NeuroImage: Clinical 6 (2014) 398–407

Contents lists available at ScienceDirect

NeuroImage: Clinical
journal homepage: www.elsevier.com/locate/ynicl

GABA and glutamate in schizophrenia: A 7 T 1H-MRS study

Anouk Marsman a,⁎,1, René C.W. Mandl a, Dennis W.J. Klompb, Marc M. Bohlken a, Vincent O. Boer b,
Anna Andreychenkoc, Wiepke Cahna, René S. Kahn a, Peter R. Luijten b, Hilleke E. Hulshoff Pol a
a
Department of Psychiatry, Brain Center Rudolf Magnus, University Medical Center Utrecht, The Netherlands
b
Department of Radiology, University Medical Center Utrecht, The Netherlands
c
Department of Radiotherapy, University Medical Center Utrecht, The Netherlands

a r t i c l e i n f o a b s t r a c t

Article history: Schizophrenia is characterized by loss of brain volume, which may represent an ongoing pathophysiological pro-
Received 21 April 2014 cess. This loss of brain volume may be explained by reduced neuropil rather than neuronal loss, suggesting abnor-
Received in revised form 30 September 2014 mal synaptic plasticity and cortical microcircuitry. A possible mechanism is hypofunction of the NMDA-type of
Accepted 11 October 2014
glutamate receptor, which reduces the excitation of inhibitory GABAergic interneurons, resulting in a disinhibi-
Available online 15 October 2014
tion of glutamatergic pyramidal neurons. Disinhibition of pyramidal cells may result in excessive stimulation by
Keywords:
glutamate, which in turn could cause neuronal damage or death through excitotoxicity.
Schizophrenia In this study, GABA/creatine ratios, and glutamate, NAA, creatine and choline concentrations in the prefrontal and
GABA parieto-occipital cortices were measured in 17 patients with schizophrenia and 23 healthy controls using proton
Glutamate magnetic resonance spectroscopy at an ultra-high magnetic field strength of 7 T. Significantly lower GABA/Cr ra-
1
H-MRS tios were found in patients with schizophrenia in the prefrontal cortex as compared to healthy controls, with
7T GABA/Cr ratios inversely correlated with cognitive functioning in the patients. No significant change in the
GABA/Cr ratio was found between patients and controls in the parieto-occipital cortex, nor were levels of gluta-
mate, NAA, creatine, and choline differed in patients and controls in the prefrontal and parieto-occipital cortices.
Our findings support a mechanism involving altered GABA levels distinguished from glutamate levels in the me-
dial prefrontal cortex in schizophrenia, particularly in high functioning patients. A (compensatory) role for GABA
through altered inhibitory neurotransmission in the prefrontal cortex may be ongoing in (higher functioning) pa-
tients with schizophrenia.
© 2014 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/3.0/).

1. Introduction 2013). Glutamate (Glu) is the most abundant excitatory neurotransmit-

ter in the central nervous system (CNS) (Fonnum, 1984). Recent meta-
Schizophrenia is characterized by loss of brain volume that is at least analyses of cross-sectional studies show that glutamate levels are
in part progressive in nature, and which may represent an ongoing path- decreased in chronic schizophrenia patients and elevated in medication-
ophysiological process (Hulshoff Pol et al., 2008). This loss of brain vol- naïve and medication free patients (Marsman et al., 2013a; Poels et al.,
ume may be explained by reduced neuropil rather than neuronal loss, 2014). It may thus be that glutamate is elevated in the early stages of
suggesting abnormal synaptic plasticity and cortical microcircuitry schizophrenia and decreases after treatment with antipsychotic medica-
(Harrison, 1999). The mechanisms underlying brain volume loss in tion. Indeed, pharmacologically induced NMDA receptor hypofunction
schizophrenia are not known, although it has been suggested that the causes increases in glutamate levels (Moghaddam et al., 1997). Also, a
glutamatergic system may be involved (Harrison and Weinberger, longitudinal study (De la Fuente-Sandoval et al., 2013) and a cross-
2005). A recent cross-sectional study showed that increased hippocam- sectional study (Kegeles et al., 2012) showed that antipsychotic medica-
pal glutamatergic levels in unmedicated schizophrenia patients were tion decreases the elevated glutamate levels in the unmedicated state. Al-
correlated with hippocampal volume reductions (Kraguljac et al., tered glutamate levels may be a result of hypofunction of the NMDA-type
of glutamate receptor (Olney et al., 1999; Kondziella et al., 2007; Stone
et al., 2009). NMDA-receptor hypofunction reduces the excitation of in-
* Corresponding author at: Department of Psychiatry, Internal address A.01.126, hibitory GABAergic interneurons, resulting in a disinhibition of gluta-
University Medical Center Utrecht, PO Box 85500, Utrecht 3508 GA, The Netherlands. matergic pyramidal neurons, which may drive a hyperdopaminergic
E-mail address: [email protected], [email protected] (A. Marsman).
1
Present address: Johns Hopkins University School of Medicine, Russell H. Morgan
state that produces psychosis. A major effect of NMDA receptor antago-
Department of Radiology and Radiological Science, 600 N Wolfe St. Park 359, Baltimore, nism may be to produce disinhibition of pyramidal cells (Lisman et al.,
MD 21287, USA. 2008; Lewis and Moghaddam, 2006). Disinhibition of pyramidal cells

http://dx.doi.org/10.1016/j.nicl.2014.10.005
2213-1582/© 2014 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).
 A. Marsman et al. / NeuroImage: Clinical 6 (2014) 398–407 399

