Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Haley, R. W.
Right arrow Articles by Fleckenstein, J. L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Haley, R. W.
Right arrow Articles by Fleckenstein, J. L.
(Radiology. 2000;215:807-817.)
© RSNA, 2000

Neuroradiology

Brain Abnormalities in Gulf War Syndrome: Evaluation with 1H MR Spectroscopy1

Robert W. Haley, MD, W. Wesley Marshall, MD, George G. McDonald, PhD, Mark A. Daugherty, RT, Frederick Petty, PhD, MD and James L. Fleckenstein, MD

1 From the Depts of Internal Medicine, Section of Epidemiology (R.W.H., W.W.M.), Radiology (G.G.M., M.A.D., J.L.F.), and Psychiatry (F.P.), University of Texas Southwestern Medical Center at Dallas, 5323 Harry Hines Blvd, Dallas, TX 75390-8874, and Dallas Veterans Affairs Medical Center, Tex (F.P.). From the 1999 RSNA scientific assembly. Received Nov 8, 1999; revision requested Jan 7, 2000; revision received Jan 28; accepted Mar 13. Supported by the U.S. Army Medical Research and Materiel Command under cooperative agreement no. DAMD17-97-2-7025; by U.S. Public Health Service grant M01-RR00633; by a grant from the Perot Foundation; and by a grant from Philips Medical Systems of North America. Address correspondence to R.W.H.


   Abstract
TOP
 Abstract
Introduction
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 APPENDIX
 References
 
PURPOSE: To test for neuronal brain damage in the basal ganglia and brainstem in Gulf War veterans by using magnetic resonance (MR) spectroscopy.

MATERIALS AND METHODS: Twenty-two Gulf War veterans with one of three factor analysis–derived syndromes (case patients); 18 well veterans matched for age, sex, and education level (control subjects); and six Gulf War veterans with syndrome 2 from a different population (replication sample) underwent long echo time (272 msec) proton (hydrogen 1) MR spectroscopy on a 4 x 2 x 2-cm voxel in the basal ganglia bilaterally and a 2 x 2 x 2-cm voxel in the pons. Syndromes 1–3 are described as "impaired cognition," "confusion-ataxia," and "central pain," respectively.

RESULTS: The N-acetylaspartate–to-creatine (NAA/Cr) ratio, which reflects functional neuronal mass, was significantly lower in the basal ganglia and brainstem of Gulf War veterans with the three syndromes than in those structures of the control subjects (P = .007). The finding was corroborated in the replication sample (P = .002). Veterans with syndrome 2 (the most severe clinically) had evidence of decreased NAA/Cr in both the basal ganglia and the brainstem; those with syndrome 1, in the basal ganglia only; and those with syndrome 3, in the brainstem only.

CONCLUSION: Veterans with different Gulf War syndromes have biochemical evidence of neuronal damage in different distributions in the basal ganglia and brainstem.

Index terms: Basal ganglia, 142.891 • Basal ganglia, MR, 142.12145 • Brain, diseases, 142.891, 1538.891 • Brain, MR, 142.12145, 1538.12145 • Brain stem, abnormalities, 1538.891 • Brain stem, MR, 1538.12145 • Epidemiology • Magnetic resonance (MR), spectroscopy, 142.12145, 1538.12145


   Introduction
TOP
 Abstract
 Introduction
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 APPENDIX
 References
 
In 1997, Haley et al (1–3) presented clinical data suggesting a neuropathologic basis for three apparently distinct Gulf War syndromes: Syndrome 1 is described as "impaired cognition;" syndrome 2, "confusion-ataxia;" and syndrome 3, "central pain." The strength of those data included the observation of differences in brainstem evoked potentials and results of audiovestibular tests of ocular pursuit and saccadic eye movement between veterans with the syndromes and matched control subjects (2,4). These objective clinical findings also differentiated the three syndromes by severity, the veterans with syndrome 2 being the most severely affected and those with syndromes 1 and 3 being only moderately affected (1–4). That these brain abnormalities may be related to chemical exposure during the Gulf War was supported by strong epidemiologic associations (relative risks of 4 to 8) between the three primary syndromes and risk factors for wartime exposures to combinations of low-level organophosphate chemical nerve agents, pesticides, pyridostigmine, and diethyltoluamide (DEET)-containing insect repellants (3).

Not only was the elevated exposure risk to these potential neurotoxins documented in the symptomatic veterans, but a biochemical explanation for heightened susceptibility to these chemicals was demonstrated in the same group of Gulf War veterans (5). Specifically, the symptomatic veterans had substantially lower blood levels of paraoxonase-1 (PON1) type Q (PON-Q) arylesterase than did the control subjects; PON-Q is a genetically controlled isoenzyme that hydrolyzes organophosphate chemical warfare nerve agents and some pesticides (5). That this genetic polymorphism may predispose to abnormalities of deep brain structures was suggested by an association (relative risk of 1.6) between the same polymorphic form of this enzyme and Parkinson disease (6), in which degeneration of the substantia nigra and basal ganglia is well recognized. These observations build upon preexistent data that indicate repetitive low-level exposure to certain organophosphates affects the neurochemistry of the basal ganglia (7–10). Also, the symptoms of Gulf War veterans (1,4) are similar to those reported by patients in the early stages of well-understood degenerative diseases of the basal ganglia (11). Hence, a potential role for neurotoxin-mediated injury of deep brain structures in genetically susceptible Gulf War veterans merits careful consideration.

Previous brain magnetic resonance (MR) imaging data that failed to demonstrate visible changes in ill Gulf War veterans (2) could be construed as refuting the hypothesis of neurotoxic brain disease. That "negative finding," however, does not imply normality of the brain in these patients, because MR imaging measures brain structure with a resolution at the millimeter level. In fact, the literature is replete with reports (12–41) of biochemical and physiologic brain abnormalities that are not consistently and reliably demonstrated at MR imaging because they do not produce demonstrable structural changes. The imperfect sensitivity of MR imaging for the detection of clinically important brain disease has contributed to the growing use of MR spectroscopy to probe underlying intracellular biochemical abnormalities, particularly in brains that appear normal at MR imaging (12–15). Examples include temporal lobe epilepsy (16), dementias (17–24), hepatic encephalopathy (25), multiple sclerosis (26–30), and a variety of psychiatric diseases (15,31–41). Documentation of the greater sensitivity of MR spectroscopy compared to MR imaging coincides in time with an increased propensity for medical insurance companies to reimburse for MR spectroscopic examinations in certain brain diseases (42).

From the similarity of the symptoms of ill Gulf War veterans with the early symptoms of patients with primary diseases of basal ganglia (11) and from the demonstrated abnormalities on audiovestibular tests in these veterans (2,4), we suspected that biochemical abnormalities exist in the deep brain structures, such as the basal ganglia and brainstem, of at least some veterans with Gulf War syndrome. In view of the absence of MR imaging–visible abnormalities, we performed MR spectroscopy in ill Gulf War veterans and a matched control group to test for biochemical evidence of neuronal brain damage in the basal ganglia and brainstem. From the previously shown differences in the clinical severity of the three syndromes (1,2,4), we hypothesized that brain abnormalities would be the most pronounced in syndrome 2 and of intermediate severity in syndromes 1 and 3.


