumetric measurements were placed on the medial temporal structures and prefrontal cortex, as presented in Figs 1 and 2, respectively.

Fig. 1

Delineations of medial temporal regions of interest taken from mutually orthogonal transaxial (A), sagittal (B) and coronal (C) planes. A three-dimensional reconstructed image of the three regions is also shown (D). Each of the regions is differentially coloured: amygdala (green), hippocampus (red) and parahippocampal gyrus (blue). a, anterior; i, inferior; l, left; p, posterior; r, right; s, superior.

Fig. 2

Three-dimensional reconstructed images of prefrontal regions of interest presenting right lateral (A), right medial (B), dorsal (C), ventral (D) and anterior (E) views of the brain. Panel F demonstrates subdivisions of superior frontal gyrus. 1, superior frontal gyrus; 2, middle frontal gyrus; 3, inferior frontal gyrus; 4, ventral medial prefrontal cortex; 5, orbitofrontal cortex; 6, straight gyrus; 7, dorsolateral part of superior frontal gyrus; 8, dorsal medial prefrontal cortex; 9, supplementary motor cortex.

Medial temporal lobe

The amygdala, hippocampus and parahippocampal gyrus were manually outlined on consecutive coronal 1 mm slices with the corresponding sagittal and axial planes simultaneously presented for reference. Volumes of grey and white matter in each of these structures were measured together. The detailed procedures for delineation of these structures were described previously (Niu et al., 2004; Suzuki et al., 2005a). The inferior border of the amygdala in contact with the hippocampal head was determined by reference to the sagittal plane since the boundary between the hippocampus and the amygdala is more readily identified on the sagittal plane (Convit et al., 1999). Anatomical boundaries for these structures are presented in Table 2.

View this table:
Table 2

Anatomical landmarks demarcating the regions of interest

RegionAnatomical landmark
Medial temporal region
    Amygdala
        Anterior borderAppearance of oval-shaped grey matter of the amygdala
        Posterior borderThin strip of grey matter of the hippocampal–amygdala transitional area
        Superior borderCerebrospinal fluid overlying the semilunar gyrus and its medial extension
        Inferior borderAlveus
        Lateral borderTemporal lobe white matter and extension of the temporal horn
        Medial borderThin strip of parahippocampal white matter (angular bundle)
    Hippocampus
        Anterior borderAlveus
        Posterior borderLevel of the last appearance of fibres of the fornix
        Superior borderAlveus
        Inferior borderWhite matter of parahippocampal gyrus
        Lateral borderInferior horn of lateral ventricle
        Medial borderMesial edge of temporal lobe
    Parahippocampal gyrus
        Anterior borderLevel of the first appearance of the temporal stem
        Posterior borderLevel of the last appearance of fibres of the fornix
        Superior borderInferior grey border of the hippocampal formation
        Lateral borderA line drawn from the most lateral border of the hippocampal flexure to the collateral sulcus
Prefrontal area
    Superior frontal gyrus (includes the paracingulate gyrus when it exists)
        Lateral inferior borderSuperior frontal sulcus
        Medial inferior borderCingulate sulcus and, in the most anterior part, superior rostral sulcus
        Anterior borderFrontomarginal sulcus, which extends from superior frontal sulcus
        Posterior borderPrecentral sulcus on the lateral surface and paracentral sulcus on the medial surface
    Dorsolateral partMedially separated by the superior margin of the hemisphere
    Medial partDorsolaterally separated by the superior margin of the hemisphere
        Dorsal medial prefrontal cortexPosteriorly demarcated by the coronal plane through the most anterior tip of the inner surface of the genu of the corpus callosum
        Supplementary motor areaAnteriorly demarcated by the same coronal plane as above
    Middle frontal gyrus
        Superior borderSuperior frontal sulcus
        Inferior borderInferior frontal sulcus
        Anterior borderFrontomarginal sulcus, which extends from superior frontal sulcus
        Posterior borderPrecentral sulcus
    Inferior frontal gyrus
        Superior borderInferior frontal sulcus
        Inferior borderFrontomarginal sulcus or lateral orbital sulcus in the anterior part and superior circular sulcus in the operculum
        Anterior borderFrontomarginal sulcus, which extends from inferior frontal sulcus
        Posterior borderPrecentral sulcus
    Ventral medial prefrontal cortex
        Superior borderSuperior rostral sulcus in the anterior part and cingulate sulcus in the posterior part
        Inferior borderInferior rostral sulcus (the lowest visible sulcus in the medial surface of the hemisphere)
        Anterior borderFrontomarginal sulcus, which extends from superior rostral sulcus
        Posterior borderMore posterior coronal plane through either of posterior extreme of cingulate sulcus or superior rostral sulcus
    Orbitofrontal cortex
        Anterior/lateral borderFrontomarginal sulcus in the anterior part, lateral orbital sulcus in the intermediate part and inferior circular sulcus in the posterior part
        Medial borderSuperior rostral sulcus, which anteriorly merges into frontomarginal sulcus in the rostral part and olfactory sulcus on the ventral surface of the hemisphere
        Posterior borderThe most posterior coronal plane containing medial orbital gyrus
    Straight gyrus
        Lateral borderOlfactory sulcus
        Medial borderInferior rostral sulcus (the lowest visible sulcus in the medial surface of the hemisphere)
        Anterior borderAnterior extreme of olfactory sulcus
        Posterior borderOlfactory trigone