may result in excessive stimulation by glutamate, which in turn could parieto-occipital GABA and glutamate levels are unchanged in schizo-
cause neuronal damage or death through excitotoxicity. One of the phrenia (Goto et al., 2009; Ongür et al., 2010; Yoon et al., 2010). To
meta-analyses showed an increase in glutamine, the precursor of gluta- our knowledge this is the first study in schizophrenia patients using
1
mate, in schizophrenia, which may lead to increased glutamate levels as H-MRS at a magnetic field strength of 7 T.
reported in medication-naïve and medication free patients (Marsman
et al., 2013a; Poels et al., 2014). However, in case of decreased glutamate 2. Methods
as seen in chronic patients, it could also suggest a deficiency in glutamin-
ase, the enzyme that converts glutamine into glutamate (Marsman et al., 2.1. Subjects
2013a).
GABA has indeed been implicated in schizophrenia based on post- A total of 40 individuals participated in the study, including 17 pa-
mortem and animal studies showing reduced expression of pre- and tients with schizophrenia, and 23 healthy control participants matched
postsynaptic markers of GABAergic neurotransmission in subpopula- to the patients for age, sex and their parents3 socio-economic status.
tions of GABAergic interneurons in the prefrontal cortex in postmortem Healthy participants had no major psychiatric or neurological history,
and animal studies (Benes and Berretta, 2001; Lewis et al., 2004; Coyle, no history of drug or alcohol abuse, and no first-degree relatives with
2004). However, postmortem studies generally include patients who psychiatric or neurological disorders. The study was approved by the
have been ill for many years. To examine glutamate and GABA levels Medical Ethics Committee of the University Medical Center Utrecht,
in the early stages of the disease, in vivo measurements using magnetic The Netherlands, and performed according to the directives of the Dec-
resonance spectroscopy (MRS) are the method of choice (De la Fuente- laration of Helsinki (amendment of Seoul, 2008). All participants pro-
Sandoval et al., 2011; Goto et al., 2009). vided written informed consent prior to the examination.
So far, a total of seven studies investigated GABA levels in vivo in the All individuals underwent psychiatric assessment using the Compre-
brains of patients with schizophrenia, using proton MRS (1H -) (Kegeles hensive Assessment of Symptoms and History (Andreasen et al., 1992).
et al., 2012; De la Fuente-Sandoval et al., 2011; Goto et al., 2009; Ongür Symptom severity was assessed using the Positive and Negative Syn-
et al., 2010; Tayoshi et al., 2010; Yoon et al., 2010; Rowland et al., 2013), drome Scale (PANSS) (Kay et al., 1987). Current daily dosage and cu-
which reported increased (Kegeles et al., 2012; Ongür et al., 2010), de- mulative life-time dosage of antipsychotic medication in haloperidol
creased (Goto et al., 2009; Yoon et al., 2010), and normal (Kegeles equivalents were calculated (Andreasen et al., 2010). In addition, cur-
et al., 2012; Goto et al., 2009; Tayoshi et al., 2010; Yoon et al., 2010; rent or prior use of benzodiazepines was established. All individuals
Rowland et al., 2013) GABA levels dependent on disease stage, medica- underwent a general cognitive assessment using the full Wechsler
tion intake and brain area studied. The only study of GABA in unmedi- Adult Intelligence Scale (WAIS) — III (Wechsler, 1997), which revealed
cated patients found elevated GABA levels in the medial prefrontal a total intelligence quotient (TIQ) as well as a verbal and performance
cortex in unmedicated patients, but not in medicated patients, suggest- IQ, and working memory, perceptual reasoning, and verbal comprehen-
ing that antipsychotic medication decreases the elevated GABA levels sion indices. Cognitive assessment could be successfully completed in
(Kegeles et al., 2012). However, since these studies have been done at all but three healthy subjects because of logistical reasons. One patient
conventional magnetic field strengths (such as 3 T) subtle changes in was not able to complete the MR acquisition because of a relapse in
these metabolite levels may have remained concealed. symptoms.
Performing 1H-MRS at ultra-high magnetic field strength (7 T) has No differences between the groups were found for age, parental
two clear advantages. The sensitivity of the measurements of metabo- completed level of education and sex. Groups differed significantly in
lite concentrations is increased because of the increased signal-to- completed level of education, with patients having lower education,
noise ratio (SNR) and the increased spectral resolution allows for a bet- and on general cognitive functioning, with patients having a lower
ter separation of the individual metabolite spectra. For instance, at a level of intelligence, as compared to the controls (Table 1).
magnetic field strength of 7 T it is now possible to adequately separate
the glutamate and glutamine signals resulting in a higher accuracy of 2.2. MR acquisition
glutamate measurement (Tkác et al., 2001). However, detection of me-
tabolites at a higher magnetic field strength is complicated by reduced ac- All investigations were performed on a 7 T whole body MR scanner
curacy in localization of the region of interest (Boer et al., 2011; (Philips, Cleveland, OH, USA). A birdcage transmit head coil was used
Andreychenko et al., 2012). The sLASER (semi-localized by adiabatic se- in dual transmit driven by 2 × 4 kW amplifiers, in combination with a
lective refocusing) sequence enables 1H-MRS of the human brain at 7 T 32-channel receive coil (both Nova Medical, Inc., Burlington, MA, USA).
with increased localization accuracy, SNR and reliability (Boer et al., MR acquisition took approximately 50 min per participant.
2011; Marsman et al., submitted). Despite the increased SNR, the mea- For anatomical reference and gray and white matter tissue classifica-
surement of GABA is not straightforward because the concentration of tion a T1-weighted magnetization prepared rapid gradient echo (MP-
GABA is low compared to other brain metabolites and the GABA signal RAGE) sequence was obtained (450 slices, slice thickness = 0.8 mm,
is obscured by other overlapping metabolites with higher signal intensity. TR = 7 ms, TE = 3 ms, flip angle = 8°, FOV = 250 × 200 × 180 mm,
To overcome this problem, spectral editing techniques can be applied to 312 × 312 acquisition matrix, SENSE factor 2.7, scan duration = 408 s).
isolate the GABA signal (Puts and Edden, 2012). In this study we used For the assessment of glutamate and the major spectral components
the sLASER sequence combined with editing techniques (MEGA-sLASER) NAA, creatine and choline, 1H-MRS experiments were conducted using
which allows for accurate and time-efficient detection of GABA a sLASER sequence (semi-localized by adiabatic selective refocusing;
(Andreychenko et al., 2012). TE = 28 ms, 32 averages, TR = 5 s, acquisition time = 2 min 40 s,
The purpose of this study was to measure GABA and glutamate levels 4 kHz acquisition bandwidth, 2048 datapoints) (Fig. 1). Voxels
in patients with schizophrenia and healthy control subjects at ultra- (2 × 2 × 2 cm3) were located in the medial prefrontal and medial
high field (7 T). We collected data from the medial prefrontal region, parieto-occipital lobes (Fig. 2). Non-water-suppressed spectra were ob-
which is known to show structural abnormalities in schizophrenia tained for quantification (carrier frequency was set to the chemical shift
(Harrison, 1999), and the medial parieto-occipital region as a control of H2O, acquisition time = 10 s).
area. Based on prior MRS studies of medicated schizophrenia patients, GABA-edited 1H-MRS experiments were conducted using a MEGA-
and structural MRI studies of schizophrenia patients which suggest sLASER sequence (TE = 74 ms, 64 averages, TR = 4 s, acquisition
alterations in the frontal brain region and only to a limited extent in time = 4 min 16 s, 4 kHz acquisition bandwidth, 2048 datapoints)
the occipital brain region (Harrison, 1999; Haijma et al., 2013), we hy- (Fig. 1). Voxels (2.5 × 2.5 × 2.5 cm3) were located in the medial frontal
pothesized that frontal GABA and glutamate levels are reduced, and and medial parieto-occipital regions (Fig. 2). GABA-editing with
400 A. Marsman et al. / NeuroImage: Clinical 6 (2014) 398–407

Table 1
Demographic characteristics.

Characteristic Patients Controls Statistic (p-value based on t, χ2)

Sex, M/F, N 13/4 16/7 0.45

Age, mean (SD), min–max, years 27.6 (6.1), 20.6–41. 7 27.7 (5.3), 21.5–40. 9 0.97
Years of education, mean (SD)
Participants 13.3 (1.9) 14.1 (2.1) 0.08
Parents 14.3 (3.0) 15.0 (1.3) 0.61
Total IQ, mean (SD) 92.8 (14.5) 108.3 (13.0) 0.002
Duration of illness, mean (SD), min–max, months 77.4 (82.1), 1–213
PANSS, mean (SD)
Total 53.1 (12.7)
Negative 12.9 (4.5)
Positive 12.7 (5.1)
General 27.5 (7.3)
Antipsychotic medication
Current daily dosage in haloperidol equivalentsa, N, mean (SD)
Typical
Clozapine 7, 4.26 (2.15)
Atypical 14, 5.48 (2.92)
Olanzapine 8, 6.84 (2.79)
Risperidone 1, 5.56
Aripiprazole 4, 4.04 (1.41)
Sulpiride 1, 0.29
Antipsychotic medication
Cumulative dosage in haloperidol equivalentsb, N, mean (SD)
Typical 5, 572 (483)
Clozapine 8, 6031 (6846)
Atypical 16, 3903 (4003)
Olanzapine 15, 2647 (3904)
Risperidone 5, 1934 (3219)
Aripiprazole 7, 1590 (1418)
Quetiapine 4, 445 (638)
Benzodiazepines current N (%), cumulative N (%) 6 (35%), 11 (65%)

The significance of bold values indicates level was set at 0.05.

a
4 patients were using two or more types of antipsychotic medication at the time of the study.
b
13 patients had used two or more types of antipsychotic medication during their lifespan.

Fig. 1. Typical metabolite spectra and fits.

 A. Marsman et al. / NeuroImage: Clinical 6 (2014) 398–407 401

macromolecule suppression was performed by shifting the editing frontal sLASER spectra, linewidth of the parieto-occipital spectra and
pulse between 1.9 (edit-on) and 1.4 ppm (edit-off) in subsequent CRLB of frontal NAA (Table 2). High SNRs, low CRLBs and small linewidths
scans (Andreychenko et al., 2012). After frequency and phase alignment were found, hence overall quality of the obtained spectra can be regarded
the subtraction of edit-on and edit-off scans yielded a GABA spectrum as excellent.
where the summation of all scans resulted in a non-edited spec-
trum which was used for the determination of the creatine peak 2.3. Spectral fitting and quantification
(Andreychenko et al., 2012). Prior to the MRS exams, second order B0
shimming was applied using the FASTERMAP algorithm at the voxel Retrospective phase and frequency alignment was performed on all
of interest (Gruetter, 1993; Gruetter and Boesch, 1992). Second, at data sets of each measurement (Waddell et al., 2007). Fitting of the
this location, a high B1 field was generated to minimize chemical shift sLASER spectra was performed with LCModel-based software imple-
displacement artifacts (Versluis et al., 2010). The highest possible B1 mented in Matlab (De Graaf, 1999), which uses a priori knowledge of
field was generated by optimizing the phase of both transmit channels the spectral components to fit metabolite resonances (Govindaraju
to locally assure constructive B1 interferences (Boer et al., 2011). et al., 2000). A fifth-order polynomial baseline was used. The follow-
Spectral quality measures, i.e. signal-to-noise ratio (SNR), Cramèr–Rao ing 16 metabolites and a measured macromolecular baseline (Behar
lower bounds (CRLBs) of the major spectral components, and linewidth, et al., 1994) were fitted to the spectra: acetate, aspartate, choline
were calculated and addressed for differences between groups and (Cho), phosphorylcholine (PC), glycerophosphorylcholine (GPC),
voxel locations. The SNR of the MEGA-sLASER was even higher than re- phosphorylethanolamine (PE), creatine (Cr), phosphocreatine (PCr),
ported in the initial paper for this sequence, which also showed standard N-acetyl aspartate (NAA), N-acetyl aspartyl glutamate (NAAG), GABA,
deviations of only 7% for GABA/Cr ratios (Andreychenko et al., 2012). Glu, glutamine (Gln), glutathione (GSH), myo-inositol (mIns), and
Small significant differences between groups were found for SNR of the taurine (Tau). Levels of Glu, total NAA (NAA + NAAG), total creatine