   MATERIALS AND METHODS
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 APPENDIX
 References
 
Subjects
Between December 1997 and June 1998, proton (hydrogen 1) MR spectroscopy was performed in the following groups of subjects from the Twenty-fourth Reserve Naval Mobile Construction Battalion (commonly known as "Seabees") who participated in the clinical case-control study reported previously by Haley et al (2): 22 Gulf War veterans with one of the three primary Gulf War syndromes (five with syndrome 1, 12 with syndrome 2, and five with syndrome 3) and 18 control subjects who were group-matched to the syndrome 2 case patients by age, sex, and education and who served in the same military unit during the Gulf War but have remained well. Details of the selection and comparability of the case patients and control subjects are given in a prior publication (2).

The nature of the three Gulf War syndromes (1), levels of a genetically determined enzyme that protects against organophosphate poisoning (5), neuropsychologic and audiovestibular evidence of brain injury (2,4,43), levels of occupational impairment (1), and epidemiologic risk factors (3) found in prior studies are summarized in Table 1. These clinical correlations led us to hypothesize that the severity of brain injury would be most pronounced in patients with syndrome 2 and of intermediate severity in patients with syndromes 1 and 3 compared with the normal brains of the control subjects. We were also aware that neurotoxic injuries could affect one side of the brain more than the other (45) or that injuries in one hemisphere might cause different symptoms from injuries in the other hemisphere (46–48). This led us to include a test (an interaction term) for different magnitudes of case-control differences in the two hemispheres.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Summary of Clinical and Epidemiologic Findings Leading to the Hypothesis of Different Levels of Severity of Brain Injury of Three Primary Gulf War Neurologic Syndromes Identified by Factor Analysis of Symptoms in Members of a Naval Reserve Battalion Who Served in the Gulf War
 
Nine of the control subjects had been deployed to the Southwest Asian Theater of Operations during the Gulf War (deployed controls), and nine had not (nondeployed controls). Of the 23 ill veterans and 20 matched control subjects in the original study (2), one ill veteran, who had also served in the Vietnam War, was unable to participate due to the development of multiple myeloma for which he underwent bone marrow transplantation; two of the original 20 control subjects, one deployed and the other nondeployed, declined to participate. All subjects gave informed consent after the nature of the procedures had been fully explained, and the research protocol was approved by the university's institutional review board.

All personnel (W.W.M., G.G.M., M.A.D., F.P., J.L.F.) involved in acquiring the images or processing the imaging data were blinded to the subjects' case- or control-group status. Approximately 1 week before MR spectroscopy, the subjects discontinued all medications that would interfere with the tests. No subjects were taking medications reported to affect MR spectroscopic results, such as choline, lithium, or cocaine.

All subjects underwent general medical and psychiatric evaluations, as well as routine clinical laboratory tests and T1-and T2-weighted MR imaging of the brain. Depression was the only medical condition that was more common in the ill veterans (13 of 22 patients) than in the controls (three of 18 subjects). Clinical laboratory tests showed no values outside normal limits for aspartate aminotransferase, alanine aminotransferase, alkaline phosphatase, and serum albumin, except for a slightly elevated alanine aminotransferase in one deployed control subject with no history of alcohol abuse or dependance. No abnormalities that could explain the symptoms were identified on brain MR images interpreted by a neuroradiologist (J.L.F.) who was blinded to the subjects' case- or control-group status (2).

To obtain preliminary evidence on whether the study's findings might be applicable to the wider population of ill Gulf War veterans, we surveyed 336 Gulf War veterans from our local Department of Veterans Affairs medical center by using the same questionnaires and factor analysis methods employed by Haley and colleagues (1). Our confirmatory factor analysis found that Haley's three-factor model fit our data well (findings to be reported in another publication). From this independent veteran population, we randomly selected six Gulf War veterans with Haley syndrome 2, the most severe clinically, to undergo MR spectroscopy as a replication sample.

All case and control subjects were men. The control subjects (mean age at the time of MR spectroscopy, 50.4 years ± 6.4 [SD]; age range, 41–66 years) were closely matched with the syndrome 2 case patients (mean, 52.2 years ± 6.7; range, 43–62 years) and with the syndrome 3 case patients (mean, 53.0 years ± 6.7; range 49–65 years) on age but were substantially older than the syndrome 1 case patients (mean, 38.6 years ± 7.5; range, 30–49 years) and the syndrome 2 patients in the replication sample (mean, 37.5 years ± 8.2; range, 27–48 years). Although the absolute signal intensities (concentrations) of the three metabolites N-acetylaspartate (NAA), choline (Cho), and creatine (Cr) measured with MR spectroscopy may decline with age in the basal ganglia, the metabolite ratios remain constant with age (49,50). Consequently, when comparing the metabolite ratios, little bias should result from the age differences between the comparison groups.

MR Spectroscopy Protocol
Proton (1H) MR spectroscopy is a noninvasive, easily tolerated radiologic technique used to explore brain chemistry in living individuals (12–15). By measuring intracellular concentrations of protons, this technique estimates the concentrations of the specific abundant brain chemicals in small volumes of brain. Among these, NAA, found only in neurons, serves as a marker of neuronal cell body and axon mass; Cho, a cell membrane component, primarily reflects brain glial mass and inflammation; and Cr, a uniformly distributed component of energy metabolism usually unaffected by pathologic processes, serves primarily as a reference chemical (51).

Data were acquired on a Gyroscan NT MR imaging—MR spectroscopy scanner operating at 1.5 T (Philips Medical Systems, Best, the Netherlands), with use of a 30-cm-diameter head coil for both excitation and reception of the 1H MR signal. The gradient system had maximum gradient strength of 23 mT/m, rise time of 200 µsec, slew rate of 105 mT/m/msec, and 100% duty cycle. MR "scout" images were acquired in the transverse, sagittal, and coronal planes to ensure that the signal acquisition volumes were placed in comparable locations in all subjects.

After standard shimming and water signal suppression were performed and four dummy scans run to attain steady state equilibrium of the nuclear magnetization, spectra were acquired (1,800/272 [repetition time msec/echo time msec]; number of signals acquired, 256) in three single-voxel views: 4 x 2 x 2-cm voxels in the basal ganglia bilaterally centered on the putamen and including parts of the head of the caudate nucleus and globus pallidus, and a single 2 x 2 x 2-cm voxel in the pons (Fig 1). Each image required approximately 8 minutes after setup, rough shim, gradient tuning, fine shim, and water suppression, amounting to approximately 60 minutes for the complete study. All images were obtained by the same experienced operator (M.A.D.), who was blinded to the subjects' case- or control-group status.



View larger version (132K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1a. Regions localized during MR spectroscopic evaluation. (a) Coronal and (b) transverse T1-weighted scout images show the position chosen for the left basal ganglia (box). A similar location was chosen for the right basal ganglia (not shown). (c) Coronal and (d) sagittal T1-weighted scout images show the position chosen for the pons (box). Parameters for a-d are repetition time, 20 msec; echo time, 5 msec; flip angle, 60°.

 


View larger version (124K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1b. Regions localized during MR spectroscopic evaluation. (a) Coronal and (b) transverse T1-weighted scout images show the position chosen for the left basal ganglia (box). A similar location was chosen for the right basal ganglia (not shown). (c) Coronal and (d) sagittal T1-weighted scout images show the position chosen for the pons (box). Parameters for a-d are repetition time, 20 msec; echo time, 5 msec; flip angle, 60°.

