proton magnetic resonance spectroscopy ( 1h-mrs) of the thalamus in schizophrenia

Proton Magnetic Resonance SpectroscopyH MRS) in Schizophrenia: Investigation of

the Right and Left Hippocampus, Thalamus,and Prefrontal Cortex

by Pascal Delamillieure, Jean-Marc Constans, Jesus Fernandez,Perine Brazo, Karim Benali, Patrick Courtheoux, Florence Thibaut,

Michel Petit, and Sonia Dollfus

Abstract

Single voxel proton magnetic resonance spectroscopy('H MRS) was used to study the metabolites N-acetyl-aspartate (NAA), choline (CHO), and myo-inositol(ml) in order to test a neurodegenerative hypothesis inschizophrenia (decrease of NAA, increase of CHO, andincrease of ml) and a cerebral asymmetry of thesemetabolites. 1H MRS was performed in 17 schizophre-nia patients and 14 healthy subjects in three cerebralareas highly involved in the pathophysiology of schizo-phrenia (the prefrontal cortex, the thalamus, and thehippocampus). The ratio amplitudes between metabo-lites and creatine plus phosphocreatine (Cr) weredetermined. No difference in the metabolites existedbetween patients and healthy subjects. However, rela-tionships were noted between NAA/Cr and age in thethalami of the schizophrenia patients (r = -0.37; p =0.14) and healthy subjects (r = -0.52; p = 0.05). A sig-nificant correlation was observed between NAA/Crand age of onset of illness in the hippocampi of schizo-phrenia patients (r = -0.59; p < 0.05). Moreover,NAA/Cr was lower in the right than in the left pre-frontal cortex in both schizophrenia patients andhealthy subjects. There was no relationship betweenthe metabolites and duration of illness or dose ofantipsychotics. These findings might suggest a neu-rodegenerative process in the hippocampi of schizo-phrenia patients with late onset of illness, and theNAA/Cr ratio could be a marker of aging in the thai-ami.

Keywords: Schizophrenia, magnetic resonancespectroscopy, brain chemistry, thalamus, hippocam-pus, prefrontal cortex.

Schizophrenia Bulletin, 28(2):329-339, 2002.

In the past decade, neuroanatomical studies documentedcerebral impairments in schizophrenia such as lateral ven-

tricular enlargement (Andreasen et al. 1982; Nasrallah etal. 1986), reduction of volume of temporal lobes (Bogertset al. 1985; Breier et al. 1992; Marsh et al. 1994;Fukuzako et al. 1996), and decrease in thalamus size(Andreasen et al. 1994). Davis et al. (1998) recentlydescribed a bilateral increase in the size of the ventricleover a 4-year interval in a Kraepelinian subgroup. DeLisiet al. (1997b) also demonstrated significant progressivecortical atrophy in a subset of schizophrenia patientsusing magnetic resonance imaging (MRI). Other abnor-malities provide evidence of a neurodegenerative processlike the gliosis observed in the periventricular structure ofthe diencephalon, the periaqueductal region of the mesen-cephalon, and the basal forebrain (Stevens 1982;Casanova 1991). These studies have provided evidence ofalterations of cerebral structures, but the nature of theneurobiological process that could be at the origin ofschizophrenia symptoms is still unknown. A neurodegen-erative process could be investigated with *H MRS, whichis commonly used to measure NAA, ml, CHO, and Cr.NAA is an intraneuronal metabolite that reflects the num-ber or the viability of the neurons or axons; it is absent inglial cells (Birken and Oldendorf 1989; Urenjak et al.1993). Up to now, most 'H MRS studies have focused onNAA in schizophrenia and have led to inconsistentresults. Several studies (Maier 1995; Yurgelun-Todd et al.1996), although not all (Buckley et al. 1994; Heimberg etal. 1998; Bartha et al. 1999), have reported a decrease ofNAA/Cr in the temporal lobes of chronic treated schizo-phrenia patients. This reduction of NAA/Cr was present infirst episode patients (Renshaw et al. 1995). In the dorso-lateral prefrontal cortex, the NAA was reduced in chronictreated schizophrenia patients in some studies (Bertolinoet al. 1996), while other authors did not observe any

Send reprint requests to Professor S. Dollfus, GIN, UMR 6095CNRS/CEA/Universites de Caen et Paris V, and Centre Esquirol, CHU,14000 Caen, FRANCE; e-mail: [email protected].

329

by guest on March 3, 2016

http://schizophreniabulletin.oxfordjournals.org/D

ownloaded from

http://schizophreniabulletin.oxfordjournals.org/

Schizophrenia Bulletin, Vol. 28, No. 2, 2002 P. Delamillieure et al.

change of NAA/Cr in frontal lobes of chronic treatedschizophrenia patients (Buckley et al. 1994; Fukuzako etal. 1995; Williamson et al. 1998). In vitro ml has beenshown to be present in only glial, not neuronal, cells(Brand et al. 1993). Gliosis might lead to increased mlconcentration. MI has been found to be abnormally ele-vated in neurodegenerative processes such as inAlzheimer disease (Miller et al. 1993). The contributionto the CHO peak is complex and includes membranephospholipids (predominantly phosphocholine and glyc-erophosphocholine). CHO is known to increase in patho-logic conditions, such as myelin breakdown and glial cellproliferation, observed in multiple sclerosis, brain tumors,and reactive gliosis (Tedeschi et al. 1995). Fukuzako et al.(1995) found an increase of CHO/Cr in the left temporallobe of chronic schizophrenia patients compared tohealthy subjects, whereas Maier (1995) found a decreaseof CHO in the left hippocampus of the treated schizophre-nia subjects.