Fig. 2. Voxel placement. A: Medial prefrontal sLASER voxel: (A) sagittal view, (B) axial view, and (C) coronal view. Medial prefrontal MEGA-sLASER voxel: (D) sagittal view, (E) axial view,
and (F) coronal view. Medial parieto-occipital sLASER voxel: (G) sagittal view, (H) axial view, and (I) coronal view. Medial parieto-occipital MEGA-sLASER voxel: (J) sagittal view, (K) axial
view, and (L) coronal view.
402 A. Marsman et al. / NeuroImage: Clinical 6 (2014) 398–407

(Cr + PCr) and total choline (Cho + PC + GPC + PE) were estimated 2.4. Statistical analysis
using the water signal as an internal reference and calculated as follows
(Gasparovic et al., 2009): Statistical analyses were performed using SPSS 21.0 (2012, Chicago,
! IL). Demographic characteristics were addressed for differences between
signalmet
ðvolGM½water GM þvolWM½water WM þvolCS F½water pure Þ
½met ¼ signalwater
volGMþvolWM
the groups using Student3s t-tests and χ2-tests. Data were checked for
normality of their distributions. No transformations were required on
where [met] is the metabolite concentration, signalmet is the fitted any of the data. Multiple univariate analyses of variance were done with
signal intensity of the metabolite, accounting for the number of protons, metabolite level as dependent variable and with group (schizophrenia
and signalwater is the fitted signal intensity of water, accounting for the patients, healthy controls), age, sex and gray and white matter fractions
number of protons; volGM, volWM and volCSF are respectively the in the voxel as independent variables. In the case of a main significant ef-
gray matter fraction, white matter fraction and cerebrospinal fluid fect for group, age-by-group, sex-by-group, intelligence, and intelligence-
(CSF) fraction in the voxel; and [waterGM], [waterWM], and [waterpure] by-group were added separately to the model to assess interaction effects.
are respectively the water concentration in gray matter, white matter In patients, possible dependencies of metabolite concentrations
or CSF. For determining the contribution of gray matter, white matter depended on clinical symptomatology, or on current and cumulative an-
and CSF of each voxel, the software package SPM8 was used to segment tipsychotic medication intake were measured through post-hoc analyses.
the T1-weighted image. In the T1-weighted image, the position of the
1
H-MRS voxel was determined, after which the amount of gray matter,
white matter and CSF in the 1H-MRS voxel was computed. To account 3. Results
for differences in transverse relaxation between water and metabolites,
3.1. GABA
a correction was applied based on reported T2 values at 7 T of 47 ms on
average for water and 107 ms assumed for the metabolites (Marjańska
A significant main effect of group on the GABA/Cr ratio was found in
et al., 2012). Statistical analysis of the gray and white matter fractions in
the medial prefrontal cortex (F(1,24) = 7.33, p = 0.012), with the pa-
the frontal and parieto-occipital MEGA-sLASER (GABA/Cr) and sLASER
tients having lower GABA/Cr ratios as compared to the healthy controls
(glutamate, NAA, creatine, choline) revealed correlations N0.95 for
(Table 3; Fig. 3A). This effect remained significant when corrections for
both gray and white matter fractions in the two voxels.
gray and white matter fractions were removed from the analyses (thus
Fitting of the MEGA-sLASER spectra was performed by frequency-
adding the GABA/Cr ratio data from an additional two healthy subjects,
domain fitting of the GABA and creatine resonances to a Lorentzian
with F(1,28) = 4.91, p = 0.035). There were no significant effects of age
line-shape function in Matlab. GABA levels were expressed as the ratios
or age-by-group interaction on the GABA/Cr ratio in the prefrontal
of their peak areas relative to the peak areas of the creatine resonance.
cortex.
Because of poor quality of the T1-weighted MP-RAGE sequence, in
GABA/Cr ratio did not reveal significant differences between patients
two healthy subjects the gray and white matter separation could not
and controls in the parieto-occipital cortex (F(1,24) = 2.35, p = n.s.)
be estimated reliably in the frontal cortex, and in two healthy subjects
(Fig. 3B). A significant main effect of age on the GABA/Cr ratio in the
the parieto-occipital cortex. For analyses with correction for gray and
parieto-occipital cortex was found (F(1,24) = 7.88, p = 0.01), with
white matter fractions these subjects were not included in the analyses.
older individuals having lower ratios, irrespective of disease.
Because of poor spectral quality as established by a Cramér–Rao
In the parieto-occipital cortex a significant correlation was
lower bound (CRLB) of more than 20% and visual inspection, some
found between the GABA/Cr ratio and glutamate level in controls
data were excluded from the study. Frontal sLASER spectra were exclud-
(r(10) = −0.56, p = 0.03), with higher GABA/Cr ratios being associated
ed for three healthy controls and two patients; parieto-occipital sLASER
with lower levels of glutamate, an association not reaching significance
spectra were excluded for three healthy subjects and one patient; fron-
inpatients (r(9) = −0.25, p = n.s.).
tal MEGA-sLASER spectra were excluded for four healthy subjects and
There were no significant effects for sex on GABA/Cr ratios in the
three patients; and parieto-occipital MEGA-sLASER spectra were ex-
frontal and parieto-occipital cortices.
cluded for four healthy subjects and one patient. Spectra were mainly
excluded because of poor shimming or technical issues. CRLBs of Glu,
NAA, Cr and Cho never exceeded 20% and CRLBs of GABA/Cr exceeded 3.2. Glutamate
20% in only two instances.
Thus, frontal MRS results are based on 18 healthy subjects and 14 pa- In the medial prefrontal cortex these was no significant main effect of
tients and parieto-occipital MRS results are based on 17 healthy subjects group on glutamate level (F(1,26) = 0.97, p = n.s.) (Table 3; Fig. 3C).
and 15 patients. Frontal GABA-edited MRS results are based on 19 healthy There were significant main effects of age and sex on glutamate levels,
subjects and 13 patients and parieto-occipital GABA-edited MRS results with older individuals having lower levels than younger individuals
are based on 19 healthy subjects and 15 patients. (F(1,26) = 11.16, p = 0.003), and with higher glutamate levels in

Table 2
Spectral quality (mean (SD)).

Medial prefrontal cortex Medial parieto-occipital cortex

Patients Controls Statistic (p-value based on t, χ2) Patients Controls Statistic (p-value based on t, χ2)
a
SNR sLASER 51.9 (15.2) 64.8 (19.7) 0.046 50.5 (12.9) 61.0 (22.8) n.s.
SNR MEGA-sLASERa 9.0 (3.2) 11.5 (5.2) n.s. 10.7 (4.0) 9.0 (3.4) n.s.
Linewidth (Hz) 9.0 (1.2) 8.4 (1.4) n.s. 9.5 (1.1)a 8.5 (0.9) 0.009
CRLB (%) GABA/Cr 3.6 (1.0) 3.1 (1.6) n.s. 3.2 (1.8) 3.5 (1.6) n.s.
CRLB (%) Glu 2.7 (0.8) 2.2 (0.9) n.s. 2.8 (1.1) 2.7 (1.6) n.s.
CRLB (%) NAA 2.3 (0.6) 1.8 (0.7) 0.049 2.2 (1.0) 1.9 (0.8) n.s.
CRLB (%) Cr 1.2 (0.3) 1.0 (0.3) n.s. 1.3 (0.5) 1.3 (0.6) n.s.
CRLB (%) Cho 2.0 (0.6) 1.7 (0.5) n.s. 3.2 (1.4) 3.4 (2.8) n.s.

The significance of bold values indicates level was set at 0.05.

a
Signal-to-noise ratios are based on non-apodized spectra.
 A. Marsman et al. / NeuroImage: Clinical 6 (2014) 398–407 403

Table 3
Metabolite levels and gray and white matter fractions (mean (SD))a.