 


View larger version (126K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1c. Regions localized during MR spectroscopic evaluation. (a) Coronal and (b) transverse T1-weighted scout images show the position chosen for the left basal ganglia (box). A similar location was chosen for the right basal ganglia (not shown). (c) Coronal and (d) sagittal T1-weighted scout images show the position chosen for the pons (box). Parameters for a-d are repetition time, 20 msec; echo time, 5 msec; flip angle, 60°.

 


View larger version (143K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1d. Regions localized during MR spectroscopic evaluation. (a) Coronal and (b) transverse T1-weighted scout images show the position chosen for the left basal ganglia (box). A similar location was chosen for the right basal ganglia (not shown). (c) Coronal and (d) sagittal T1-weighted scout images show the position chosen for the pons (box). Parameters for a-d are repetition time, 20 msec; echo time, 5 msec; flip angle, 60°.

 
Metabolite Quantification
The postprocessing of the spectra was automated (52) and performed in the time domain by an analyst (G.G.M.) who was blinded to the subjects' case- or control-group status. Signals from the repetitive image iterations were summed by data point and exported to a SPARC workstation (SunMicrosystems, Mountain View, Calif) for analysis.

Hankle single value deconvolution, or HSVD, filtering was used to remove residual water signal in the time domain (53). Then signal intensities (metabolite concentration estimates) of NAA, Cho, and Cr in each volume of interest were estimated with the FITMASTERS (Fast Interpretation of Time Domain Data by Multi-component Analysis of Selectively Truncated Exponential Resonance Signals) program (Phillips Medical Systems), which automatically fit separate damped sine waves to the NAA, Cho, and Cr peaks of the acquired spectra in the time domain to estimate the integral (amplitude in the time domain), frequency, and T2 time constant (53). Chemical shifts were calculated relative to the NAA peak at 2.01 ppm. For model fitting, the frequencies of the peaks for NAA, Cho, and Cr were fixed at 2.01, 3.019, and 3.202 ppm, respectively, and their 1/T2 values were constrained to vary together at 8.84, 5.86, and 4.05 Hz, respectively. Intensity fitting was unconstrained.

The automatic FITMASTERS program was able to fit and make signal intensity estimates for all three peaks on 135 (98%) of the 138 single-voxel images. Of the 46 participating subjects (22 case patients, 18 control subjects, and six replication case patients), signal intensity estimates for all three metabolites were obtained from the left basal ganglia in 46, from the right basal ganglia in 44, and from the pons in 45. The missing metabolite measurements were due to technical issues that prevented separation of the Cho and Cr peaks. The missing measurements involved one veteran with syndrome 3 and one deployed control subject, each having incomplete estimates from the right basal ganglia only, and one deployed control subject with incomplete estimates from the pons.

Display of Spectra
The Hankle single value deconvolution– filtered spectra, the fitted model, and the difference (model minus data), generated in the time domain, were processed with 1-Hz exponential filter, Fourier transformation, and phase correction with Xunspec software (Philips Medical Systems) for presentation in the frequency domain (Fig 2). Visual inspection of the three nonfitting spectra showed entirely overlapping signal from the adjacent Cho and Cr peaks, which precluded their being individually fit.



View larger version (19K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2. Representative 1H MR spectra (echo time, 272 msec) from the right basal ganglia of a Gulf War veteran with syndrome 2 (right) compared with that from a well control veteran of the same age (left). The areas under the three major peaks indicate the signal intensities (concentrations) of Cho, Cr, and NAA in the brain volume of interest. The model spectra (top), mathematically fit to the original spectra (bottom), are scaled to equal heights of the Cr peaks to facilitate comparison of the NAA/Cr and Cho/Cr ratios of the two subjects. The ill veteran (top right) has greatly reduced neuronal mass (NAA/Cr ratio) and glial mass (Cho/Cr ratio) compared with those ratios in the control subject (top left).

 
Statistical Methods
Since metabolite measurements from proton MR spectroscopy are not generally considered to be absolute quantitative measures, the relative signal intensities of NAA and Cho were expressed for group comparisons by their ratios with Cr (ie, NAA/Cr and Cho/Cr ratios) (12–15). The primary objective was to test the difference in the NAA/Cr ratio between each of the three syndrome groups and the control group in the right and left basal ganglia and in the pons, with the prior expectation, based on the previously demonstrated relative severity of the three syndromes (Table 1) (1–4), that the NAA/Cr ratio would be highest in the control subjects, intermediate in the Seabees with syndromes 1 and 3, and lowest in those with syndrome 2.

To minimize the risk of falsely positive (type I) errors from testing multiple hypotheses while avoiding falsely negative (type II) errors from excessively conservative Bonferroni corrections, a single family-wise (global) test of the hypothesis (54) was first performed with a univariate repeated-measures analysis of variance (ANOVA) under the generalized linear model (55). This analysis tested the hypothesis that the mean NAA/Cr ratio would be highest in the control group, intermediate in patients with syndromes 1 and 3, and lowest in patients with syndrome 2 in all three brain regions (right and left basal ganglia and pons), while controlling for differences in the NAA/Cr ratio among the three brain regions. (In statistical terminology of a repeated-measures analysis, the test of differences among the syndrome and control groups is the between-subjects effect, and the test among anatomic regions is the within-subject effect [55].) An interaction term was included to test whether the magnitude of the difference in the NAA/Cr ratio across the syndrome and control groups varied significantly by brain region.

After rejecting the family-wise null hypothesis and demonstrating a difference by syndrome and control groups, we attempted to replicate the finding by performing a confirmatory repeated-measures ANOVA testing for a difference in the NAA/Cr ratio between the six syndrome 2 case patients in the replication sample and the original control group, while controlling for differences by anatomic region and the interaction as before.

After firmly rejecting the global null hypothesis in both the Seabees sample and the replication group (54), detailed analyses were performed to describe the NAA/Cr differences between each syndrome group and the control group in each of the brain regions. First, the differences in the left and right basal ganglia were tested simultaneously with a univariate repeated-measures ANOVA (55), testing the difference between each syndrome group and the control group (between-subjects effect) while controlling for differences by hemisphere (within-subject effect) and testing for different magnitudes of case-control difference in the two hemispheres (hemisphere-by-group interaction). Second, the NAA/Cr differences between each syndrome group and the control group were tested separately in each brain region with the Student t test or the Mann-Whitney U test (with continuity correction) (56), depending on whether the normality assumption was satisfied. The same analyses on the Cho/Cr ratio were done to assist in interpreting the primary differences in the NAA/Cr ratio.

Before each significance test, the distributions of the ratios in each group of case patients and the control group were tested for normality by the Shapiro-Wilk normality test (57). If either distribution being compared failed the normality test (P < .2), the log-transformed distributions were tested for normality (57). When either distribution in the analysis could not be normalized by transformation, the nonparametric Mann-Whitney U test (with continuity correction) (56) was reported. All distributions of the NAA/Cr ratio satisfied the normality test; wherever the nonparametric test was required for testing the Cho/Cr ratio, it is indicated by a footnote.

To address the question of whether the NAA/Cr ratio or the absolute Cr signal intensity constituted the primary difference between the case patients and the control subjects, two multivariate logistic regression analyses were performed to determine which metabolite measure more strongly discriminated each syndrome group from the control group. In the first, the residual effect of the NAA/Cr ratio was assessed after controlling for Cr; in the second, the residual effect of Cr was assessed after controlling for the NAA/Cr ratio. These analyses were repeated for the left and right basal ganglia and the pons.