In order to test the neurodegenerative hypothesis inschizophrenia, in this study we examined whether adecrease of NAA and an increase of ml and CHO occur inschizophrenia patients compared to healthy subjects.

In addition, the purpose of the present investigationwas to study metabolic lateralization effects in the differ-ent volumes of interest. In fact, a defect of asymmetry wasinvestigated because disturbances in asymmetry are partic-ularly striking in schizophrenia and may provide a neuro-logical substrate for the pathophysiology of the illness(Petty 1999). Frontal asymmetry reversal and reductionwere described in chronic schizophrenia patients (Luchinset al. 1979). Hemispherical differences in metabolite con-centration with higher Phosphocreatine (PCr)/p-ATP wereobserved in the left temporal lobe compared to the righttemporal lobe of chronic schizophrenia patients (Calabreseet al. 1992). The aim of the study was also to investigatecerebral asymmetry or reverse asymmetry in schizophreniapatients compared to healthy subjects.

These abnormalities were studied in the prefrontalcortex, the thalamus, and the hippocampus, which are par-ticularly involved in schizophrenia (Weinberger 1987;Andreasen et al. 1994; Silbersweig et al. 1995).

Methods

Subjects. Seventeen right-handed schizophrenia patients(14 males and 3 females, 5 inpatients and 12 outpatients)who met DSM-IV (American Psychiatric Association1994) criteria were recruited. The diagnoses were estab-lished by two senior psychiatrists using all the informa-tion available from clinical observations, medical records,and key informants. All patients were examined using thePositive and Negative Syndrome Scale (PANSS; Kay et

al. 1987). The mean age of the schizophrenia group was31.25 years (standard deviation [SD] = 6.09), with a meanage of onset of illness (age at first episode) of 22.82 years(SD = 5.41) and a length of illness (time between theonset of illness and the MRS examination) of 8.42 years(SD = 5.45). All the treated patients were in a stabilizedphase, defined by no change in antipsychotic doses for 2months. All medicated patients were treated with antipsy-chotics (antipsychotic daily doses were converted intochlorpromazine equivalents [Davis 1976]), and five alsoreceived adjunctive anticholinergic drugs; five patientswere drug-naive.

The patients were classified as residual (n = 9), para-noid (n = 4), disorganized (n = 1), and undifferentiated(n = 3).

The handedness was denned by the Edinburgh inven-tory (Oldfield 1971). Patients had no history of headinjury, past or present neurological or organic disorders,alcoholism, or drug abuse that was evaluated by urinetests. Table 1 shows the characteristics of the patients.

Fourteen healthy subjects (11 males and 3 females)were recruited by advertisement. The mean age was 30.14years (SD = 6.39). They were matched with patients forage, educational level (elementary, secondary, universityand up), sex, and handedness (all were right-handed).None of the controls had any psychiatric disorders evalu-ated by a psychiatrist with the Diagnostic InterviewSchedule (Robins et al. 1981), organic illness, alcoholism,or drug abuse. Seven healthy subjects (50%) had not com-pleted high school. After complete description of thestudy to the subjects, written informed consent wasobtained.

Table 1. Characteristics of the patients

Educational level (n = 17)Elementary (n, %)Secondary (n, %)University (n, %)

Antipsychotics (n= 12)Amisulpride (n, %)Flupentixol (n, %)Haloperidol (n, %)Loxapine (n, %)Penfluridol (n, %)Pimozide (n, %)Pipotiazine (n, %)

PANSS (n = 17), mean ± SDPositive scaleNegative scaleGeneral psychopathology scale

1 (5.89)9 (52.94)7(41.18)

3(25)2(16.67)2(16.67)2(16.67)1 (8.33)1 (8.33)1 (8.33)

12.35 ±4.9316.35 ±6.8330.29 ±8.14

Note.—PANSS = Positive and Negative Syndrome Scale; SDstandard deviation.

330



ownloaded from


'H MRS in Schizophrenia Schizophrenia Bulletin, Vol. 28, No. 2, 2002

MRS Examination. MRI and spectroscopy were per-formed on a General Electric (GE) Signa 1.5 Tesla MRImaging system (GE Medical Systems, Milwaukee, WI)using a standard quadrature head coil. T2 weighted fast-spin echo images were used to obtain sagittal and axialviews of the brain. The average voxel size studied was 9cm3 in the left and the right medial prefrontal cortex, thala-mus, and hippocampus. The volumes of interest were pre-scribed from the axial slices and an image of the volumesof interest was obtained to confirm accurate localization(figure 1). Water unsuppressed and water suppressed withchemical shift selective (CHESS) and stimulated echoacquisition mode (STEAM)-localized spectra (RepetitionTime [TR] = 1,500 msec, Echo Time [TE] = 30 msec, mix-

ing time = 13.7 msec, 128 averages) were acquired in 4minutes, 12 seconds (Frahm et al. 1989a, 1989fc). Therewere 2,048 data points collected during an aquisition witha spectral width of 2,500 Hz. The spectral processing wasperformed on a SUN Sparc Station 10, with SA/GE soft-ware provided by GE (Spectroscopic Application for GE).Before processing the spectra, the value of the line width(Lw) of Cr was computed by fitting the peak in the fre-quency domain by Fourier transformation and withoutusing any other function (no apodization, for example).The spectra were processed using zero-filling to 8K data-points, (1-Lw) Hz line broadening in the time domain andzero-order phase, and a simple baseline correction (directcurrent [DC] offset) in the frequency domain after Fourier

Figure 1. VOIs localized on series of T1 axial 1H MRS images1

Note.—1 H MRS = proton magnetic resonance spectroscopy; VOIs = volumes of interest.1 VOIs (3 x 1.5 x 2, 5-10 ml) were placed respectively in the right and left prefrontal cortex (A, D), thalamus (B, E), and hippocampus (C, F).