Medial prefrontal cortex Medial parieto-occipital cortex

Patients Controls Statistic (F, p) Patients Controls Statistic (F, p)

GABA/Cr ratio 0.12 (0.02) 0.14 (0.03) 7.33, 0.012 0.15 (0.04) 0.14 (0.03) 2.59, n.s.
Glutamate (mM) 8.48 (1.34) 8.65 (1.14) 0.97, n.s. 8.45 (1.07) 8.48 (1.26) 0.00, n.s.
NAA (mM) 9.36 (1.23) 9.85 (1.39) 1.04, n.s. 10.35 (1.26) 10.85 (1.13) 2.67, n.s.
Creatine (mM) 8.35 (1.45) 8.26 (1.43) 0.08, n.s. 8.48 (1.24) 8.28 (1.25) 0.13, n.s.
Choline (mM) 5.02 (1.96) 5.14 (1.51) 0.49, n.s. 3.48 (1.12) 3.35 (1.23) 0.00, n.s.
GM (%) sLASER 72.1 (10.3) 71.4 (13.6) 0.47, n.s. 72.1 (7.4) 70.1 (12.1) 1.79, n.s.
WM (%) sLASER 12.9 (8.8) 19.5 (14.9) 2.29, n.s. 22.1 (6.8) 25.2 (13.9) 4.37, n.s.
GM (%) MEGA-sLASER 68.4 (9.5) 68.1 (10.8) 0.06, n.s. 69.2 (7.8) 68.0 (10.4) 0.06, n.s.
WM (%) MEGA-sLASER 18.5 (7.0) 24.0 (11.2) 1.46, n.s. 24.5 (7.4) 27.7 (12.2) 1.20, n.s.

The significance of bold values indicates level was set at 0.05.

a
Data are based on two separate 1H-MRS measurements each done in two brain areas (prefrontal and parieto-occipital). A MEGA-sLASER sequence was performed for assessment of
GABA/Cr ratios and successfully completed in 13 patients and 19 controls in the medial prefrontal cortex and in 15 patients and 19 controls in the medial parieto-occipital cortex. A sLASER
sequence was performed for assessment of glutamate, NAA, creatine, and choline levels (corrected for gray and white matter fractions in the voxel) and successfully completed in 14 pa-
tients and 18 controls in the medial prefrontal cortex and in 15 patients and 17 controls in the medial parieto-occipital cortex.

females as compared to males (F(1,26) = 4.64, p = 0.04), irrespec- 3.3. Other metabolites
tive of disease.
Glutamate did not show any significant effects for group, age, and No significant findings were revealed for any of the other metabolites,
sex on glutamate level in the parieto-occipital cortex (Fig. 3D). except for significant effects of age on prefrontal NAA levels (F(1,26) =

Fig. 3. GABA/Cr ratios and glutamate concentrations in healthy controls and patients with schizophrenia. The blue (controls) and red (patients) bars indicate group averages. (A) Patients
show significantly lower medial prefrontal GABA/Cr ratios as compared to healthy controls (F(1,24) = 7.33, p = 0.012), when correcting for age, sex, and gray and white matter fractions.
(B) There is no significant difference in medial parieto-occipital GABA/Cr ratio between patients and controls (F(1,24) = 2.35, p = n.s.). (C) There is no significant difference in medial
prefrontal glutamate concentration between patients and controls (F(1,26) = 0.97, p = n.s.). (D) There is no significant difference in medial parieto-occipital glutamate concentration
between patients and controls (F(1,26) = 0.002, p = n.s.).
404 A. Marsman et al. / NeuroImage: Clinical 6 (2014) 398–407

7.22, p = 0.01) and prefrontal creatine levels (F(1,26) = 5.21, p = 0.03), added to the analyses as a covariate. Cannabis was used in the past
with lower levels being found in older individuals, irrespective of dis- month by 2 (9%) patients and were not used by any of the healthy sub-
ease; and for a significant effect of sex on parieto-occipital NAA levels jects. Cannabis use did not alter the findings when added as a covariate
with higher levels in females than in males (F(1,26) = 6.18, p = 0.02), in the significant analyses.
irrespective of disease.
4. Discussion
3.4. Associations with intelligence
To our knowledge, this study presents the first proton magnetic reso-
GABA/Cr ratio showed a significant interaction effect of intelligence- nance spectroscopy (MRS) measurements of GABA and glutamate levels
by-group in the prefrontal cortex (F(1,20) = 4.77, p = 0.04), when in- in vivo in the brains of patients with schizophrenia at a magnetic field
telligence and intelligence-by-group interaction were added to the strength of 7 T. The main finding is a significantly lower GABA/Cr ratio
model. This effect remained significant when corrections for gray and in the prefrontal cortex in patients with schizophrenia as compared to
white matter fractions were removed from the analyses (adding the healthy controls. In addition, the lower prefrontal GABA/Cr ratios were as-
GABA/Cr ratio data from an additional one healthy subject) with sociated with higher levels of general cognitive functioning in the pa-
F(1,23) = 4.79, p = 0.039. This interaction effect was due to pa- tients. No significant difference in the GABA/Cr ratio was found
tients with a higher intelligence level having a lower GABA/Cr ratio between patients and controls in the parieto-occipital cortex. Also, no sig-
(r = −0.92, p b 0.001) (Fig. 4), whereas in healthy individuals no signif- nificant differences in glutamate, NAA, creatine, and choline, were found
icant association between intelligence level and GABA/Cr was found. In in patients and controls in the prefrontal and parieto-occipital cortices.
patients, the correlation between higher level of intelligence and lower The main finding of this study is a significantly lower GABA/Cr ratio
GABA/Cr ratio was found for all aspects of intelligence but was most in the prefrontal cortex in patients as compared to healthy controls
prominently reflected in performance IQ (r = −0.89, p b 0.001), followed based on a GABA dedicated MRS sequence operating at 7 T. Our finding
by working memory index (r = −0.76, p = 0.02), verbal IQ (r = −0.74, is consistent with postmortem studies in schizophrenia, suggesting
p = 0.02), perceptual reasoning index (r = −0.72, p = 0.03), and verbal diminished GABA production based on decreased levels of mRNA
comprehension index (r = −0.70, p = 0.04). encoding for GAD67 (Olney et al., 1999; Kondziella et al., 2007; Stone
et al., 2009; Lisman et al., 2008) — the enzyme that facilitates GABA syn-
3.5. Associations with symptomatology, medication intake and substance thesis from glutamate, decreased levels of mRNA encoding for GAT-1
use (De la Fuente-Sandoval et al., 2011; Goto et al., 2009) – the transporter
that removes GABA from the synaptic cleft – and increased expression
Post-hoc analyses revealed that parieto-occipital glutamate of the α2 subunits of the GABAA receptor — the major inhibitory recep-
(r = −0.85, p = 0.003), parieto-occipital NAA (r = −0.84, p = 0.005), tor in the brain mediating most of the physiological actions of GABA
and parieto-occipital choline (r = −0.79, p = 0.012) levels were nega- (Sieghart et al., 1999). Moreover, the finding of lower GABA/Cr ratios in
tively correlated with the severity of negative symptoms. patients agrees with the hypothesis that NMDA-receptor hypofunction
No significant associations between any of the metabolite levels in reduces the excitation of GABAergic interneurons (Kondziella et al.,
the prefrontal and parieto-occipital cortices with current or cumulative 2007; Stone et al., 2009). However, we found an inverse correlation be-
dosage of classical or atypical antipsychotic medication were found. tween GABA/Cr ratios and glutamate levels in the parieto-occipital cortex
Also, no significant associations were found between metabolite levels in healthy controls but not in patients. Considering the tight coupling of
and benzodiazepine use. GABA and glutamate through GAD, this is unexpected and contrasts
Smoking status, alcohol use in the past month, and cannabis use in with previous studies reporting positive correlations between GABA and
the past month could be retrieved in 34 individuals (85%). 7 (41%) pa- glutamate in patients in the anterior cingulate and parieto-occipital corti-
tients and 2 (9%) healthy subjects were smokers. Smoking did not ces (Ongür et al., 2010) and in a mixed sample of patients and healthy
alter the findings when added as a covariate to the significant analyses. controls in the prefrontal cortex (Kegeles et al., 2012).
7 (41%) patients and 13 (57%) healthy subjects consumed alcohol in the Interestingly, we find the decrease in GABA/Cr ratios in patients with
past month. Alcohol use did not alter the significant findings when schizophrenia to be associated with a higher level of general cognitive
functioning. In patients, the association between higher general level
of intelligence and lower GABA/Cr ratio was found for all aspects of in-
telligence. Since we find an association between GABA and intelligence
in the patients and not in the healthy controls, it could reflect a compen-
satory mechanism to (continue) functioning at an (above) average
level. When separated out for several aspects of IQ, the decreased
GABA/Cr ratio in patients was associated with performance IQ, followed
by higher working memory functioning, perceptual reasoning, and ver-
bal comprehension (all ≥0.70). The medial prefrontal cortex serves as a
hub in (performance) IQ as part of the brain3s functional network (Van
den Heuvel et al., 2009). Moreover, this brain area revealed disruption
of brain connectivity in schizophrenia, as one of the brain3s hub areas
potentially leading to a reduced communication capacity and altered
functional brain dynamics (Van den Heuvel et al., 2013). As observed
in many earlier studies, we also found a lower intelligence level in pa-
tients with schizophrenia, which possibly exists prior to the disease
onset (Hedman et al., 2013) and is associated with genetic risk for
schizophrenia (Toulopoulou et al., 2007), thus confirming contributions
from the medial prefrontal cortex to both schizophrenia and intelli-
gence. However, the inverse correlation between GABA and intelligence
Fig. 4. Medial prefrontal GABA/Cr ratios in patients with schizophrenia significantly
contrasts with findings that GABA deficits occur as part of the impair-
decrease with increasing total IQ, when correcting for age, sex, and gray and white matter ment of schizophrenia (Enomoto et al., 2011; Menzies et al., 2007). An
fractions. (r = −0.91, p = 0.001). alternative explanation might be that patients with a higher IQ are
 A. Marsman et al. / NeuroImage: Clinical 6 (2014) 398–407 405