Since in adults NAA (neuronal mass) in brain tissue can only be reduced by disease (51,58), a one-tailed test of significance was used when analyzing it; however, since Cho can be either increased or reduced by disease (51), a two-tailed test was used for it. In these analyses, P values ranging from greater than .05 to .10 were regarded as indicating marginal significance, while P values less than or equal to .05 were regarded as indicating a significant difference.

Analyses were performed with the General Linear Models (GLM), Univariate Statistics (UNIVARIATE), and Nonparametric (NPAR1WAY) procedures of SAS software (SAS Institute, Cary, NC). Since MR spectroscopic protocols and imaging equipment are rapidly evolving, there are no stable normal ranges for the metabolite ratios, making comparisons with well-matched, contemporaneously imaged control subjects the most useful for research.


   RESULTS
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 APPENDIX
 References
 
Global Hypothesis Test
In the repeated-measures ANOVA, analyzing all three brain regions simultaneously, the NAA/Cr ratio was found to be lowest in the syndrome 2 group, intermediate in the syndrome 1 and 3 groups, and highest in the control group as hypothesized (P = .007), while controlling for the differences between brain region and the group-by-region interaction. In the test of the region effect, the NAA/Cr level was higher in the pons than in the basal ganglia across case patients and control subjects (P < .001). The group-by-region interaction, however, was not statistically significant (P = .32), indicating that the three-level difference in the NAA/Cr ratio across the syndrome and control groups appeared to be of similar magnitude in all three brain regions. Individual chemical spectra illustrating the reduction in the NAA/Cr ratio are shown in Figure 2.

In the confirmatory analysis, the difference in the NAA/Cr ratios between the syndrome 2 cases in the replication sample and the control group was also statistically significant irrespective of brain region (P = .002). As before, the NAA/Cr ratio varied significantly by brain region, being higher in the pons than the basal ganglia (P < .001), and the magnitude of the difference in NAA/Cr ratio between the syndrome 2 group and the control group did not appear to differ by brain region (P = .12).

Basal Ganglia
In detailed analysis of the basal ganglia, the NAA/Cr ratio was significantly lower in veterans with syndrome 2 than in the control subjects (group effect P < .001 by repeated-measures ANOVA), and this finding was corroborated in the veterans with syndrome 2 in the replication sample (group effect P = .005) (Fig 3, Appendix Table A1). The difference was larger and more highly statistically significant in the right basal ganglia (18% difference from controls, P < .001) than the left (9% difference, P = .09), but in the repeated-measures ANOVA the evidence for a bilateral case-control difference (group effect P < .001) was stronger than that for a unilateral difference (hemisphere-by-group interaction, P = .19) (Fig 3, Appendix Table A1).



View larger version (40K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3. Mean (± SEM) metabolite ratios in the basal ganglia and pons of veterans with Gulf War syndromes 1-3 (Syn 1-Syn 3) and a replication sample of veterans with syndrome 2 from an independent veteran population (Syn 2 R), compared with an age-sex-education-matched group of well veterans (Controls). The NAA/Cr ratios in the basal ganglia (upper left) are significantly lower than that in the control group for the Gulf War veteran sample with syndrome 2 and for the veterans of the replication sample with syndrome 2, and there is a similar, though nonsignificant, trend for those with syndrome 1. The NAA/Cr ratios in the pons (upper right) are significantly lower than that in the control group for the veterans with syndromes 2 and 3 and in the replication sample with syndrome 2, but not in those with syndrome 1. The Cho/Cr ratios in basal ganglia (lower left) are lower than that in the control group for syndrome 1 and the replication sample but not for syndromes 2 and 3. The Cho/Cr ratios in the pons (lower right) tended to be lower than that in the control group for all syndrome groups, but none was statistically significant. For the basal ganglia, the P value is for a difference from the control group in both left and right basal ganglia (group effect), controlling for hemisphere effect and the group-by-hemisphere interaction by repeated-measures ANOVA. For the pons, the P value came from a t test of the difference from the control group.

 
The data also suggested a slight, though nonsignificant, bilateral reduction of the NAA/Cr ratio in the basal ganglia of the veterans with syndrome 1 (3% difference in both hemispheres, group effect P = .19), but there was little evidence for a reduction in those with syndrome 3 (group effect, P = .33) (Fig 3, Appendix Table A1). In both basal ganglia, the mean NAA/Cr ratio was lower in the combined group of all Seabees with any of the three Gulf War syndromes than that in the control group (5% difference on the left and 12% on the right, group effect P = .01) (Appendix Table A1).

By repeated-measures ANOVA, the Cho/Cr ratio in the basal ganglia was significantly lower than that in the control subjects only for syndrome 1 patients (group effect P = .02) and for the replication sample (group effect P = .03) (Fig 3, Appendix Table A2). None of the Cho/Cr ratios was significantly higher in ill Gulf War veterans than in the control subjects.

Pons
In the pons where the variance of the metabolite ratios was greater, the mean NAA/Cr ratio tended to be lower in the combined group of all Seabees with any of the three Gulf War syndromes than that in the control group (20% overall difference, P = .10) (Appendix Table A3). The NAA/Cr ratio was substantially lower in veterans with syndrome 2 than in the control subjects (26% difference, P = .026), and this finding was corroborated in the replication sample of veterans with syndrome 2 (28% difference, P = .05) (Fig 3, Appendix Table A3). There was also a large, possibly significant, reduction of the NAA/Cr ratio in the veterans with syndrome 3 (24% difference, P = .09), but there was no evidence for a reduction in those with syndrome 1 (P = .36) (Fig 3, Appendix Table A3).

A tendency for the Cho/Cr ratio in the pons to be lower in the ill Gulf War veterans was most evident for syndromes 2 and 3 and in the replication sample, but none of these differences was statistically significant (Fig 3, Appendix Table A3). None of the mean Cho/Cr ratios was higher in the pons of ill Gulf War veterans than that in the control subjects.

Decreased NAA/Cr Ratio or Increased Cr Concentration?
The finding that both the NAA/Cr and Cho/Cr ratios were decreased in case patients compared with control subjects is compatible with either a reduction in the relative signal intensities of NAA and Cho or with an increase in that of Cr. That the NAA/Cr reduction was found predominantly in syndrome 2 and the Cho/Cr reduction was predominantly in syndrome 1 argues against a generalized increase in creatine. Moreover, in a series of multivariate logistic regression analyses predicting case versus control status with the metabolite ratios and Cr, case status was found to be far more strongly associated with reductions in the metabolite ratios than with variations in Cr levels. For example, in a logistic regression analysis predicting syndrome 2 versus control status in the right basal ganglia, after controlling for the Cr concentration the NAA/Cr ratio remained strongly associated (P < .001); whereas, after controlling for the NAA/Cr ratio the Cr concentration was not significantly associated (P = .5). Results were similar for the other brain areas and for the Cho/Cr ratio (data not shown).