331



ownloaded from



transformation. The (1-Lw) Hz line broadening was usedto normalize the Lw of all peaks. An Lw of 1 for Cr wastherefore obtained. Before the fitting procedure, the spec-tra were downscaled by the value of the amplitude of Cr.In this way a normalization of the amplitude and Lw ofcreatine to 1 was obtained. Peak amplitudes were deter-mined using a Levenberg-Marquardt fit to a Gaussian lineshape. MRS metabolite concentration was expressed asthe ratio of peak amplitudes (NAA/Cr, CHO/Cr, and

ml/Cr) (figure 2). The total time of acquisition was 1 to 2hours; it was not always possible to acquire spectra suc-cessfully from all regions of interest. Spectra were rejectedfor different reasons: first, if the resolution between theCHO and creatine peaks was not at least 50 percent of thedistance from peak maximum to apparent baseline; andsecond, if the spectral resolution or the signal-to-noiseratio was not acceptable for a good analysis (ratio ampli-tude peak/maximum noise > 4).

Figure 2. Proton spectra of healthy subject's left thalamus, acquired with the STEAM pulse sequenceat TE = 30 ms

VOX

Chemical ihift (ppm)0

Note.—CHO = choline; Cr = creatine plus phosphocreatine; ml = myo-inositol; NAA = N-acetyl-aspartate; STEAM = stimulated echoacquisition mode; TE = echo time.

332



ownloaded from



Statistical Analysis. Statistical analyses were carriedout on Statview 5.7 software. The ratios for each regionserved as dependent measures in a multivariate analysis ofvariance (MANOVA), with two grouping factors(patients, controls) by region group (prefrontal cortex,hippocampus, and thalamus) and on repeated measures(NAA/Cr, CHO/Cr, and ml/Cr). Interactions were decom-posed with univariate analysis of variance (ANOVA). Allthese tests were considered significant when p valueswere less than 0.05. Correlations between metaboliteratios and age, age of onset of illness, length of illness,and dose of antipsychotics were performed with Pearson'stest.

Results

The overall MANOVA did not show significant maineffects of diagnosis. Followup ANOVA showed signifi-cant main effects of lateralization in the prefrontal cortex.

Thalamus. Table 2 shows the means of the metaboliteratios in the right and left thalamus in schizophrenia andhealthy subjects. No difference was found between the leftand the right and between schizophrenia patients andhealthy subjects. Figure 3 shows the relationships betweenNAA/Cr and age in healthy (r = -0.52; p - 0.05) andschizophrenia (r = -0.37; p = 0.14) subjects. No correla-tion was noted between metabolite ratios and age of onsetof illness, length of illness, and dose of antipsychotics.

Prefrontal cortex. Fourteen schizophrenia patients (12males and 2 females) were analyzed using MRS. For threesubjects, the spectra in this region were not analyzedbecause they did not meet the criteria for a good analysis(very bad signal-to-noise ratio). Table 3 shows the meansof the metabolite ratios in the right and left prefrontal cor-tex in schizophrenia and healthy subjects. The NAA/Crwas significantly lower in the right than in the left pre-

frontal cortex in both schizophrenia patients and healthysubjects, but no difference was noted between patientsand controls.

There was no relationship between metabolite ratiosand age, age of onset of illness, length of illness, and doseof antipsychotics.

Hippocampus. Fourteen schizophrenia patients (12males and 2 females) were analyzed with MRS. As withthe prefrontal cortex, the spectra were not analyzed forthree subjects because they did not meet the criteria for agood analysis (very bad signal-to-noise ratio). Table 4shows the means of the metabolite ratios in the right andleft hippocampus in schizophrenia and healthy subjects.No difference was found between the right and left hemi-spheres in healthy and schizophrenia subjects. No differ-ence was noted between patients and controls.

Figure 4 shows a significant correlation between hip-pocampic NAA/Cr and age of onset of illness in schizo-phrenia patients (r = -0.59; p < 0.05). No relationship wasnoted between metabolite ratios and in particular NAAand age (r = -0.24), length of illness (r = 0.30), and doseof antipsychotics (r = 0.28).

Discussion

The major findings of this study were a significant corre-lation between NAA/Cr and age of onset of illness in thehippocampi of schizophrenia patients, a lower NAA/Cr inthe right than in the left prefrontal cortex in both schizo-phrenia patients and healthy subjects, and a negative rela-tionship between NAA/Cr and age in the thalami ofhealthy subjects. In contrast, we did not find any differ-ence of NAA/Cr, ml/Cr, and CHO/Cr between schizo-phrenia patients and healthy subjects. There were no cor-relations between age of onset of illness and metaboliteratios in the prefrontal cortex and thalamus or between thelength of illness and metabolite ratios in the prefrontalcortex, thalamus, and hippocampus.