more adherent to treatment, which may lower GABA/Cr ratios to a may be associated with age-dependent neuronal pruning (Marsman
greater extent. Although there were no correlations found between et al., 2013b; Segovia et al., 2001). Lower NAA levels with increasing
GABA/Cr ratios and medication intake in this study, a previous study re- age have been reported before (Angelie et al., 2001; Raininko and
ports an inverse correlation between GABA levels and antipsychotic Mattsson, 2010; Sailasuta et al., 2008) and may reflect a reduction of
medication intake (Tayoshi et al., 2010). brain volume (Raininko and Mattsson, 2010) or neuronal function. A
We found no significant changes in glutamate levels in patients with lower creatine level with increasing age may be a sign of maturation
schizophrenia as compared to healthy controls based on 1H-MRS at 7 T. of neural systems in the brain (Steen et al., 2005). More likely, the
This is consistent with a recent large study in which no changes in glu- decreasing levels reflect different aspects of neural system integrity
tamate levels were found in schizophrenia (Bustillo et al., 2014). A re- and energy. However, there have been reports on increased creatine
cent meta-analysis showed an overall reduction of frontal glutamate and choline levels with age in healthy subjects and increased Gln/Glu
levels in patients as compared to healthy controls (Fonnum, 1984). ratios and choline levels in schizophrenia patients (Bustillo et al.,
However, in that study it was also suggested that glutamate levels are 2014; Maudsley et al., 2009). Thus, overall there were significant
increased in the early stage of the disease and begin to decrease below lower levels of all metabolites with older age, particularly in the pre-
control levels around the age of 25 years when most patients have frontal and not in the parieto-occipital cortex, with the exception of
been ill for a few years. Moreover, another meta-analysis suggested GABA which had lower levels in the parieto-occipital but not prefrontal
that glutamatergic levels are elevated in medication-naïve and medica- cortex with increasing age. We found no significant differences in the
tion free patients (Poels et al., 2014). Since the average age of the schizo- extent of decline with age in patients with schizophrenia as com-
phrenia patients in this study is 27.6 years with an average illness pared to controls, although some levels (NAA) did seem to decline
duration of 77 months (Table 2), it is possible that glutamate levels in disproportionally in the patients.
these patients are approaching control levels because of their disease Irrespective of disease, significant differences between the sexes
stage. Other levels of brain metabolites, including those of prefrontal were found in prefrontal glutamate and parieto-occipital NAA levels,
and parieto-occipital NAA, a marker of neuronal integrity, and of crea- with males having lower levels as compared to females. We did not
tine, a marker of energy metabolism, did not differ significantly be- find any differential findings for males and females in patients as com-
tween patients and healthy controls. Although meta-analyses showed pared to controls. Thus, GABA/Cr differences between patients and con-
reduced NAA levels in patients with schizophrenia in the frontal lobe trols could not be explained by differences between the sexes.
(Steen et al., 2005), most of the studies that were included in these Although there were no significant associations found between
meta-analyses were in patients who had been ill for many years medication use and metabolite levels, a limitation of this study is the
(Hulshoff Pol and Kahn, 2008; Harrison, 1999; Steen et al., 2005; lack of control for medication use during the lifespan. Also, this study
Brugger et al., 2011; Kraguljac et al., 2012). We did find that parieto- has a rather small sample size. Substance use, in particular smoking
occipital glutamate, NAA and choline were negatively correlated with (Gallinat et al., 2007), has been found to affect neurometabolite levels.
negative symptom severity. An inverse relationship between negative Therefore, in a post-hoc analysis we added smoking, alcohol and canna-
symptoms and NAA has been reported before (Brugger et al., 2011). bis intake as a covariate. Smoking, alcohol and cannabis intake did not
The correlation between glutamate and negative symptoms agrees alter the findings. In vivo MRS measurements have some limitations to
with previous findings of correlations between negative symptoms and take into account. One, with MRS one cannot distinguish between intra-
measures of abnormal glutamatergic neurotransmission (Pilowsky et al., cellular and extracellular metabolite levels. Two, because of its low con-
2006; Szulc et al., 2005). Moreover, NAA and glutamatergic neurotrans- centration a large voxel size is needed to reliably and time-efficiently
mission are probably closely related (Reynolds and Harte, 2007). measure GABA. Hence the voxel contained both gray and white matter.
Antipsychotic medication may alter metabolite levels in the brain However, the gray and white matter fractions in the individual voxels
(McLoughlin et al., 2009); including reductions in GABA and Glx (the were controlled for in the analyses making a large influence on the find-
sum of glutamate and glutamine) levels in the medial prefrontal cortex ings unlikely. The findings did not alter considerably when leaving out
(Kegeles et al., 2012). However, antipsychotic use and metabolite levels the corrections for gray and white matter fractions from the analyses.
did not reveal significant associations in our study. The intake of benzo- Three, the creatine peak that is derived from the MEGA-sLASER se-
diazepines may increase GABA levels, since benzodiazepines are GABAA quence to calculate the GABA/Cr ratio, contains contributions of GABA
receptor agonists (Sieghart and Sperk, 2002) (but see Goddard et al., and macromolecules that were not accounted for since it is difficult to
2004 for a decrease in GABA levels by down-modulation of GAD func- determine the relative contribution of macromolecules. However, the
tion and gene expression). Two-thirds of the patients had used benzodi- contributions of GABA and macromolecules to the creatine peak are rel-
azepines since they had become ill and one-third of the patients were atively small. Unlike the water signal, the creatine signal is obtained at
currently taking benzodiazepines. However, we found no significant exactly the same time as the GABA signal, so in the presence of subtle
associations with cumulative or current intake, although we did observe instability artifacts. Also, because of the long TE in the MEGA-sLASER se-
that patients that were currently taking benzodiazepines had (non- quence, the water signal is partially decayed due to T2 relaxation. The
significantly) higher GABA/Cr ratios compared to those who did creatine signal is also partially decayed due to T2 relaxation, but this ef-
not. Thus, it seems unlikely that medication intake explained our find- fect is smaller compared to the effect T2 relaxation has on the water sig-
ing of decreased GABA/Cr ratio in the frontal cortex in patients. This is nal. Moreover, the T2 relaxation of water differs between white matter,
however in contrast with a previous study which found a negative cor- gray matter and CSF, thus it is more difficult to calculate a correction fac-
relation between frontal GABA and antipsychotic medication intake tor because fractions of white matter, gray matter and CSF differ among
(Tayoshi et al., 2010). individuals. For the data to be as accurate as possible, creatine was used
Irrespective of disease and with increasing age, we find significantly as a reference instead of water. Although there were no differences in
lower parieto-occipital GABA and prefrontal glutamate, NAA and crea- creatine levels found between patients and healthy controls in this
tine levels in the current 7 T 1H-MRS study. Parieto-occipital but not study, one study does report changes in creatine levels in schizophrenia
frontal GABA levels decreased with increasing age in these relatively (Ongür et al., 2009).
young adult individuals, whereas for all the other metabolites the fron- Despite the use of the sensitivity optimized sLASER technique, in
tal but not parieto-occipital levels were lower with older age. Possibly some of the spectra glutamine and NAAG, which are important compo-
high levels of the inhibitory neurotransmitter GABA are still important nents in glutamatergic metabolism, could not be assessed accurately.
for development and plasticity of the prefrontal cortex in young adult- While the relatively short scan time could have been increased to im-
hood (Ben-Ari et al., 2007). Decreases in glutamate level with age are prove SNR, our current study shows that in 9% of the spectra the CRLB
consistent with previous studies (Marsman et al., 2013a;2013b) and for glutamine was above 20% and 53% the CRLB for NAAG was above 20%.
406 A. Marsman et al. / NeuroImage: Clinical 6 (2014) 398–407