   DISCUSSION
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
APPENDIX
 References
 
The results of this study, in which brain biochemical measurements obtained with MR spectroscopy were compared in ill veterans and healthy control subjects, confirm a reduction of functioning neuronal mass in the basal ganglia and brainstem of greatest severity in Gulf War syndrome 2 ("confusion-ataxia") and of intermediate severity in syndromes 1 ("impaired cognition") and 3 ("central pain"). Measurements with MR spectroscopy demonstrated an 18% reduction in NAA/Cr in the right basal ganglia and a 26% reduction in the brainstem of the veterans with Haley syndrome 2 (1), changes that are similar in magnitude to those seen in other brain diseases (18,20,22,26,27,59–64). Under the MR spectroscopic pulsing conditions that we employed (long echo time), resonance signals for three primary metabolites—NAA, Cho and Cr—are expected in normal brain, as described in Materials and Methods. Although the exact physiologic role of NAA is unknown, it is abundant in neurons and axons throughout the nervous system (65,66), being synthesized in neuronal mitochondria and transported to the cytosol (58,67). For group comparisons, it is commonly expressed as a ratio relative to the Cr signal. When abnormal, the NAA/Cr ratio is virtually always reduced, reflecting a nonspecific, generalized reduction of functioning neuronal mass (18,20,22,26,27,59–64). Although nonspecific, it is a very sensitive measure, frequently revealing abnormalities in tissue that appears normal on conventional MR imaging studies.

The finding of reduced neuronal mass in the basal ganglia and brainstem has at least three important implications for understanding Gulf War syndrome. It provides a plausible anatomic explanation for the previous clinical, genetic, and epidemiologic findings that is consistent with well-understood diseases of deep brain structures. It completes a chain of evidence linking Gulf War syndrome with injury from wartime organophosphate exposure. And, it highlights MR spectroscopy as a potentially useful biomarker in the diagnosis of Gulf War syndrome.

The first important implication is that the degree of reduction in neuronal mass in the different brain regions provides a plausible anatomic explanation for prior clinical and epidemiologic findings (1–5,43). The largest and clearest reduction of neuronal mass (NAA/Cr ratio) in both basal ganglia and brainstem was found in veterans with Haley syndrome 2. The veterans in this group have been shown to have the most severe cognitive and vestibular symptoms (1,2,4); the greatest degree of occupational disability (1); the greatest neuropsychologic evidence of brain dysfunction (2); the worst abnormalities of visual pursuit, saccadic velocity, ocular response to rotation, and spontaneous nystagmus at rest (2,4); the most profound genetic predisposition to damage from chemical nerve agents (5); and, by history, the greatest likelihood of exposure to chemical nerve agents and the worst adverse reactions to the pyridostigmine anti–nerve gas tablets (3) (Table 1).

In contrast, the veterans with Haley syndrome 3 have been shown to have a reduction in neuronal mass of a similar magnitude, though of marginal statistical significance, in the brainstem but not in the basal ganglia. These veterans typically complain of intractable body pain and other sensory disturbances but less cognitive impairment (1), have less occupational disability (1), demonstrate a similarly high degree of neuropsychologic brain dysfunction (2), have the greatest audiovestibular slowing of ocular velocity in response to caloric stimulation but little or no abnormalities on the other audiovestibular tests (2,4), have an intermediate level of genetic predisposition to organophosphate injury (5), and, by history, used greater amounts of highly concentrated DEET insect repellants and also had severe adverse reactions to pyridostigmine tablets (3).

The veterans with Haley syndrome 1 had just the opposite pattern—a slight, nonsignificant reduction of neuronal mass in the basal ganglia but no abnormality in the brainstem. They complained of only mild cognitive impairment (1); had the least abnormality on neuropsychologic and audiovestibular tests (2,4), resembling syndrome 2 veterans only in patterns of ocular response to rotation and in abnormalities of saccadic velocity (2,4); had an intermediate degree of genetic predisposition (5); and, by history, were more likely to have worn pesticide-containing pet flea collars to repel insects (3). The lack of clear statistical significance of the trends for syndromes 1 and 3, possibly due to the small sample sizes and low statistical power, limits our confidence in the findings for these two groups. This problem is mitigated, however, by the fact that the earlier clinical and audiovestibular testing patterns (1,2,4,5) predicted the intermediate degrees of biochemical abnormality found at MR spectroscopy (Table 1).

Our rationale for focusing MR spectroscopy on the basal ganglia and brainstem was that the range of symptoms reported by these Gulf War veterans is consistent with the early presenting symptoms of well-understood degenerative diseases of the basal ganglia and brainstem. These include Huntington disease, which initially affects the caudate nucleus; Wilson hepatolenticular degeneration, which initially affects the putamen; and Fahr disease, which initially involves calcification of the globus pallidus (11). These classic neurodegenerative diseases often begin with personality changes; irritability; cognitive impairments particularly in executive function, memory, concentration, and attention; changes in speech; attacks of vertigo; abnormalities of ocular pursuit and saccadic eye movement; central pain; and mood disorders (11,68). Since these symptoms often begin before the development of objective neurologic findings, the diagnosis is often delayed until the inevitable appearance of objective signs (11,68–70). Veterans with Gulf War syndrome originally presented with symptoms that strongly suggested a primary disease of basal ganglia. Now we have demonstrated biochemical evidence of organic abnormality in the basal ganglia of ill Gulf War veterans. The fact that patients with Gulf War syndrome have generally not, up to now, progressed to the point of having objective neurologic signs, as those with Huntington, Wilson, or Fahr disease inevitably do, is consistent with their having experienced a time-limited neurotoxic chemical exposure in the war that resulted in a less severe, possibly nonprogressive basal ganglia injury.

In addition, the high frequency of vertigo in Gulf War veterans ill with these syndromes (1,2,4) and the abnormalities on the previously reported brainstem evoked potentials and audiovestibular tests (2,4) are also compatible with the biochemical evidence of damage to the basal ganglia and brainstem demonstrated at MR spectroscopy. The complex neural pathways of the vestibulo-ocular reflex traverse the brainstem and are controlled by important projections from the caudate nucleus, explaining why they can be affected by diseases of either the brainstem or the basal ganglia (71,72). Thus, the symptoms observed in Gulf War syndrome overlap with well-understood degenerative diseases of deep brain structures, and the degree of clinical impairment is proportional to the decrement in neuronal mass within these structures. Together, these findings make a compelling case for regional brain injury as the explanation for symptoms in some Gulf War veterans.

The second major implication of our study is that the biochemical injury we have found in the deep brain structures completes a chain of evidence linking Gulf War syndrome with injury from organophosphate exposure. The fact that large numbers of U.S. and coalition troops were exposed during the Gulf War to diverse, potentially neurotoxic chemicals, including low-level sarin, has been thoroughly documented (73–78). Only the numbers of troops exposed, the effective dose levels, and the long-term health consequences remain in question. Epidemiologic analysis has shown unusually strong statistical associations (relative risks of 4 to 8) between wartime exposures to different chemical combinations and unique Gulf War symptom complexes (3) (Table 1). The epidemiologically implicated chemicals were subsequently proved to cause permanent neurologic injury in laboratory animals when administered either alone (79,80) or in combinations (81,82). The human relevance of these associations was amplified by the discovery that having one of the three Haley Gulf War syndromes was strongly associated with low blood levels of the isomorphic Q type of the paraoxonase/arylesterase enzyme (PON-Q) but not with the R type (PON-R) (5). PON-Q has high hydrolytic activity against chemical nerve agents, such as sarin and soman; whereas, PON-R is most active against common pesticides such as parathion (83). Moreover, the level of the enzyme deficiency was associated with the severity of neurologic symptoms. Compared with that in the control subjects, the blood levels were most profoundly decreased in Gulf War veterans with syndrome 2 and moderately decreased in those with syndromes 1 and 3, fitting plausibly with the different degrees of clinical severity, occupational impairment, neuropsychologic and audiovestibular dysfunction, and the regional distributions of reduced neuronal mass (Table 1, Fig 3).