Table 2. Means ± standard deviations of metabolite peak amplitude ratios in left and right thalamus inschizophrenia patients and healthy subjects

Ratio Thalamus

Metabolite Ratios

ControlsSchizophrenia

subjects

NAA/Cr

CHO/Cr

ml/Cr

LeftRight

LeftRight

LeftRight

1.36±0.151.41 ±0.14

0.89 ±0.080.94 ±0.10

0.57 ± 0.080.59 ±0.06

1.36 ±0.151.36 ±0.22

0.88 ±0.100.88 ±0.15

0.53 + 0.080.55 ±0.09

333



ownloaded from



Figure 3. Negative correlation between age andNAA/Cr in the thalamus of controls (r = -0.52; p =0.05) and schizophrenia patients (r=-0.37; p = 0.14)

• Controls —

® Patients

1,9 •

1,8 "

1,7 •

1,6 "

1.5 "

1,4 '

1.3 "

1.1 •

•——-

•

_i 1 —

©

« ^

• o

_i 1 1——i 1 1

•

•

0

• ^

22.5 25 27,5 30 32.5 3S 37,5 40 42,5 45Age (years)

Note.—Cr = creatine plus phosphocreatine; NAA = N-acetyl-aspartate.

Decreased NAA levels have been observed by ]HMRS in some cerebral pathologies involving cell damageand loss (Arnold et al. 1992). The absence of change ofNAA/Cr in schizophrenia patients compared to healthysubjects confirms some previous results. Fukuzako et al.(1995) and Buckley et al. (1994) did not find a decrease ofNAA/Cr in the frontal lobes in chronic schizophreniapatients. Williamson et al. (1998) did not find a decreaseof NAA in the medial prefrontal cortex in chronic schizo-phrenia patients. Bertolino et al. (1996) did not observeany difference of metabolite ratios between chronicpatients and controls in thalami, but ml/Cr was not tested.However, a decrease of either NAA/Cr or NAA wasobserved by other authors in the frontal areas (Bertolinoet al. 1996, 1998; Deicken et al. 1997; Cecil et al. 1999)or temporal areas (Maier 1995; Bertolino et al. 1996,1998; Fukuzako et al. 1996; Yurgelun-Todd et al. 1996;Cecil et al. 1999) of chronic schizophrenia patients. Therewas a trend toward a higher NAA/Cr in the temporal lobein schizophrenia patients compared to controls. This trendcould be due to the patients with early onset of illness(figure 4), suggesting an active process or neuronal dys-function in this subtype of patients. Our result concerningthe CHO/Cr is consistent with several earlier reports

Table 3. Means ± standard deviations of metabolite peak amplitude ratios in left and right prefrontalcortex in schizophrenia patients and healthy subjects

Ratio

NAA/Cr1

CHO/Cr

ml/Cr

1 Significant difference between right and left (F = 4.27, df = 1, 24, p = 0.05).

Prefrontal cortex

LeftRight

LeftRight

LeftRight

Controls

1.40 ± 0.251.33 ±0.17

0.90 ±0.150.94 ± 0.08

0.65 ±0.140.66 ±0.11

Metabolite Ratios

Schizophreniasubjects

1.41 ±0.231.25 ±0.23

0.99 ±0.180.92 ±0.17

0.65 ± 0.220.60 ±0.16

Table 4. Means ± standard deviations of metabolite peak amplitude ratios in left and righthippocampus in schizophrenia patients and healthy subjects

Ratio

NAA/Cr

CHO/Cr

ml/Cr

Hippocampus

LeftRight

LeftRight

LeftRight

Controls

1.14 ±0.161.16 ± 0.13

0.95 ± 0.090.94 ±0.10

0.79 + 0.110.74 ± 0.08

Metabolite Ratios

Schizophreniasubjects

1.26 ±0.291.25±0.16

0.90 ±0.120.90 ±0.13

0.69 ±0.180.69 ±0.14

334



ownloaded from



Figure 4. Negative correlation between age ofonset of illness and hippocampic NAA/Cr ofschizophrenia patients (r = -0.59; p< 0.05)

14 16 18 20 22 24 26 28 30 32 34 36Age of onset of illness (years)

Note.—Cr = creatine plus phosphocreatine; NAA = N-acetyl-aspartate.

(Fukuzako et al. 1996; Williamson et al. 1998) but is dif-ferent from those of Cecil et al. (1999), who reported anincrease of the CHO/Cr ratio in the frontal lobe of chronicschizophrenia patients compared to healthy subjects.There was also no difference in temporal CHO/Cr inpatients compared to controls. This is consistent withsome previous studies (Yurgelun-Todd et al. 1996). CHOis known to increase in neurodegenerative process asdemyelinating lesions (DeStephano et al. 1995).

In our study, prefrontal, thalamic, and hippocampicml/Cr were similar in patients and control subjects. Thediscrepancies between our findings and those of otherresearchers could be due to the differences of size andlocalization of the volume of interest (VOI) or the clinicalcharacteristics of the patients. Indeed, in our study, onlythe medial and not the dorsolateral prefrontal cortex wasstudied, and the VOI was wider than in other studies.Another explanation is the heterogeneity of schizophre-nia. These points can be considered as limitations of thestudy and could contribute to these differences.