Prefrontal cortex

Delineation of the prefrontal cortex was partially based on the works of Rademacher et al. (1992) and Crespo-Facorro et al. (1999a). Parcellation of the frontal lobe into subcomponents was performed according to the anatomical landmarks that were, in principle, intrinsic to the brain (sulci/gyri). With the availability of synchronous–orthogonal views in three dimensions in conjunction with the context of gyri/sulci on successive slices, decisions about the landmarks could be made readily. First, the whole frontal lobe was separated from the rest of the brain by the central sulcus. The prefrontal area was demarcated by subtracting the precentral gyrus and the cingulate gyrus from the whole frontal lobe. By this definition of the prefrontal area, it inevitably includes the premotor cortex [Brodmann area (BA) 6 and part of BA 8]. The paracingulate gyrus (approximately corresponding to BA 32), when present, was included in the prefrontal area. After the extraction of the prefrontal area, it was subdivided into six subregions: the superior frontal gyrus, which was further subdivided into three parts (dorsolateral part, dorsal medial prefrontal cortex and supplementary motor cortex); middle frontal gyrus; inferior frontal gyrus; ventral medial prefrontal cortex; orbitofrontal cortex; and straight gyrus. Anatomical boundaries for each region are described in Table 2. All the volumetric measurements were performed on reformatted consecutive 1 mm coronal slices by manual outlining. Grey matter volumes of the regional cortices were calculated by applying the segmentation procedure described previously.

Three trained raters (S.Z., H.H. and L.N.), who were blinded to the subjects' identities, measured the volumes of the prefrontal regions, the amygdala, and the hippocampus and parahippocampal gyrus, respectively. Inter- and intra-rater intraclass correlation coefficients in five randomly selected brains were over 0.92 for the prefrontal ROIs and over 0.93 for the medial temporal ROIs.

Statistical analysis

Statistical differences in the regional volume measures were analysed by repeated measures multivariate analysis of covariance (MANCOVA) with ICV and age as covariates for each region, with diagnosis group (schizophrenia patients, schizotypal disorder patients, control subjects) and gender (male, female) as between-subject factors and hemisphere (right, left) as a within-subject factor. For the comparison of ICV, only age was treated as a covariate. Post hoc Tukey's tests were employed to follow up the significant main effects or interactions yielded by MANCOVAs. Pearson's partial correlation coefficients, controlling for ICV and age, were calculated to examine relationships between the ROI volumes and the clinical variables. Statistical significance was defined as P < 0.05 (two-tailed). To prevent a possible type I error due to multiple tests, a Bonferroni correction was applied for correlation analyses.

Results

Volumes of measured ROIs and results of MANCOVAs for the main effect of diagnosis are presented in Tables 3, 4 and 5. We report below the results concerning main effects of diagnosis or interactions involving diagnosis only when they were significant or had a nearly significant trend, and subsequent post hoc analyses.