Whereas the foregoing evidence makes a strong circumstantial case for brain injury from chemical exposure, direct evidence of injury in brain regions known to be susceptible to organophosphate neurotoxic damage has to date been lacking. It is well documented, particularly in agricultural workers, that poisoning by organophosphate pesticides can cause a chronic psychiatric syndrome involving fatigue, cognitive disturbance, vertigo, central pain, depression, and even psychosis, but little or no peripheral neuropathy (84), similar to the symptoms of both Gulf War syndrome (1) and the early stages of primary basal ganglia diseases (11). When the same chemicals are given to laboratory animals repetitively, coincident with the chronic behavioral changes they suffer damage to the basal ganglia manifested by reductions in striatal cholinesterase, neurotoxic esterase, dopamine, {gamma}-aminobutyric acid (or GABA), and the number of cholinergic muscarinic receptors (7–10). Now our biochemical evidence of damage to the basal ganglia in genetically susceptible, organophosphate-exposed veterans links their disease to the brain region known to be the most affected by organophosphate neurotoxicity.

Other unlikely organic causes of symptoms in Gulf War syndrome that might be entertained run the gamut of inflammatory disease, trauma, ischemia, neoplasia, and psychiatric diseases such as major depressive disorder and bipolar disorder. Although there is no direct evidence in the literature linking these to Gulf War syndrome, the normal to low values for the Cho/Cr ratio found in this study impose a restriction on the number of possibilities. The Cho peak primarily measures water-soluble Cho compounds in the cytoplasm of oligodendroglia and in cell membranes and is generally increased by diseases that either injure glial cell membranes or stimulate increased membrane synthesis (51). The fact that the Cho/Cr ratio is typically elevated in active brain infection (20,22), acute infarction (61,64), inflammation (85), trauma (86), radiation (62), neoplasm (63), demyelination (27), and major depressive and bipolar disorders (15,32,33,87) but is normal to reduced in these patients with Gulf War syndrome argues against these causes. A few neurologic diseases may be associated with a normal or reduced Cho/Cr ratio, however, including later stages of infarction (61,88), atrophy (18,26,60,61), and chronic alcoholism (89), which may have unaffected or reduced glial mass. The normal brain MR images and our clinical evaluations (including normal liver function test results) ruled out these conditions as possible confounders. Thus, the ill Gulf War veterans in our sample appear to have lost neuronal mass in the basal ganglia and brainstem, in the absence of evidence for alternative causes such as demyelinating, inflammatory, infectious, or neoplastic processes or primary depressive disorders.

A limitation of many MR imaging and MR spectroscopic studies is the performance of statistical tests on a large number of brain regions without properly adjusting for multiple comparisons and without confirming findings on independent sets of subjects. We incorporated design features to avoid these pitfalls. First, we limited the number of hypotheses tested by obtaining single-voxel images on only the three brain regions suggested clinically by the similarity of the veterans' symptoms with those of well-understood diseases of deep brain structures (11) and the previously published neuropsychologic and audiovestibular test findings (1,2,4). Second, we rejected the overall null hypothesis in a family-wise (global) test in all three brain regions with a three-level case-control effect, also based on prior clinical findings (1,2,4,5), before proceeding with the descriptive analysis of the detailed effects. This is the preferred method of avoiding type I errors from multiple comparisons while also minimizing the chance of type II errors from overly conservative Bonferroni P value adjustments in families of comparisons that are correlated (54,90). Third, we replicated our main finding in a new sample of veterans with Haley syndrome 2 from an independent Gulf War veteran population and obtained a highly similar result, thus increasing the confidence in our main finding.

Our findings in a limited number of subjects are adequate to show important differences between ill and well Gulf War veterans. Larger studies are needed, however, before the techniques can be applied to diagnosis in individual patients. If our results are substantiated by additional larger studies, brain MR spectroscopy, along with other indicators such a blood enzyme levels and audiovestibular tests, might prove useful to help define a new nosologic classification or service-connected brain injury in Gulf War veterans.


   APPENDIX
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 APPENDIX
References
 


View this table:
[in this window]
[in a new window]

  		TABLE A1. Mean NAA/Cr Ratios (±SEM) in the Left and Right Basal Ganglia in Ill Gulf War Veterans and Matched Control Subjects
 

View this table:
[in this window]
[in a new window]

  		TABLE A2. Mean Cho/Cr Ratios (±SEM) in the Left and Right Basal Ganglia in Ill Gulf War Veterans and Matched Control Subjects
 

View this table:
[in this window]
[in a new window]

  		TABLE A3. Mean NAA/Cr and Cho/Cr Ratios (±SEM) in the Pons in Ill Gulf War Veterans and Matched Control Subjects
 


   Acknowledgments
 
The authors acknowledge the important contributions of Frederick J. Bonte, MD, and Michael D. Devous, Sr, PhD, for helpful suggestions on anatomic localization of the illness; Patricia Thompson, Maria Drake, and Alexandria Nugent for logistical support; and the medical and nursing staffs at our university's imaging facilities and our National Institutes of Health–supported General Clinical Research Center who cared for the subjects during the study.


   Footnotes
 
The content of this paper does not necessarily reflect the position or the policy of the U.S. government, and no official endorsement should be inferred.

Abbreviations: ANOVA = analysis of variance, Cho = choline, Cr = creatine, DEET = diethyltoluamide, NAA = N-acetylaspartate, PON1 = paraoxonase-1 gene, PON-Q = Q isoenzyme of paraoxonase, PON-R = R isoenzyme of paraoxonase

Author contributions: Guarantors of integrity of entire study, R.W.H., J.L.F.; study concepts, R.W.H., J.L.F., F.P., G.G.M.; study design, all authors; definition of intellectual content, all authors; literature research, R.W.H., J.L.F.; clinical studies, all authors; data acquisition, W.W.M., M.A.D., R.W.H.; data analysis, R.W.H., J.L.F., G.G.M.; statistical analysis, R.W.H.; manuscript preparation, R.W.H., J.L.F., G.G.M.; manuscript editing and review, all authors.


   References
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 APPENDIX
 References
 