As ml has been identified as a glial cell marker(Urenjak et al. 1993), the absence of increase of ml is anargument against a glial reaction. Therefore, our resultsdid not support the hypothesis of a neurodegenerativeprocess in schizophrenia because no decrease of NAA/Crand no increase of ml/Cr and CHO/Cr were found in thethree areas (hippocampus, thalamus, and prefrontal cor-tex). However, a significant correlation between NAA/Crand age of onset of illness in the hippocampi of schizo-phrenia patients was found (r = -0.59; p < 0.05) (figure4). This result was close to Fukuzako et al. (1995), whoobserved this correlation in the temporal lobes of chronic

schizophrenia patients. Thus, assuming that NAA is aneuronal or axonal number or viability marker, a decreaseof NAA/Cr with age of onset might suggest a neurode-generative process in hippocampi but only in a subtype ofschizophrenia characterized, as some authors have alreadysuggested (DeLisi 1992; Murray et al. 1992), by a lateonset of illness. This result was not due to the age of thesubjects, because no correlation between age and metabo-lite ratios and in particular NAA in the hippocampus wasfound in the patients (r = -0.24) or the healthy subjects (r= -0.19). These results, which could reflect a neuronalloss or dysfunction without gliosis, might suggest a neu-rodegenerative process by apoptosis (there was no corre-lation between ml and CHO/Cr and age of onset of ill-ness) in a subtype of schizophrenia with late onset ofillness. Contrarily to Ende et al. (2000), who found adecrease of NAA with the duration of illness in the ante-rior cingulate region of chronic schizophrenia patients, nocorrelation between length of illness and metabolite ratioswas found in the three areas.

The ratio NAA/Cr was significantly lower in the rightthan in the left prefrontal cortex in healthy subjects andpatients, suggesting an asymmetry in the prefrontal cortexin both schizophrenia and healthy subjects. Measurementson computed tomography scans and postmortem brainshave long suggested that the right frontal region in normalbrains extends further forward and is wider than the left(LeMay 1976; Weinberger et al. 1982; Kertesz et al.1990). Golden et al. (1981) found that there was a trendfor a greater density in the left hemisphere than in theright hemisphere in both schizophrenia patients and con-trol subjects. In the same way, Gur et al. (1980) foundwith a xenon-133 inhalation method a greater density ofcells in the left than in the right hemisphere in schizophre-nia patients (there were no controls). Thus, this asymme-try of NAA/Cr ratios could suggest a lower neuronal den-sity in the right than in the left prefrontal cortexcompatible with a wider right than left prefrontal cortex.However, an asymmetry reversal or reduction was notobserved in schizophrenia patients, contrary to the obser-vations of Bilder et al. (1994) and DeLisi et al. (1997a).These discrepancies could be due to the fact that thesestudies were structural imaging and not metabolic studies.

There was no asymmetry in the temporal lobe.Structural MRI studies have reported lateralized changesin the temporal lobes, such as a reduction of the left tem-poral lobe (DeLisi et al. 1990; Rossi et al. 1990). Indeed,these studies were focused on the planum temporale,while the VOI in our study was centered on the hip-pocampus. The planum temporale is well known to beasymmetric in healthy subjects, and a large study of litera-ture has reported a reduction or a reverse asymmetry ofthis structure in schizophrenia patients (Rossi et al. 1990).

335



ownloaded from



In consequence, the absence of asymmetry in metaboliteratios in our study was probably due to a difference oflocalization of the VOL

A negative relationship between NAA/Cr and age inhealthy subjects and a correlation (although not signifi-cant) in schizophrenia patients (figure 3) were found inthe thalami. This suggests that NAA might be a marker ofaging in particular in the thalamus and that the older sub-jects were, the lower ratios of NAA/Cr were. This result isalso consistent with Charles et al. (1994), who found thatNAA was lower in older subjects in the voxel represent-ing cortical and subcortical gray matter.

There was no relationship between metaboliteratios and dose of antipsychotics, contrary to the resultsof Shioiri et al. (1996), who found a significant positivecorrelation between the chlorpromazine equivalentantipsychotic dosage and the level of NAA. However,their result was found in striata, areas particularlyinvolved in extrapyramidal symptoms secondary toantipsychotics.

Several technical limitations of the present studyhave to be considered. First, the voxels contained vari-ous proportions of gray matter and white matter.Determining these proportions could be informativebecause some studies in schizophrenia have reported adecrease in NAA in white but not in gray matter (Lim etal. 1998). It was also not possible to estimate the extentto which the prefrontal, hippocampal, and thalamic vox-els contained tissue that was not within these structures.Therefore, the asymmetry that we found in the pre-frontal cortex has to be confirmed in future studies withsegmentation. The second limitation is the large size ofthe VOI. Third, absolute metabolic quantification,although preferable, was not feasible in our protocol.We used metabolite ratios because of the best reliabilityobtained with this method compared to the method ofabsolute concentration. The fourth limitation involvesthe power of statistical analyses: for a significant differ-ence of about 0.1 or 0.3 in metabolite ratio, the noncen-trality parameter is 0.5 or 1.5, that is to say an estimat-ing calculated power of 40 percent or 80 percent,respectively; the small size of our sample of patientsand the great value of standard deviations may have dis-missed true differences between patients and controlsbecause of type II error in the statistical analyses. Thefifth limitation is the possibility of effects of antipsy-chotic agents on NAA, although there was no differencebetween drug-naive and treated patients.

In conclusion, the present study might provide supportfor a neurodegenerative process, but only in some schizo-phrenia patients with a late onset of illness. However, thishypothesis needs to be confirmed in future studies withsegmentation and a larger number of subjects.

ReferencesAmerican Psychiatric Association. DSM-IV: Diagnosticand Statistical Manual of Mental Disorders. 4th ed.Washington, DC: APA, 1994.

Andreasen, N.C.; Arndt, S.; Swayze, V.W., 2nd; Cizadlo,T ; Flaum, M.; O'Leary, D.; Ehrhardt, J.C.; and Yuh,W.T.C. Thalamic abnormalities in schizophrenia visual-ized through magnetic resonance image averaging.Science, 266:294-298, 1994.