View this table:
Table 3

Volumes of intracranial cavity, cerebral hemisphere and cerebral grey and white matter in patients with schizotypal disorder, patients with schizophrenia and healthy comparison subjects

Regions of interestSchizotypal disorder patientsSchizophrenia patientsHealthy comparison subjectsDiagnosis effect
FdfP
Intracranial volume1526 ± 1501496 ± 1551509 ± 1280.712,1300.492
Whole cerebral hemisphere1.842,1290.162
    Left§559.0 ± 56.6538.6 ± 59.5553.8 ± 48.8
    Right566.1 ± 57.2545.8 ± 60.4561.3 ± 49.2
Whole cerebral grey matter4.892,1290.008
    Left#362.1 ± 40.5339.8 ± 39.3†,356.4 ± 36.2
    Right353.6 ± 40.3329.9 ± 37.5†,347.6 ± 35.6
Whole cerebral white matter1.692,1290.187
    Left§197.0 ± 26.3198.8 ± 33.1197.3 ± 31.5
    Right212.5 ± 29.3215.9 ± 39.4213.7 ± 36.0
  • Values represent mean ± SD of measured volume (cm3). Post hoc comparisons following multivariate analysis of variance with age and intracranial volume as covariates (MANCOVA) revealed:

  • P < 0.01, smaller than in controls;

  • P < 0.01, smaller than in schizotypal disorder patients;

  • § P < 0.01, smaller than on right hemisphere;

  • # P < 0.01, larger than on right hemisphere.

View this table:
Table 4

Volumes of medial temporal lobe structures in patients with schizotypal disorder, patients with schizophrenia and healthy comparison subjects

Regions of interestSchizotypal disorder patientsSchizophrenia patientsHealthy comparison subjectsDiagnosis effect
FdfP
Amygdala19.082,129<0.001
    Left§0.96 ± 0.130.99 ± 0.151.13 ± 0.14
    Right0.97 ± 0.151.05 ± 0.171.15 ± 0.14
Hippocampus3.242,1290.042
    Left§2.83 ± 0.372.89 ± 0.423.04 ± 0.40
    Right3.03 ± 0.393.09 ± 0.563.24 ± 0.35
Parahippocampal gyrus0.342,1290.706
    Left7.22 ± 0.737.01 ± 1.137.15 ± 0.90
    Right7.22 ± 0.577.09 ± 1.087.31 ± 0.76
  • Values represent mean ± SD of measured volume (cm3). Post hoc comparisons following multivariate analysis of variance with age and intracranial volume as covariates (MANCOVA) revealed:

  • P < 0.01,

  • P < 0.05, smaller than in controls;

  • § P < 0.01, smaller than on right hemisphere.

View this table:
Table 5

Volumes of whole prefrontal grey and white matter and prefrontal cortex subcomponents in patients with schizotypal disorder, patients with schizophrenia and healthy comparison subjects

Regions of interestSchizotypal disorder patientsSchizophrenia patientsHealthy comparison subjectsDiagnosis effect
Fdfp
Prefrontal grey matter3.512,1290.032
    Left‡‡97.18 ± 13.6591.05 ± 12.3296.38 ± 9.88
    Right94.99 ± 13.33††88.16 ± 11.8292.38 ± 9.54
Prefrontal white matter1.342,1290.264
    Left§§46.72 ± 7.0744.54 ± 8.6947.13 ± 7.66
    Right49.73 ± 7.2649.14 ± 10.0550.28 ± 9.05
Superior frontal gyrus3.492,1290.033
    Left‡‡28.91 ± 5.3127.49 ± 4.2129.64 ± 4.05
    Right27.65 ± 5.2925.63 ± 4.37,#27.76 ± 3.95
Dorsolateral part0.532,1290.589
    Left‡‡12.42 ± 3.1912.34 ± 2.3313.02 ± 2.81
    Right12.27 ± 3.4611.76 ± 2.8312.34 ± 2.33
Dorsal medial prefrontal cortex4.842,1290.009
    Left‡‡10.19 ± 2.359.16 ± 1.99,#10.16 ± 1.81
    Right9.39 ± 1.928.72 ± 1.779.73 ± 1.79
Supplementary motor cortex3.482,1290.033
    Left‡‡6.30 ± 1.096.00 ± 1.006.47 ± 1.19
    Right5.99 ± 1.565.14 ± 1.07,#5.69 ± 1.23
Middle frontal gyrus2.902,1290.058
    Left‡‡29.34 ± 5.67††25.87 ± 4.9527.31 ± 4.62
    Right28.44 ± 5.66††25.50 ± 4.41§26.53 ± 4.78
Inferior frontal gyrus4.922,1290.008
    Left‡‡13.02 ± 2.6512.58 ± 2.1413.87 ± 2.10
    Right13.18 ± 2.2212.05 ± 2.1912.89 ± 1.83
Ventral medial prefrontal cortex0.922,1290.397
    Left5.61 ± 1.225.51 ± 1.215.84 ± 1.12
    Right5.56 ± 1.035.48 ± 1.175.69 ± 1.23
Orbitofrontal cortex0.472,1290.622
    Left15.69 ± 1.9115.11 ± 2.0315.44 ± 1.57
    Right15.58 ± 1.8715.07 ± 2.0715.47 ± 1.45
Straight gyrus15.452,129<0.001
    Left3.06 ± 0.492.89 ± 0.453.31 ± 0.49
    Right3.02 ± 0.512.92 ± 0.433.31 ± 0.50
  • Values represent mean ± SD of measured volume (cm3). Post hoc comparisons following multivariate analysis of variance with age and intracranial volume as covariates (MANCOVA) revealed:

  • P < 0.01;

  • P < 0.05, smaller than in controls;

  • § P < 0.01,

  • # P < 0.05, smaller than in schizotypal disorder patients;

  • †† P < 0.05, larger than in controls;

  • ‡‡ P < 0.01, larger than on right hemisphere;

  • §§ P < 0.01, smaller than on right hemisphere.

Volumes of global brain structures

There were no significant differences in the volumes of intracranial cavity, whole cerebral hemisphere or whole cerebral white matter among diagnostic groups (Table 3). MANCOVAs revealed a significant main effect of diagnosis for the whole cerebral grey matter (Table 3). The volumes of whole cerebral grey matter were significantly smaller in the schizophrenia patients compared with the controls (post hoc tests, P < 0.001 for both hemispheres) and the schizotypal patients (P < 0.001 for both hemispheres).

Volumes of medial temporal lobe

A significant main effect of diagnosis in MANCOVA was revealed in the amygdala and the hippocampus (Table 4). Post hoc analyses demonstrated that, compared with the controls, the volume of the amygdala was significantly smaller in the patients with schizotypal disorder (P < 0.001 for both hemispheres) and schizophrenia (P < 0.001 for both hemispheres). The volume of the hippocampus was also significantly smaller in the patients with schizotypal disorder (P = 0.039 for the left and P = 0.020 for the right) and schizophrenia (P = 0.009 for the left and P = 0.005 for the right) than in the controls. There was no significant difference in the amygdala or hippocampus volume between schizotypal disorder and schizophrenia. The parahippocampal gyrus measures did not differ among diagnostic groups (Table 4).

Volumes of prefrontal cortex

A significant main effect of diagnosis in MANCOVA was observed in the total prefrontal grey matter but not in the total prefrontal white matter (Table 5). Post hoc analyses demonstrated that the prefrontal grey matter volume was significantly smaller in the schizophrenia patients compared with the controls (P < 0.001 for both hemispheres) and the schizotypal patients (P < 0.001 for both hemispheres). In contrast, the schizotypal disorder patients had larger prefrontal grey matter than the controls in the right hemisphere (P = 0.040).

Among the prefrontal cortex subcomponents, MANCOVA revealed a significant main effect of diagnosis in the superior frontal gyrus, inferior frontal gyrus and straight gyrus, and an insignificant trend for main effect of diagnosis in the middle frontal gyrus (Table 5). When the superior frontal gyrus was further subdivided, a significant main effect of diagnosis was found only in the medial parts, such as the dorsal medial prefrontal cortex and supplementary motor cortex (Table 5). A significant interaction between diagnosis and gender was observed only in the inferior frontal gyrus [F(2,129) = 3.97, P = 0.021].