  1. Haley RW, Kurt TM, Hom J. Is there a Gulf War syndrome? Searching for syndromes by factor analysis of symptoms. JAMA 1997; 277:215-222.[Abstract]
  2. Haley RW, Hom J, Roland PS, et al. Evaluation of neurologic function in Gulf War veterans: a blinded case-control study. JAMA 1997; 277:223-230.[Abstract]
  3. Haley RW, Kurt TL. Self-reported exposure to neurotoxic chemical combinations in the Gulf War: a cross-sectional epidemiologic study. JAMA 1997; 277:231-237.[Abstract]
  4. Roland PS, Haley RW, Yellin W, Owens K. Vestibular dysfunction in Gulf War syndrome. Otolaryngol Head Neck Surg 2000; 122:319-329Available at: http: //www1.Mosby.com/oto. Accessed March 29, 2000..[Medline]
  5. Haley RW, Billecke S, La Du BN. Association of low PON1 type Q (type A) arylesterase activity with neurololgic symptom complexes in Gulf War veterans. Toxicol Appl Pharm 1999; 157:227-233.[Medline]
  6. Konda I, Yamamoto M. Genetic polymorphism of paraoxonase 1 (PON1) and susceptibility to Parkinson's disease. Brain Res 1998; 806:271-273.[Medline]
  7. Freed VH, Matin MA, Fang SC, Kar PP. Role of striatal dopamine in delayed neurotoxic effects of organophosphorus compounds. Eur J Pharmacol 1976; 35:229-232.[Medline]
  8. Fernando JC, Hoskins B, Ho IK. Effect on striatal dopamine metabolism and differential motor behavioral tolerance following chronic cholinesterase inhibition with diisopropylfluorophosphate. Pharmacol Biochem Behav 1984; 20:951-957.[Medline]
  9. Matin MA, Hussain K. Striatal neurochemical changes and motor dysfunction in mipafox-treated animals. Methods Find Exp Clin Pharmacol 1985; 7:79-81.[Medline]
  10. McDonald BE, Costa LG, Murphy SD. Spatial memory impairment and central muscarinic receptor loss following prolonged treatment with organophosphates. Toxicol Lett 1988; 40:42-56.
  11. Lauterbach EC, Cummings JL, Duffy J, et al. Neuropsychiatric correlates and treatment of lenticulostriatal diseases: a review of the literature and overview of research opportunities in Huntington's, Wilson's, and Fahr's diseases—a report of the American Neuropsychiatric Association Committee on Research. J Neuropsychiatry Clin Neurosci 1998; 10:249-266.[Abstract/Free Full Text]
  12. Narayana PA, Johnston D, Flamig DP. In vivo proton magnetic resonance spectroscopy studies of human brain. Magn Reson Imaging 1991; 9:303-308.[Medline]
  13. Ross B, Kreis R, Ernst T. Clinical tools for the 90s: magnetic resonance spectroscopy and metabolite imaging. Eur J Radiol 1992; 14:128-140.[Medline]
  14. Cousins JP. Clinical MR spectroscopy: fundamentals, current applications and future potential. AJR Am J Roentgenol 1995; 164:1337-1347.[Abstract/Free Full Text]
  15. Soares JC, Krishnan KR, Keshavan MS. Nuclear magnetic resonance spectroscopy: new insights into the pathophysiology of mood disorders. Depression 1996; 4:14-30.[Medline]
  16. Duncan JS. Progress in epilepsy research: magnetic resonance spectroscopy. Epilepsia 1996; 37:598-605.[Medline]
  17. Shiino A, Matsuda M, Morikawa S, Inubushi T, Akiguchi I, Handa J. Proton magnetic resonance spectroscopy with dementia. Surg Neurol 1993; 39:143-147.[Medline]
  18. Ross BD, Bluml S, Cowan R, Danielsen E, Farrow N, Tan J. In vivo MR spectroscopy of human dementia. Neuroimaging Clin N Am 1998; 8:809-822.[Medline]
  19. Lazeyras F, Charles HC, Tupler LA, Erickson R, Boyko OB, Krishnan KR. Metabolic brain mapping in Alzheimer's disease using proton magnetic resonance spectroscopy. Psychiatry Res 1998; 82:95-106.[Medline]
  20. Cecil KM, Lenkinski RE. Proton MR spectroscopy in inflammatory and infectious brain disorders. Neuroimaging Clin N Am 1998; 8:863-880.[Medline]
  21. Menon DK, Baudouin CJ, Tomlinson D, Hoyle C. Proton MR spectroscopy and imaging of the brain in AIDS: evidence of neuronal loss in regions that appear normal with imaging. J Comput Assist Tomogr 1990; 14:882-885.[Medline]
  22. Menon DK, Ainsworth JG, Cox IJ, et al. Proton MR spectroscopy of the brain in AIDS dementia complex. J Comput Assist Tomogr 1992; 16:538-542.[Medline]
  23. Meyerhoff DJ, Weiner MW, Fein G. Deep gray matter structures in HIV infection: a proton MR spectroscopic study. Am J Neuroradiol 1996; 17:973-978.[Abstract]
  24. Jarvik JG, Lenkinski RE, Grossman RI, Gomori JM, Schnall MD, Frank I. Proton MR spectroscopy of HIV-infected patients: characterization of abnormalities with imaging and clinical correlation. Radiology 1993; 186:739-744.[Abstract/Free Full Text]
  25. Kreis R, Ross BD, Farrow NA, Ackerman Z. Metabolic disorders of the brain in chronic hepatic encephalopathy detected with H-1 MR spectroscopy. Radiology 1992; 182:19-27.[Abstract/Free Full Text]
  26. van der Knaap MS, van der Grond J, Luyten PR, den Hollander JA, Nauta JJ, Valk J. 1H and 31P magnetic resonance spectroscopy of the brain in degenerative cerebral disorders. Ann Neurol 1992; 31:202-211.[Medline]
  27. Richards TL. Proton MR spectroscopy in multiple sclerosis: value in establishing diagnosis, monitoring progression, and evaluating therapy. AJR Am J Roentgenol 1991; 157:1073-1078.[Abstract/Free Full Text]
  28. Rooney WD, Goodkin DE, Schuff N, Meyerhoff DJ, Norman D, Weiner MW. 1H MRSI of normal appearing white matter in multiple sclerosis. Mult Scler 1997; 3:231-237.[Medline]
  29. De Stefano N, Matthews PM, Fu L, et al. Axonal damage correlates with disability in patients with relapsing-remitting multiple sclerosis: results of a longitudinal magnetic resonance spectroscopy study. Brain 1998; 121:1469-1477.[Abstract/Free Full Text]
  30. Fu L, Wolfson C, Worsley KJ, et al. Statistics for investigation of multimodal MR imaging data and an application to multiple sclerosis patients. NMR Biomed 1996; 9:339-346.[Medline]
  31. Bertolino A, Nawroz S, Mattay VS, et al. Regionally specific pattern of neurochemical pathology in schizophrenia as assessed by multislice proton magnetic resonance spectroscopic imaging. Am J Psychiatry 1996; 153:1554-1563.[Abstract/Free Full Text]
  32. Charles HC, Lazeyras F, Krishnan KR, Boyko OB, Payne M, Moore D. Brain choline in depression: in vivo detection of potential pharmacodynamic effects of antidepressant therapy using hydrogen localized spectroscopy. Prog Neuropsychopharmacol Biol Psychiatry 1994; 18:1121-1127.[Medline]
  33. Kato T, Hamakawa H, Shioiri T, et al. Choline-containing compounds detected by proton magnetic resonance spectroscopy in the basal ganglia in bipolar disorder. J Psychiatry Neurosci 1996; 21:248-254.[Medline]
  34. Moore CM, Christensen JD, Lafer B, Fava M, Renshaw PF. Lower levels of nucleoside triphosphate in the basal ganglia of depressed subjects: a phosphorous-31 magnetic resonance spectroscopy study. Am J Psychiatry 1997; 154:116-118.[Abstract]
  35. Kato T, Inubushi T, Kato N. Magnetic resonance spectroscopy in affective disorders. J Neuropsychiatry Clin Neurosci 1998; 10:133-147.[Abstract/Free Full Text]