Andreasen, N.C.; Olsen, S.A.; Dennert, J.W.; and Smith,M.R. Ventricular enlargement in schizophrenia:Relationship to positive and negative symptoms.American Journal of Psychiatry, 139:297-302, 1982.

Arnold, D.L.; Matthews, P.M.; Francis, G.S.; O'Connor,J.; and Antel, J.P. Proton magnetic resonance spectro-scopic imaging for metabolic characterization of demyeli-nating plaques. Annals of Neurology, 31:235-241, 1992.Bartha, R.; Al-Semaan, Y.M.; Williamson, P.C.; Drost,D.J.; Malla, A.K.; and Carr, T.J. A short echo proton mag-netic resonance spectroscopy study of the left medial tem-poral lobe in first-onset schizophrenia patients. BiologicalPsychiatry, 45:1403-1411, 1999.

Bertolino, A.; Callicott, J.H.; Elman, I.; Mattay, VS.;Tedeschi, G.; Frank, J.A.; Breier, A.; and Weinberger,D.R. Regionally specific neuronal pathology in untreatedpatients with schizophrenia: A proton magnetic resonancespectroscopic imaging study. Biological Psychiatry,43:641-648, 1998.

Bertolino, A.; Nawroz, S.; Mattay, VS.; Barnett, A.S.;Duyn, J.H.; Moonen, C.T.W.; Frank, J.A.; Tedeschi, G.;and Weinberger, D.R. Regionally specific pattern of neu-rochemical pathology in schizophrenia as assessed bymultislice proton magnetic resonance spectroscopic imag-ing. American Journal of Psychiatry, 153:1554-1563,1996.

Bilder, R.M.; Wu, H.; Bogerts, B.; Degreef, G.; Ashtari,M.; Alvir, J.M.J.; Snyder, P.J.; and Lieberman, J.A.Absence of regional hemispheric volume asymmetries infirst-episode schizophrenia. American Journal ofPsychiatry, 151:1437-1447, 1994.

Birken, D.L., and Oldendorf, W.H. N-acetyl-L-asparticacid: A literature review of a compound prominent in 1H-NMR spectroscopic studies of brain. Neuroscience andBiobehavioral Reviews, 13:23-31, 1989.

Bogerts, B.; Meertz, E.; and Schonfeldt-Bausch, R. Basalganglia and limbic system pathology in schizophrenia: Amorphometric study of brain volume and shrinkage.Archives of General Psychiatry, 42:784-791, 1985.

Brand, A.; Richter-Landsberg, C ; and Leibfritz, D.Multinuclear NMR studies on the energy metabolism of

336



ownloaded from



glial and neuronal cells. Development in Neuroscience,15:289-298, 1993.

Breier, A.; Buchanan, R.W.; Elkashef. A.; Munson, R.C.;Kirkpatrick, B.; and Cellad, F. Brain morphology andschizophrenia: A magnetic resonance imaging study oflimbic, prefrontal cortex, and caudate structures. Archivesof General Psychiatry, 49:921-926, 1992.

Buckley, P.F.; Moore, C ; Long, H.; Larkin, C ;Thompson, P.; Mulvany, F.; Redmond, O.; Stack, J.P.;Ennis, J.T.; and Waddington, J.L. 1H magnetic reso-nance spectroscopy of the left temporal and frontal lobesin schizophrenia: Clinical, neurodevelopmental, andcognitive correlates. Biological Psychiatry, 36:792-800,1994.

Calabrese, G.; Deicken, R.F.; Fein, G.; Merrin, E.L.;Schoenfeld, F.; and Weiner, M.W. 31Phosphorus magneticresonance spectroscopy of the temporal lobes in schizo-phrenia. Biological Psychiatry, 32:26-32, 1992.

Casanova, M.F. Astrocytosis and schizophrenia.Schizophrenia Research, 5:186-187, 1991.

Cecil, K.M.; Lenkenski, R.E.; Gur, R.E.; and Gur, R.C.Proton magnetic resonance spectroscopy in the frontaland temporal lobes of neuroleptic-naive patients withschizophrenia. Neuropsychopharmacology, 20:131-140,1999.

Charles, H.C.; Lazeyras, F.; Krishnan, K.R.; Boyko, O.B.;Patterson, L.J.; Doraiswamy, P.M.; and McDonald, W.M.Proton spectroscopy of human brain: Effects of age andsex. Progress in Neuropsychopharmacology andBiological Psychiatry, 18(6):995-1004, 1994.

Davis, J.M. Comparative doses and costs of antipsychoticmedication. Archives of General Psychiatry, 33:858-861,1976.

Davis, K.L.; Buchsbaum, M.S.; Shihabuddin, L.;Spiegel-Cohen, J.; Metzger, M.; Frecska, E.; Keefe,R.S.; and Powchik, P. Ventricular enlargement in poor-outcome schizophrenia. Biological Psychiatry,43:783-793, 1998.

Deicken, R.F.; Zhou, L.; Corwin, F.; Vinagradov, S.; andWeiner, M.W. Decreased left frontal lobe N-acetylaspar-tate in schizophrenia. American Journal of Psychiatry,154:688-690, 1997.

DeLisi, L.E. The significance of age of onset for schizo-phrenia. Schizophrenia Bulletin, 18(2):209-215, 1992.

DeLisi, L.E.; Gupta, S.M.; Hoff, A.; Shields, A.;Schwartz, J.; Halthore, S.; and Anand, A. A brain mor-phology in first episode cases of schizophrenia.Schizophrenia Research, 3:20, 1990.