Post hoc analyses demonstrated that the superior frontal gyrus grey matter volume was significantly reduced in the schizophrenia patients compared with the controls (P < 0.001 for both hemispheres) and the schizotypal patients (P = 0.014 for the right) (Fig. 3A). In the superior frontal gyrus subdivisions, the schizophrenia patients had a significantly smaller dorsal medial prefrontal cortex volume than the controls (P < 0.001 for both hemispheres) and the schizotypal patients (P = 0.016 for the left). The supplementary motor cortex volume in the schizophrenia patients was also significantly smaller than in the controls (P = 0.022 for the left and P = 0.026 for the right) and the schizotypal patients (P = 0.013 for the right).

Fig. 3

Scatter plots of absolute volumes of grey matter for each prefrontal subcomponent in patients with schizotypal disorder, patients with schizophrenia and healthy comparison subjects: superior frontal gyrus (A), middle frontal gyrus (B), inferior frontal gyrus (C) and straight gyrus (D). Horizontal bars indicate means of each group. *P < 0.05; **P < 0.01: post hoc comparisons followed multivariate analysis of variance with age and intracranial volume as covariates.

The middle frontal gyrus volume was significantly smaller in the schizophrenia patients compared with the controls (P = 0.002 for the left) and the schizotypal patients (P < 0.001 for both hemisphere) (Fig. 3B). Further, the schizotypal patients had significantly larger middle frontal gyrus volume than the controls (P = 0.026 for both hemispheres) (Fig. 3B).

The inferior frontal gyrus volume was significantly reduced in the schizophrenia patients compared with the controls (P < 0.001 for both hemispheres) and the schizotypal patients (P < 0.001 for the right) (Fig. 3C). As a significant diagnosis × gender interaction was also found, we made post hoc comparisons separately in each gender. In the male subjects, significant volume reductions of the left inferior frontal gyrus were found in the patients with schizotypal disorder (P = 0.001) and schizophrenia (P = 0.020) compared with the controls. The female patients with schizophrenia had a significantly smaller volume than the patients with schizotypal disorder (P < 0.001 for both hemispheres) and the controls (P < 0.001 for the left and P = 0.001 for the right).

Compared with the controls, the straight gyrus volume was significantly smaller in the patients with schizotypal disorder (P = 0.037 for the right) and schizophrenia (P < 0.001 for both hemispheres) (Fig. 3D).

Correlations between volume measures and clinical variables

Partial correlation analyses controlling for ICV and age did not reveal any significant correlation between the volume measures of each ROI and daily dosage of neuroleptic medication or duration of medication in either the schizotypal disorder or the schizophrenia group. In addition, the volume measures were not significantly correlated with age at onset of illness or duration of illness in the schizophrenia patients.

To test the possibility that increases in the prefrontal cortex volumes in the schizotypal group reflect the compensatory mechanism secondary to the medial temporal lobe abnormalities, partial correlation coefficients were calculated between the volume of the right prefrontal grey matter or the bilateral middle frontal gyri and the volume of the amygdala or the hippocampus. The right hippocampal volume was significantly correlated with the right prefrontal grey matter volume (r = −0.620, P = 0.002) and the left middle frontal gyrus volume (r = −0.607, P = 0.002) even after Bonferroni correction (P < 0.004).

Discussion

There are two main points in this study: (i) volumes of the amygdala and the hippocampus were commonly reduced in patients with schizophrenia and schizotypal disorder; (ii) volumes of the subcomponents of the prefrontal cortex were widely reduced in schizophrenia patients, whereas those in schizotypal subjects were mostly preserved.

Temporolimbic pathology as vulnerability

Consistent with the previous VBM study (Kawasaki et al., 2004), the present results suggest that the volume reduction of the amygdala and hippocampus is a common morphological basis for the schizophrenia spectrum. Studies of family members of patients with schizophrenia have also revealed evidence of medial temporal abnormalities similar to those found in schizophrenia patients (Lawrie et al., 1999; Seidman et al., 1999, 2002; Van Erp et al., 2002). Schizotypal disorder has dual aspects that are contradictory in relation to the liability to schizophrenia. Schizotypal subjects are generally spared overt psychosis in spite of the presence of incipient psychotic symptoms. On the other hand, they have a higher incidence of developing schizophrenia than the general population (Fenton and McGlashan, 1989). Thus they are assumed to have vulnerability to schizophrenia but are simultaneously protected from developing full-blown psychosis. Our findings support the notion that reduced temporolimbic volume represents a vulnerability marker, which is necessary but not sufficient for developing schizophrenia (Seidman et al., 2002; Kurachi, 2003a, b).