  36. Ebert D, Speck O, Konig A, Berger M, Hennig J, Hohagen F. 1H-magnetic resonance spectroscopy in obsessive-compulsive disorder: evidence for neuronal loss in the cingulate gyrus and the right striatum. Psychiatry Res 1997; 74:173-176.[Medline]
  37. Bartha R, Stein MB, Williamson PC, et al. A short echo 1H spectroscopy and volumetric MRI study of the corpus striatum in patients with obsessive-compulsive disorder and comparison subjects. Am J Psychiatry 1998; 155:1584-1591.[Abstract/Free Full Text]
  38. Dager SR, Strauss WL, Marro KI, Richards TL, Metzger GD, Artru AA. Proton magnetic resonance spectroscopy investigation of hyperventilation in subjects with panic disorder and comparison subjects. Am J Psychiatry 1995; 152:666-672.[Abstract/Free Full Text]
  39. Dager SR, Richards T, Strauss W, Artru A. Single-voxel 1H-MRS investigation of brain metabolic changes during lactate-induced panic. Psychiatry Res 1997; 76:89-99.[Medline]
  40. Dager SR, Friedman SD, Heide A, et al. Two-dimensional proton echo-planar spectroscopic imaging of brain metabolic changes during lactate-induced panic. Arch Gen Psychiatry 1999; 56:70-77.[Abstract/Free Full Text]
  41. Freeman TW, Cardwell D, Karson CN, Komoroski RA. In vivo proton magnetic resonance spectroscopy of the medial temporal lobes of subjects with combat-related posttraumatic stress disorder. Magn Reson Med 1998; 40:66-71.[Medline]
  42. Prost RW, Mark LP, Rand SD, Kim TA, Haughton VM. Recent advances in MR spectroscopy expand its applications in neurologic disease. Wisc Med J 1997; 96:41-45.
  43. Hom J, Haley RW, Kurt TL. Neuropsychological correlates of Gulf War syndrome. Arch Clin Neuropsychol 1997; 12:531-544.
  44. Reitan RM, Wolfson D. The Halstead-Reitan Neuropsychology Test Battery: theory and clinical interpretation 2nd ed. Tucson, Ariz: Neuropsychology Press, 1993.
  45. Sullivan RM, Szechtman H. Left/right nigrostriatal asymmetry in susceptibility to neurotoxic dopamine depletion with 6-hydroxydopamine in rats. Neurosci Lett 1994; 170:83-86.[Medline]
  46. Fleminger S. Left-sided Parkinson's disease is associated with greater anxiety and depression. Psychol Med 1991; 21:629-638.[Medline]
  47. Carlson JN, Visker KE, Keller RWJ, Glick SD. Left and right 6-hydroxydopamine lesions of the medial prefrontal cortex differentially alter subcortical dopamine utilization and the behavioral response to stress. Brain Res 1996; 711:1-9.[Medline]
  48. Louilot A, Choulli MK. Asymmetrical increases in dopamine turn-over in the nucleus accumbens and lack of changes in locomotor responses following unilateral dopaminergic depletions in the entorhinal cortex. Brain Res 1997; 778:150-157.[Medline]
  49. Christiansen P, Toft P, Larsson HB, Stubgaard M, Henriksen O. The concentration of N-acetyl aspartate, creatine + phosphocreatine, and choline in different parts of the brain in adulthood and senium. Magn Reson Imaging 1993; 11:799-806.[Medline]
  50. Charles HC, Lazeyras F, Krishnan KR, et al. Proton spectroscopy of human brain: effects of age and sex. Prog Neuropsychopharmacol Biol Psychiatry 1994; 18:995-1004.[Medline]
  51. Miller BL. A review of chemical issues in 1H NMR spectroscopy: N-acetyl-L-aspartate, creatine and choline. NMR Biomed 1991; 4:47-52.[Medline]
  52. Webb PG, Sailasuta N, Kohler SJ, Raidy T, Moats RA, Hurd RE. Automated single-voxel proton MRS: technical development and multisite verification. Magn Reson Med 1994; 31:365-373.[Medline]
  53. de Beer R, van den Boogaart A, van Ormondt D, et al. Application of time-domain fitting in the quantification of in vivo 1H spectroscopic imaging data sets. NMR Biomed 1992; 5:171-178.[Medline]
  54. Ludbrook J. Multiple comparison procedures updated. Clin Exp Pharmacol Physiol 1998; 25:1032-1037.[Medline]
  55. Winer BJ. Statistical principles in experimental design 2nd ed. New York, NY: McGraw-Hill, 1971.
  56. Siegel S. Nonparametric statistics for the behavioral sciences New York, NY: McGraw-Hill, 1956; 184-194.
  57. Shapiro SS, Wilk MB. An analysis of variance test for normality (complete samples). Biometrika 1965; 52:591-611.[Free Full Text]
  58. Birken DL, Oldendorf WH. N-acetyl-L-aspartic acid: a literature review of a compound prominent in 1H-NMR spectroscopic studies of brain. Neurosci Biobehav Rev 1989; 13:23-31.[Medline]
  59. Burtscher IM, Stahlberg F, Holtas S. Proton (1H) MR spectroscopy for routine diagnostic evaluation of brain lesions. Acta Radiol 1997; 38:953-960.[Medline]
  60. Pioro EP. MR spectroscopy in amyotrophic lateral sclerosis/motor neuron disease. J Neurol Sci 1997; 152(suppl):S49-S53.
  61. Friedman SD, Stidley CA, Brooks WM, Hart BL, Sibbitt WLJ. Brain injury and neurometabolic abnormalities in systemic lupus erythematosus. Radiology 1998; 209:79-84.[Abstract/Free Full Text]
  62. Esteve F, Rubin C, Grand S, Kolodie H, Le Bas JF. Transient metabolic changes observed with proton MR spectroscopy in normal human brain after radiation therapy. Int J Radiat Oncol Biol Phys 1998; 40:279-286.[Medline]
  63. Rand SD, Prost R, Haughton V, et al. Accuracy of single-voxel proton MR spectroscopy in distinguishing neoplastic from nonneoplastic brain lesions. AJNR Am J Neuroradiol 1997; 18:1695-1704.[Abstract]
  64. Mathews VP, Barker PB, Blackband SJ, Chatham JC, Bryan RN. Cerebral metabolites in patients with acute and subacute strokes: concentrations determined by quantitative proton MR spectroscopy. AJR Am J Roentgenol 1995; 165:633-638.[Abstract/Free Full Text]
  65. Koller KJ, Zaczek R, Coyle JT. N-acetyl-aspartyl-glutamate: regional levels in rat brain and the effects of brain lesions as determined by a new HPLC method. J Neurochem 1984; 43:1136-1142.[Medline]
  66. Tedeschi G, Bertolino A, Righini A, et al. Brain regional distribution pattern of metabolite signal intensities in young adults by proton magnetic resonance spectroscopic imaging. Neurology 1995; 45:1384-1391.[Abstract]
  67. Truckenmiller ME, Namboodiri MA, Brownstein MJ, Neale JH. N-acetylation of L-aspartate in the nervous system: differential distribution of a specific enzyme. J Neurochem 1985; 45:1658-1662.[Medline]
  68. Barker RA. The basal ganglia and pain. Int J Neurosci 1988; 41:29-34.[Medline]
  69. Giroud M, Lemesle M, Madinier G, Billiar T, Dumas R. Unilateral lenticular infarcts: radiological and clinical syndromes, aetiology, and prognosis. J Neurol Neurosurg Psychiatry 1997; 63:611-615.[Abstract/Free Full Text]
  70. Casey BJ, Castellanos FX, Giedd JN, et al. Implication of right frontostriatal circuitry in response inhibition and attention-deficit/hyperactivity disorder. J Am Acad Child Adolesc Psychiatry 1997; 36:374-383.[Medline]
  71. Afifi AK. Basal ganglia: functional anatomy and physiology: part 2. J Child Neurol 1994; 9:352-361.[Medline]
  72. Baloh RW, Honrubia V. Clinical neurophysiology of the vestibular system 2nd ed. Philadelphia, Pa: Davis, 1990; 182-185.