DeLisi, L.E.; Sakuma, M.; Kushner, M.; Finer, D.L.;Hoff, A.L.; and Crow, T.J. Anomalous cerebral asymme-

try and language processing in schizophrenia.Schizophrenia Bulletin, 23(2):255-271, 1997a.

DeLisi, L.E.; Sakuma, M.; Tew, W.; Kushner, M.; Hoff,A.L.; and Grimson, R. Schizophrenia as a chronic activebrain process: A study of progressive brain structuralchange subsequent to the onset of schizophrenia.Psychiatry Research, 74:129-140, 19976.

DeStephano, N.; Matthews, P.M.; Antel, J.P.; Preul, M.;Francis, G.; and Arnold, D.L. Chemical pathology ofacute demyelinating lesions and its correlation with dis-ability. Annals of Neurology, 38:901-909, 1995.Ende, G.; Braus, D.F.; Walter, S.; Weber-Fahr, W.; Soher,B.; Maudsley, A.A.; and Henn, F.A. Effects of age, med-ication and illness duration on the N-acetyl-aspartate sig-nal of the anterior cingulated region in schizophrenia.Schizophrenia Research, 41:389-395, 2000.

Frahm, J.; Bruhn, H.; Gyngell, M.L.; Merboldt, K.D.;Hanicke, W.; and Sauter, R. Localized high resolutionproton NMR spectroscopy using stimulated echos: Initialapplication to human brain in vivo. Magnetic Resonancein Medicine, 9:73-93, 1989a.

Frahm, J.; Bruhn, H.; Gyngell, M.L.; Merboldt, K.D.;Hanicke, W.; and Sauter, R. Localized proton NMRspectroscopy in different regions of the human brain invivo: Relaxation times and concentrations of cerebralmetabolites. Magnetic Resonance in Medicine,11:47-63, 19896.

Fukuzako, H.; Fukuzako, T.; Hashiguchi, T.; Hokazono,Y.; Takeuchi, K.; Hirakawa, K.; Ueyama, K.; Takigawa,M.; Kajima, Y.; Nakajo, M.; and Fujimoto, T. Reductionin hippocampal formation volume is caused mainly by itsshortening in chronic schizophrenia: Assessment by MRI.Biological Psychiatry, 39:938-945, 1996.

Fukuzako, H.; Takeuchi, K.; Hokazono, Y.; Fukuzako, T.;Yamada, K.; Hashiguchi, T.; Obo, Y.; Ueyama, K.;Takigawa, M.; and Fujimoto, T. Proton magnetic reso-nance spectroscopy of the left medial temporal and frontallobes in chronic schizophrenia: Preliminary report.Psychiatry Research, 61:193-200, 1995.

Golden, C.J.; Graber, B.; Coffman, J.; Berg, R.A.;Newlin, D.B.; and Block, S. Structural brain deficits inschizophrenia: Identification by computed tomographicscan density measurements. Archives of GeneralPsychiatry, 38:1014-1017, 1981.

Gur, R.C; Packer, I.K.; Hungerbuhler, J.P.; Reivich, M.;Obrist, W.D.; Amarneck, W.S.; and Sackheim, H.A.Differences in the distribution of gray and white matter inhuman cerebral hemispheres. Science, 207:1226-1228,1980.

Heimberg, C ; Komoroski, R.A.; Lawson, W.B.;Cardwell, D.; and Karson, C.N. Regional proton mag-

337



ownloaded from



netic resonance spectroscopy in schizophrenia andexploration of drug effect. Psychiatry Research,83:105-115, 1998.

Kay, S.R.; Fiszbein, A.; and Opler, L.A. The Positive andNegative Syndrome Scale (PANSS) for schizophrenia.Schizophrenia Bulletin, 13(2):261-276, 1987.

Kertesz, A.; Polk, M.; and Black, S.E.J.H. Sex, handed-ness, and the morphometry of cerebral asymmetries onmagnetic resonance imaging. Brain Research, 530:40-48,1990.

LeMay, M. Morphological cerebral asymmetries of mod-ern man, fossil man, and non-human primate. Annals ofNY Academy of Sciences, 280:349-356, 1976.

Lim, K.O.; Adalsteinsson, E.; Spielman, D.; Sullivan,E.V.; and Rosenbloom, M.J. Proton magnetic resonancespectroscopic imaging of cortical gray and white matter inschizophrenia. Archives of General Psychiatry,55:346-352, 1998.

Luchins, D.J.; Weinberger, D.R.; and Wyatt, R.J.Schizophrenia: Evidence of a subgroup with reversalcerebral asymmetry. American Journal of Psychiatry,36:1309-1311, 1979.

Maier, M. In vivo magnetic resonance spectroscopy:Applications in psychiatry. British Journal of Psychiatry,167:299-306, 1995.

Marsh, L.; Suddath, R.L.; Higgins, N.; and Weinberger,D.R. Medial temporal lobe structures in schizophrenia:Relationship of size to duration of illness. SchizophreniaResearch, 11:225-238, 1994.

Miller, B.L.; Moats, R.; Shonk, T.; Ernst, T.; Wooley, S.;and Ross, B.D. Alzheimer's disease: Depiction ofincreased cerebral myo-inositol with proton MR spec-troscopy. Radiology, 187:433^137, 1993.

Murray, R.M.; O'Callaghan, E.; Castle, D.J.; and Lewis,S.W. A neurodevelopmental approach to the classificationof schizophrenia. Schizophrenia Bulletin, 18(2):319—332,1992.