In conclusion, using 1H-MRS at 7 T, prefrontal GABA/Cr ratios appear Lewis, D.A., Moghaddam, B., 2006. Cognitive dysfunction in schizophrenia: convergence
of {gamma}-aminobutyric acid and glutamate alterations. Archives of Neurology 63
to be decreased in medicated patients with schizophrenia as compared (10), 1372–1376. http://dx.doi.org/10.1001/archneur.63.10.137217030651.
to healthy controls, in contrast to previously reported elevations in un- Benes, F.M., Berretta, S., 2001. GABAergic interneurons: implications for understanding
medicated patients. Moreover, GABA/Cr ratios are associated with the schizophrenia and bipolar disorder. Neuropsychopharmacology: Official Publication
of the American College of Neuropsychopharmacology 25 (1), 1–27. http://dx.doi.
level of cognitive functioning, with high functioning patients having org/10.1016/S0893-133X(01)00225-111377916.
lower GABA/Cr ratios. This suggests a role for GABA in the earlier Lewis, D.A., Volk, D.W., Hashimoto, T., 2004. Selective alterations in prefrontal cortical
stages of the disease in schizophrenia, and replication studies are GABA neurotransmission in schizophrenia: a novel target for the treatment of work-
ing memory dysfunction. Psychopharmacology 174 (1), 143–150. http://dx.doi.org/
needed to assess the effects of antipsychotic medication use, aging, 10.1007/s00213-003-1673-x15205885.
and symptomatology. Coyle, J.T., 2004. The GABA–glutamate connection in schizophrenia: which is the proxi-
mate cause? Biochemical Pharmacology 68 (8), 1507–1514. http://dx.doi.org/10.
1016/j.bcp.2004.07.03415451393.
Conflict of interest De la Fuente-Sandoval, C., León-Ortiz, P., Favila, R., Stephano, S., Mamo, D., Ramírez-
Bermúdez, J., Graff-Guerrero, A., 2011. Higher levels of glutamate in the associative-
striatum of subjects with prodromal symptoms of schizophrenia and patients with
All authors declare that they have no conflicts of interest. first-episode psychosis. Neuropsychopharmacology: Official Publication of the
American College of Neuropsychopharmacology 36 (9), 1781–1791. http://dx.doi.
org/10.1038/npp.2011.6521508933.
Funding Goto, N., Yoshimura, R., Moriya, J., Kakeda, S., Ueda, N., Ikenouchi-Sugita, A., Umene-
Nakano, W., Hayashi, K., Oonari, N., Korogi, Y., 2009. Reduction of brain gamma-
aminobutyric acid (GABA) concentrations in early-stage schizophrenia patients: 3 T
This research was funded by the Netherlands Organisation for Scien- Proton MRS study. Schizophrenia Research 112 (1–3), 192–193. http://dx.doi.org/
tific Research (NWO) VIDI Grant 917-46-370 (to H.H.); and Utrecht Uni- 10.1016/j.schres.2009.04.02619464152.
Ongür, D., Prescot, A.P., McCarthy, J., Cohen, B.M., Renshaw, P.F., 2010. Elevated gamma-
versity High Potential Grant (to H.H.). aminobutyric acid levels in chronic schizophrenia. Biological Psychiatry 68 (7),
667–670. http://dx.doi.org/10.1016/j.biopsych.2010.05.01620598290.
Tayoshi, S., Nakataki, M., Sumitani, S., Taniguchi, K., Shibuya-Tayoshi, S., Numata, S., Iga, J.,
References Ueno, S., Harada, M., Ohmori, T., 2010. GABA concentration in schizophrenia patients
and the effects of antipsychotic medication: a proton magnetic resonance spectrosco-
Hulshoff Pol, H.E., Kahn, R.S., 2008. What happens after the first episode? A review of pro- py study. Schizophrenia Research 117 (1), 83–91. http://dx.doi.org/10.1016/j.schres.
gressive brain changes in chronically ill patients with schizophrenia. Schizophrenia 2009.11.01120022731.
Bulletin 34 (2), 354–366. http://dx.doi.org/10.1093/schbul/sbm16818283048. Yoon, J.H., Maddock, R.J., Rokem, A., Silver, M.A., Minzenberg, M.J., Ragland, J.D., Carter,
Harrison, P.J., 1999. The neuropathology of schizophrenia. A critical review of the data and C.S., 2010. GABA concentration is reduced in visual cortex in schizophrenia and cor-
their interpretation. Brain: A Journal of Neurology 122 (4), 593–624. http://dx.doi. relates with orientation-specific surround suppression. Journal of Neuroscience: the
org/10.1093/brain/122.4.59310219775. Official Journal of the Society for Neuroscience 30 (10), 3777–3781. http://dx.doi.
Harrison, P.J., Weinberger, D.R., 2005. Schizophrenia genes, gene expression, and neuro- org/10.1523/JNEUROSCI.6158-09.201020220012.
pathology: on the matter of their convergence. Molecular Psychiatry 10 (1), 40–68. Rowland, L.M., Kontson, K., West, J., Edden, R.A., Zhu, H., Wijtenburg, S.A., Holcomb, H.H.,
http://dx.doi.org/10.1038/sj.mp.400155815263907. Barker, P.B., 2013. In vivo measurements of glutamate, GABA, and NAAG in schizo-
Kraguljac, N.V., White, D.M., Reid, M.A., Lahti, A.C., 2013. Increased hippocampal glu- phrenia. Schizophrenia Bulletin 39 (5), 1096–1104. http://dx.doi.org/10.1093/
tamate and volumetric deficits in unmedicated patients with schizophrenia. JAMA schbul/sbs09223081992.
Psychiatry 70 (12), 1294–1302. http://dx.doi.org/10.1001/jamapsychiatry.2013. Tkác, I., Andersen, P., Adriany, G., Merkle, H., Ugurbil, K., Gruetter, R., 2001. In vivo 1H
243724108440. NMR spectroscopy of the human brain at 7 T. Magnetic Resonance in Medicine:
Fonnum, F., 1984. Glutamate: a neurotransmitter in mammalian brain. Journal of Neuro- Official Journal of the Society of Magnetic Resonance in Medicine / Society of
chemistry 42 (1), 1–11. http://dx.doi.org/10.1111/j.1471-4159.1984.tb09689. Magnetic Resonance in Medicine 46 (3), 451–456. http://dx.doi.org/10.1002/
x6139418. mrm.121311550235.
Marsman, A., van den Heuvel, M.P., Klomp, D.W.J., Kahn, R.S., Luijten, P.R., Hulshoff Pol, Boer, V.O., van Lier, A.L., Hoogduin, J.M., Wijnen, J.P., Luijten, P.R., Klomp, D.W., 2011. 7-T