Prefrontal involvement in schizophrenia

There seems general agreement that total prefrontal grey matter is reduced in patients with schizophrenia compared with healthy subjects (Shenton et al., 2001; Selemon et al., 2002). However, findings in previous studies that have parcellated the prefrontal cortex into subcomponents have varied in spatial distribution of the gross anatomical changes within the prefrontal cortex in schizophrenia (Buchanan et al., 1998, 2004; Baaré et al., 1999; Goldstein et al., 1999; Crespo-Facorro et al., 2000; Gur et al., 2000; Sanfilipo et al., 2000; Convit et al., 2001; Yamasue et al., 2004). These inconsistencies may be due, in large part, to the use of different image measurement procedures. In particular, there has been substantial variability among studies in definitions of boundaries subdividing the prefrontal cortex into subcomponents.

The present study demarcated the prefrontal ROIs by fully taking account of the anatomical landmarks intrinsic to the frontal lobe, and revealed widespread alterations in volume of the prefrontal cortex in schizophrenia. This is consistent with the observation that schizophrenia patients have deficits in extensive neurobehavioural domains involving the prefrontal cortex, such as cognition including executive functions, motivation and emotion (Goldman-Rakic and Selemon, 1997).

The present study also suggested a considerable preference for anatomical involvement of the prefrontal cortex in schizophrenia. When the superior frontal gyrus was subdivided into dorsolateral and medial parts, significant volume reduction was observed only in the medial part. This finding should be interpreted with caution because the corpus callosum, which has been reported to be abnormal in schizophrenia (Shenton et al., 2001), was used as a landmark to define the subregions of the medial part of the superior frontal gyrus. Thus, the volume differences found may reflect differences in shape or volume of the corpus callosum between groups. However, decreased blood flow in the medial prefrontal region has been shown in schizophrenia patients during the performance of memory tasks (Andreasen et al., 1996; Crespo-Facorro et al., 1999b). Moreover, functional neuroimaging studies have demonstrated that the medial prefrontal cortex, including the paracingulate cortex, is activated by tasks involving autonomic arousal, many forms of self-monitoring and social cognition (Gallagher and Frith, 2003; Ridderinkhof et al., 2004). The possible relevance of medial prefrontal dysfunction to the pathophysiology of schizophrenia seems worthy of examination in future studies.

Prefrontal involvement in schizotypal disorder

To our knowledge, this study is the first to report comprehensive volumetric results of the prefrontal cortex subcomponents in schizotypal subjects. In all parts except the right straight gyrus, prefrontal cortical volumes in the schizotypal patients were not reduced, while the bilateral middle frontal gyrus and right prefrontal grey matter as a whole were even larger than those of the control subjects. These findings support the model proposed by Siever and Davis (2004) and provide more compelling morphological evidence for their predictions. Preserved volume of the prefrontal cortical regions is consistent with the findings that performance in tasks involving the frontal lobe functions is better in schizotypal individuals than that in patients with schizophrenia (Mitropoulou et al., 2002; Matsui et al., 2004). Increases in the prefrontal grey matter volume might reflect functional compensation, which a few functional brain imaging studies have suggested to occur in the prefrontal cortex of schizotypal subjects (Buchsbaum et al., 1997, 2002). The possible compensatory mechanism will be discussed further below.

Prefrontal pathology and manifestation of psychosis

Differential involvement of the prefrontal cortex between patients with schizophrenia and schizotypal disorder in the present study strongly suggests that prefrontal pathology is critical for overt manifestation of psychosis in schizophrenia spectrum patients. Previous literature on MRI, however, has highlighted several non-frontal regions involving the positive psychotic symptoms in schizophrenia. In particular, volume loss in the superior temporal gyrus has been related to a variety of psychotic symptoms (Barta et al., 1990; Shenton et al., 1992; Menon et al., 1995; Kim et al., 2003). Other studies revealed an inverse relationship between the amygdala–hippocampal complex and overall positive symptoms (Bogerts et al., 1993) or between the paralimbic cortices and Schneiderian symptoms (Suzuki et al., 2005b). It may be notable that volume reduction of the superior temporal gyrus was reported commonly in schizotypal subjects (Dickey et al., 1999, 2002b; Kawasaki et al., 2004) and patients with schizophrenia (Shenton et al., 2001).