Nasrallah, H.A.; Olson, S.C.; McCalley-Whitters, M.;Chapman, S.; and Jacoby, C.G. Cerebral ventricularenlargement in schizophrenia. Archives of GeneralPsychiatry, 43:157-159, 1986.

Oldfield, R.C. The assessment and analysis of handed-ness: The Edinburgh inventory. Neuropsychologia,9:97-113, 1971.

Petty, R.G. Structural asymmetries of the human brain andtheir disturbance in schizophrenia. Schizophrenia Bulletin,25(1):121-139, 1999.

Renshaw, P.F.; Yurgelun-Todd, D.A.; Tohen, M.; Gruber,S.A.; and Cohen, B.M. Temporal lobe proton magneticresonance spectroscopy of patients with first-episode psy-

chosis. American Journal of Psychiatry, 152:444-446,1995.

Robins, L.N.; Helzer, J.E.; Groughan, J.; and Ratcliff,K.S. National Institute of Mental Health DiagnosticInterview Schedule: Its history, characteristics and valid-ity. A rchives of General Psychiatry, 38:381-389, 1981.

Rossi, G.F.; Stratta, P.; D'Albenzio, L.; Tartaro, A.;Schiazza, G.; Di Michele, V.; Bolino, F; and Casacchia,M. Reduced temporal lobe areas in schizophrenia:Preliminary evidence from a controlled multiplanar mag-netic resonance study. [Abstract]. SchizophreniaResearch, 3:21, 1990.

Shioiri, T; Hamakawa, H.; Kato, T.; Murashita, J.; Fujii,K.; Inubushi, T; and Takahashi, S. Proton magnetic reso-nance spectroscopy of the basal ganglia in patients withschizophrenia: A preliminary report. SchizophreniaResearch, 22:19-26, 1996.

Silbersweig, D.A.; Stern, E.; Frith, C ; Cahill, C ; Holmes,A.; Grootoonk, S.; Seaward, J.; McKenna, P.; Chua, S.E.;Schnorr, L.; Jones, T.; and Frackowiak, R.S.J. A func-tional neuroanatomy of hallucinations in schizophrenia.Nature, 378:176-179, 1995.

Stevens, J.R. Neuropathology of schizophrenia. Archivesof General Psychiatry, 39:1131-1139, 1982.

Tedeschi, G.; Bertolino, A.; Righini, A.; Campbell, G.;Raman, R.; Duyn, J.H.; Moonen, C.T.W.; Alger, J.R.; andDiChiro, G. Brain regional distribution pattern of metabo-lite signal intensities in young adults by proton magneticresonance spectroscopic imaging. Neurology,45:1384-1391, 1995.

Urenjak, J.; Williams, S.R.; Gadian, D.G.; and Noble, M.Proton nuclear magnetic resonance spectroscopy unam-biguously identifies different neural cell types. Journal ofNeuroscience, 13:981-989, 1993.

Weinberger, D.R. Implications of normal brain develop-ment for the pathogenesis of schizophrenia. Archives ofGeneral Psychiatry, 44:660-669, 1987.

Weinberger, D.R.; Luchins, D.J.; Morihisa, J.; and WyattR.J. Asymmetrical volumes of the right and left frontaland occipital regions of the human brain. Annals ofNeurology, 11:97-100, 1982.

Williamson, P C ; Bartha, R.; Drost, D.J.; Malla, A.; Carr,T.; Canaran, G.; Densmore, M.; and Neufeld, R.W.J.Medial prefrontal 1H MRS and MRI findings in never-treated schizophrenics and healthy controls. [Abstract].Schizophrenia Research, 29:101, 1998.

Yurgelun-Todd, D.A.; Renshaw, P.F.; Gruber, S.A.;Watemaux, CM.; and Cohen, B.M. Proton magnetic reso-nance spectroscopy of the temporal lobes in schizophren-ics and normal controls. Schizophrenia Research,19:55-59, 1996.

338



ownloaded from



The Authors

Pascal Delamillieure, M.D., Ph.D., is Psychiatrist, Grouped'Imagerie Neurofonctionelle (GIN), UMR 6095CNRS/CEA/Universites de Caen et Paris V, Caen, France;and Centre Esquirol, Centre Hospitalier et Universitaire,Caen, France. Jean-Marc Constans, M.D., isNeuroradiologist; Jesus Fernandez, Engineer in Informatics;and Patrick Courtheoux, M.D., is Professor ofNeuroradiology, Departement d'IRM, Centre Hospitalier etUniversitaire, Caen, France. Perine Brazo, M.D., is

Psychiatrist, GIN, UMR 6095 CNRS/CEA/Universites deCaen et Paris V; and Centre Esquirol, Centre Hospitalier etUniversitaire. Karim Benali, M.D., is head of the Departmentof Psychiatry, GIN, UMR 6095 CNRS/CEA/Universites deCaen et Paris V. Florence Thibaut, M.D., Ph.D., is Professorof Psychiatry; and Michel Petit, M.D., is Professor ofPsychiatry, Unite Inserm EPI 9906, IFR 23, CentreHospitalier et Universitaire de Rouen, Rouen, France. SoniaDollfus, M.D., Ph.D., is Professor of Psychiatry, GIN, UMR6095 CNRS/CEA/Universites de Caen et Paris V; and CentreEsquirol, Centre Hospitalier et Universitaire.

339



ownloaded from




ownloaded from


proton magnetic resonance spectroscopy ( 1h-mrs) of the thalamus in schizophrenia

Documents