H.E., 2013a. Glutamate in schizophrenia: a focused review and meta-analysis of 1H- (1) H MRS with adiabatic refocusing at short TE using radiofrequency focusing with a
MRS studies. Schizophrenia Bulletin 39 (1), 120–129. http://dx.doi.org/10.1093/ dual-channel volume transmit coil. NMR in Biomedicine 24 (9), 1038–1046. http://
schbul/sbr06921746807. dx.doi.org/10.1002/nbm.164121294206.
Poels, E.M.P., Kegeles, L.S., Kantrowitz, J.T., Javitt, D.C., Lieberman, J.A., Abi-Dargham, A., Andreychenko, A., Boer, V.O., Arteaga de Castro, C.S., Luijten, P.R., Klomp, D.W., 2012.
Girgis, R.R., 2014. Glutamatergic abnormalities in schizophrenia: a review of proton Efficient spectral editing at 7 T: GABA detection with MEGA-sLASER. Magnetic Reso-
MRS findings. Schizophrenia Research 152 (2-3), 325–332. http://dx.doi.org/10. nance in Medicine: Official Journal of the Society of Magnetic Resonance in Medicine /
1016/j.schres.2013.12.01324418122. Society of Magnetic Resonance in Medicine 68 (4), 1018–1025. http://dx.doi.org/10.
Moghaddam, B., Adams, B., Verma, A., Daly, D., 1997. Activation of glutamatergic neuro- 1002/mrm.2413122213204.
transmission by ketamine: a novel step in the pathway from NMDA receptor block- Marsman A., Boer V.O., Luijten P.R., Hulshoff Pol H.E., Klomp D.W.J., Mandl R.C.W. Detec-
ade to dopaminergic and cognitive disruptions associated with the prefrontal tion of glutamate alterations in the human brain using 1H-MRS with sLASER at 7 T
cortex. Journal of Neuroscience: the Official Journal of the Society for Neuroscience (submitted)
17 (8), 2921–29279092613. Puts, N.A., Edden, R.A., 2012. In vivo magnetic resonance spectroscopy of GABA: a meth-
De la Fuente-Sandoval, C., León-Ortiz, P., Azcárraga, M., Stephano, S., Favila, R., Díaz- odological review. Progress in Nuclear Magnetic Resonance Spectroscopy 60, 29–41.
Galvis, L., Alvarado-Alanis, P., Ramírez-Bermúdez, J., Graff-Guerrero, A., 2013. Gluta- http://dx.doi.org/10.1016/j.pnmrs.2011.06.00122293397.
mate levels in the associative striatum before and after 4 weeks of antipsychotic Haijma, S.V., Van Haren, N., Cahn, W., Koolschijn, P.C.M.P., Hulshoff Pol, H.E., Kahn, R.S.,
treatment in first-episode psychosis: a longitudinal proton magnetic resonance spec- 2013. Brain volumes in schizophrenia: a meta-analysis in over 18 000 subjects.
troscopy study. JAMA Psychiatry 70 (10), 1057–1066. http://dx.doi.org/10.1001/ Schizophrenia Bulletin 39 (5), 1129–1138. http://dx.doi.org/10.1093/schbul/
jamapsychiatry.2013.28923966023. sbs11823042112.
Kegeles, L.S., Mao, X., Stanford, A.D., Girgis, R., Ojeil, N., Xu, X., Gil, R., Slifstein, M., Abi- Andreasen, N.C., Flaum, M., Arndt, S., 1992. The comprehensive assessment of symptoms
Dargham, A., Lisanby, S.H., Shungu, D.C., 2012. Elevated prefrontal cortex {gamma}- and history (CASH). An instrument for assessing diagnosis and psychopathology. Ar-
aminobutyric acid and glutamate–glutamine levels in schizophrenia measured chives of General Psychiatry 49 (8), 615–623. http://dx.doi.org/10.1001/archpsyc.
in vivo with proton magnetic resonance spectroscopy. Archives of General Psychiatry 1992.018200800230041637251.
69 (5), 449–459. http://dx.doi.org/10.1001/archgenpsychiatry.2011.151922213769. Kay, S.R., Fiszbein, A., Opler, L.A., 1987. The positive and negative syndrome scale (PANSS)
Olney, J.W., Newcomer, J.W., Farber, N.B., 1999. NMDA receptor hypofunction model of for schizophrenia. Schizophrenia Bulletin 13 (2), 261–276. http://dx.doi.org/10.1093/
schizophrenia. Journal of Psychiatric Research 33 (6), 523–533. http://dx.doi.org/10. schbul/13.2.2613616518.
1016/S0022-3956(99)00029-110628529. Andreasen, N.C., Pressler, M., Nopoulos, P., Miller, D., Ho, B.C., 2010. Antipsychotic dose
Kondziella, D., Brenner, E., Eyjolfsson, E.M., Sonnewald, U., 2007. How do glial–neuronal equivalents and dose–years: a standardized method for comparing exposure to dif-
interactions fit into current neurotransmitter hypotheses of schizophrenia? Neuro- ferent drugs. Biological Psychiatry 67 (3), 255–262. http://dx.doi.org/10.1016/j.
chemistry International 50 (2), 291–301. http://dx.doi.org/10.1016/j.neuint.2006.09. biopsych.2009.08.04019897178.
00617084946. Wechsler D. Wechsler adult intelligence scale. 3rd ed (1997). Lisse, The Netherlands.
Stone, J.M., Day, F., Tsagaraki, H., Valli, I., McLean, M.A., Lythgoe, D.J., O3Gorman, R.L., [Dutch version (2005)].
Barker, G.J., McGuire, P.K., 2009. Glutamate dysfunction in people with prodromal Gruetter, R., 1993. Automatic, localized in vivo adjustment of all first- and second-order
symptoms of psychosis: relationship to gray matter volume. Biological Psychiatry shim coils. Magnetic Resonance in Medicine: Official Journal of the Society of Magnet-
66 (6), 533–539. http://dx.doi.org/10.1016/j.biopsych.2009.05.00619559402. ic Resonance in Medicine / Society of Magnetic Resonance in Medicine 29 (6),
Lisman, J.E., Coyle, J.T., Green, R.W., Javitt, D.C., Benes, F.M., Heckers, S., Grace, A.A., 2008. 804–811. http://dx.doi.org/10.1002/mrm.19102906138350724.
Circuit-based framework for understanding neurotransmitter and risk gene interac- Gruetter, R., Boesch, C., 1992. Fast, non-iterative shimming on spatially localized signals:
tions in schizophrenia. Trends in Neurosciences 31 (5), 234–242. http://dx.doi.org/ in vivo analysis of the magnetic field along axes. Journal of Magnetic Resonance 96,
10.1016/j.tins.2008.02.00518395805. 323–334.
 A. Marsman et al. / NeuroImage: Clinical 6 (2014) 398–407 407