The prefrontal cortex has a high density of interconnections with almost all other sectors of the cerebral cortex, including the limbic areas. One of the integrative functions of the prefrontal cortex, through these widespread connections, is thought to be the inhibitory control of interference (Fuster, 1997; Mesulam, 2000). It probably protects the structure of behaviour or thought from external or internal interfering influences. From our results, as has been suggested in previous literature (Frith et al., 2000; Kurachi, 2003b), it might be possible to imply that deficits in the inhibitory function of the prefrontal cortex result in emergence of prominent psychotic symptoms, which might have a source in the dysfunctional medial and/or lateral temporal regions.

A few functional brain imaging studies have provided supporting evidence for this notion that prefrontal cortex dysfunction is correlated with exaggerated subcortical dopaminergic transmission in schizophrenia (Bertolino et al., 2000; Meyer-Lindenberg et al., 2002). Animal studies have also shown that neonatal excitotoxic lesions of the medial temporal lobe lead to developmental abnormality of the prefrontal cortex (Bertolino et al., 1997, 2002) in association with postpubertal emergence of excessive subcortical dopamine transmission (Lipska et al., 1993; Saunders et al., 1998; Uehara et al., 2000). Any significant correlation between the prefrontal cortical volumes and positive psychotic symptoms in schizophrenia, if present, would support a possible critical role of the prefrontal cortex in the manifestation of overt psychosis. However, we could not examine the symptom–morphology relationships because our schizophrenia sample, with varying clinical status, was not suitable for such analysis. This should be noted as a limitation of the present study.

Taken the results together, however, it is tempting to speculate that some genetic or environmental factor, which enables the prefrontal cortex to compensate for the medial temporal lobe abnormality, e.g. increases in synapses secondary to reduced inputs from the medial temporal lobe, may contribute to avoiding prominent and persistent psychotic symptoms in schizotypal disorder. The significant negative correlations found between the prefrontal grey matter volume and the hippocampal volume in the schizotypal group may lend support to this view. It cannot be stated, however, that the prefrontal cortex is specifically involved in the compensatory mechanism, because the implications of the present study are limited by the lack of volume measures of other neocortical regions, such as the temporal neocortex and the parietal cortex, where morphological changes have also been reported in schizophrenia (Shenton et al., 2001; Buchanan et al., 2004).

Possible confounding factors

A few possible confounding factors in the present study must be taken into account. First, significant differences in the medication status between the schizophrenia and schizotypal groups might have affected the volumetric results. However, the dosage or duration of neuroleptic medication was not correlated with any of the volume measures of the medial temporal and prefrontal structures. Furthermore, sustained neuroleptic treatment could not easily explain the fact that the medial temporal volumes were comparably reduced both in the patients with schizophrenia and in those with schizotypal disorder. Secondly, in the present study young patients with schizotypal disorder were included for comparison with the schizophrenia patients with relatively short durations of illness and medication. This has made it difficult to eliminate the possibility of including schizotypal subjects who would develop overt schizophrenia later on. All the patients included have continued to receive prospective clinical follow-up.

Conclusions

Detailed volumetric comparisons of the medial temporal structures and the prefrontal cortex subcomponents revealed differential morphological alterations in these structures between the patients with schizotypal disorder and those with schizophrenia. Volume reductions in the amygdala and the hippocampus common to both patient groups may represent the vulnerability to schizophrenia, while prefrontal volume loss preferentially observed in schizophrenia may be a critical factor for overt manifestation of psychosis. Although the specificity of this relationship should further be clarified, possible differential contributions of prefrontal and temporolimbic pathologies to the mechanisms of psychosis provide a framework for further studies investigating the pathogenesis of schizophrenia.

Acknowledgments

This research was supported in part by Grants-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (16500215) and the Japanese Ministry of Education, Culture, Sports, Science and Technology (12210009).

References

View Abstract