Versluis, M.J., Kan, H.E., van Buchem, M.A., Webb, A.G., 2010. Improved signal to noise in pro- Biological Psychiatry 69 (5), 495–503. http://dx.doi.org/10.1016/j.biopsych.2010.10.
ton spectroscopy of the human calf muscle at 7 T using localized B1 calibration. Magnet- 00421145039.
ic Resonance in Medicine: Official Journal of the Society of Magnetic Resonance in Kraguljac, N.V., Reid, M., White, D., Jones, R., Den Hollander, J., Lowman, D., Lahti, A.C.,
Medicine / Society of Magnetic Resonance in Medicine 63 (1), 207–211. http://dx.doi. 2012. Neurometabolites in schizophrenia and bipolar disorder — a systematic review
org/10.1002/mrm.2219519918906. and meta-analysis. Psychiatry Research 203 (2-3), 111–125. http://dx.doi.org/10.
Waddell, K.W., Avison, M.J., Joers, J.M., Gore, J.C., 2007. A practical guide to robust detec- 1016/j.pscychresns.2012.02.00322981426.
tion of GABA in human brain by J-difference spectroscopy at 3 T using a standard vol- Pilowsky, L.S., Bressan, R.A., Stone, J.M., Erlandsson, K., Mulligan, R.S., Krystal, J.H., Ell, P.J.,
ume coil. Magnetic Resonance Imaging 25 (7), 1032–1038. http://dx.doi.org/10.1016/ 2006. First in vivo evidence of an NMDA receptor deficit in medication-free schizo-
j.mri.2006.11.02617707165. phrenic patients. Molecular Psychiatry 11 (2), 118–119. http://dx.doi.org/10.1038/
De Graaf, R.A., 1999. NMR Processing Software for Spectroscopy, Imaging and Spectro- sj.mp.400175116189506.
scopic Imaging. Szulc, A., Galinska, B., Tarasow, E., Dzienis, W., Kubas, B., Konarzewska, B., Walecki, J.,
Govindaraju, V., Young, K., Maudsley, A.A., 2000. Proton NMR chemical shifts and coupling Alathiaki, A.S., Czernikiewicz, A., 2005. The effect of risperidone on metabolite mea-
constants for brain metabolites. NMR in Biomedicine 13 (3), 129–15310861994. sures in the frontal lobe, temporal lobe, and thalamus in schizophrenic patients. A
Behar, K.L., Rothman, D.L., Spencer, D.D., Petroff, O.A., 1994. Analysis of macromole- proton magnetic resonance spectroscopy (1H MRS). Pharmacopsychiatry 38 (5),
cule resonances in 1H NMR spectra of human brain. Magnetic Resonance in Med- 214–219. http://dx.doi.org/10.1055/s-2005-87315616189748.
icine: Official Journal of the Society of Magnetic Resonance in Medicine / Society Reynolds, G.P., Harte, M.K., 2007. The neuronal pathology of schizophrenia: molecules
of Magnetic Resonance in Medicine 32 (3), 294–302. http://dx.doi.org/10.1002/ and mechanisms. Biochemical Society Transactions 35 (2), 433–436. http://dx.doi.
mrm.19103203047984061. org/10.1042/BST035043317371293.
Gasparovic, C., Yeo, R., Mannell, M., Ling, J., Elgie, R., Phillips, J., Doezema, D., Mayer, A.R., McLoughlin, G.A., Ma, D., Tsang, T.M., Jones, D.N., Cilia, J., Hill, M.D., Robbins, M.J.,
2009. Neurometabolite concentrations in gray and white matter in mild traumatic Benzel, I.M., Maycox, P.R., Holmes, E., Bahn, S., 2009. Analyzing the effects of psy-
brain injury: an 1H-magnetic resonance spectroscopy study. Journal of Neurotrauma chotropic drugs on metabolite profiles in rat brain using 1H NMR spectroscopy.
26 (10), 1635–1643. http://dx.doi.org/10.1089/neu.2009-089619355814. Journal of Proteome Research 8 (4), 1943–1952. http://dx.doi.org/10.1021/
Marjańska, M., Auerbach, E.J., Valabrègue, R., Van de Moortele, P.F., Adriany, G., Garwood, pr800892u19714815.
M., 2012. Localized 1H NMR spectroscopy in different regions of human brain in vivo Sieghart, W., Sperk, G., 2002. Subunit composition, distribution and function of
at 7 T: T2 relaxation times and concentrations of cerebral metabolites. NMR in Bio- GABA(A) receptor subtypes. Current Topics in Medicinal Chemistry 2 (8), 795–816.
medicine 25 (2), 332–339. http://dx.doi.org/10.1002/nbm.175421796710. http://dx.doi.org/10.2174/156802602339350712171572.
Sieghart, W., Fuchs, K., Tretter, V., Ebert, V., Jechlinger, M., Höger, H., Adamiker, D., 1999. Goddard, A.W., Mason, G.F., Appel, M., Rothman, D.L., Gueorguieva, R., Behar, K.L., Krystal,
Structure and subunit composition of GABA(A) receptors. Neurochemistry International J.H., 2004. Impaired GABA neuronal response to acute benzodiazepine administration
34 (5), 379–385. http://dx.doi.org/10.1016/S0197-0186(99)00045-510397365. in panic disorder. American Journal of Psychiatry 161 (12), 2186–2193. http://dx.doi.
Van den Heuvel, M.P., Stam, C.J., Kahn, R.S., Hulshoff Pol, H.E., 2009. Efficiency of function- org/10.1176/appi.ajp.161.12.218615569888.
al brain networks and intellectual performance. Journal of Neuroscience: the Official Ben-Ari, Y., Gaiarsa, J.L., Tyzio, R., Khazipov, R., 2007. GABA: a pioneer transmitter that ex-
Journal of the Society for Neuroscience 29 (23), 7619–7624. http://dx.doi.org/10. cites immature neurons and generates primitive oscillations. Physiological Reviews
1523/JNEUROSCI.1443-09.200919515930. 87 (4), 1215–1284. http://dx.doi.org/10.1152/physrev.00017.200617928584.
Van den Heuvel, M.P., Sporns, O., Collin, G., Scheewe, T., Mandl, R.C., Cahn, W., Goñi, J., Marsman, A., Mandl, R.C.W., Van den Heuvel, M.P., Boer, V.O., Wijnen, J.P., Klomp,
Hulshoff Pol, H.E., Kahn, R.S., 2013. Abnormal rich club organization and functional D.W.J., Luijten, P.R., Hilleke E, H.P., 2013b. Glutamate changes in healthy young
brain dynamics in schizophrenia. JAMA Psychiatry 70 (8), 783–792. http://dx.doi. adulthood. European Neuropsychopharmacology: the Journal of the European
org/10.1001/jamapsychiatry.2013.132823739835. College of Neuropsychopharmacology 23 (11), 1484–1490. http://dx.doi.org/
Hedman, A.M., van Haren, N.E., van Baal, C.G., Kahn, R.S., Hulshoff Pol, H.E., 2013. IQ 10.1016/j.euroneuro.2012.11.00323245833.
change over time in schizophrenia and healthy individuals: a meta-analysis. Schizo- Segovia, G., Porras, A., Del Arco, A., Mora, F., 2001. Glutamatergic neurotransmission in
phrenia Research 146 (1–3), 201–208. http://dx.doi.org/10.1016/j.schres.2013.01. aging: a critical perspective. Mechanisms of Ageing and Development 122 (1),
02723490758. 1–29. http://dx.doi.org/10.1016/S0047-6374(00)00225-611163621.
Toulopoulou, T., Picchioni, M., Rijsdijk, F., Hua-Hall, M., Ettinger, U., Sham, P., Murray, R., Angelie, E., Bonmartin, A., Boudraa, A., Gonnaud, P.M., Mallet, J.J., Sappey-Marinier, D.,
2007. Substantial genetic overlap between neurocognition and schizophrenia: genet- 2001. Regional differences and metabolic changes in normal aging of the human
ic modeling in twin samples. Archives of General Psychiatry 64 (12), 1348–1355. brain: proton MR spectroscopic imaging study. AJNR. American Journal of Neuroradi-
http://dx.doi.org/10.1001/archpsyc.64.12.134818056542. ology 22 (1), 119–12711158897.
Enomoto, T., Tse, M.T., Floresco, S.B., 2011. Reducing prefrontal gamma-aminobutyric acid Raininko, R., Mattsson, P., 2010. Metabolite concentrations in supraventricular white mat-
activity induces cognitive, behavioral, and dopaminergic abnormalities that resemble ter from teenage to early old age: a short echo time 1H magnetic resonance spectros-
schizophrenia. Biological Psychiatry 69 (5), 432–441. http://dx.doi.org/10.1016/j. copy (MRS) study. Acta Radiologica (Stockholm, Sweden: 1987) 51 (3), 309–315.
biopsych.2010.09.03821146155. http://dx.doi.org/10.3109/0284185090347656420170295.
Menzies, L., Ooi, C., Kamath, S., Suckling, J., McKenna, P., Fletcher, P., Bullmore, E., Stephenson, Sailasuta, N., Ernst, T., Chang, L., 2008. Regional variations and the effects of age and gen-
C., 2007. Effects of gamma-aminobutyric acid-modulating drugs on working memory der on glutamate concentrations in the human brain. Magnetic Resonance Imaging
and brain function in patients with schizophrenia. Archives of General Psychiatry 64 26 (5), 667–675. http://dx.doi.org/10.1016/j.mri.2007.06.00717692491.
(2), 156–167. http://dx.doi.org/10.1001/archpsyc.64.2.15617283283. Maudsley, A.A., Domenig, C., Govind, V., Darkazanli, A., Studholme, C., Arheart, K.,
Bustillo, J.R., Chen, H., Jones, T., Lemke, N., Abbott, C., Qualls, C., Canive, J., Gasparovic, C., Bloomer, C., 2009. Mapping of brain metabolite distributions by volumetric proton
2014. Increased glutamine in patients undergoing long-term treatment for schizo- MR spectroscopic imaging (MRSI). Magnetic Resonance in Medicine: Official Journal
phrenia: a proton magnetic resonance spectroscopy study at 3 T. JAMA Psychiatry of the Society of Magnetic Resonance in Medicine / Society of Magnetic Resonance in
71 (3), 265–272. http://dx.doi.org/10.1001/jamapsychiatry.2013.393924402128. Medicine 61 (3), 548–559. http://dx.doi.org/10.1002/mrm.2187519111009.
Steen, R.G., Hamer, R.M., Lieberman, J.A., 2005. Measurement of brain metabolites by 1H Gallinat, J., Lang, U.E., Jacobsen, L.K., Bajbouj, M., Kalus, P., Von Haebler, D., Seifert, F.,
magnetic resonance spectroscopy in patients with schizophrenia: a systematic re- Schubert, F., 2007. Abnormal hippocampal neurochemistry in smokers: evidence
view and meta-analysis. Neuropsychopharmacology: Official Publication of the from proton magnetic resonance spectroscopy at 3 T. Journal of Clinical Psychophar-
American College of Neuropsychopharmacology 30 (11), 1949–1962. http://dx.doi. macology 27 (1), 80–84. http://dx.doi.org/10.1097/JCP.0b013e31802dffde17224719.
org/10.1038/sj.npp.130085016123764. Ongür, D., Prescot, A.P., Jensen, J.E., Cohen, B.M., Renshaw, P.F., 2009. Creatine abnormal-
Brugger, S., Davis, J.M., Leucht, S., Stone, J.M., 2011. Proton magnetic resonance spectros- ities in schizophrenia and bipolar disorder. Psychiatry Research 172 (1), 44–48.
copy and illness stage in schizophrenia — a systematic review and meta-analysis. http://dx.doi.org/10.1016/j.pscychresns.2008.06.00219239984.