Introduction: Perfusion studies using SPECT (1) and ASL (2) imaging methods have demonstrated significantly different baseline hippocampal perfusion and hippocampal perfusion responses to cholinergic challenge via physostigmine infusion in veterans with Gulf War Illness, consistent with the memory loss and movement disorders reported for ill Gulf War veterans (3,4), and corroborating the hippocampal dysfunction reported in MRS studies (5). The previous SPECT and ASL perfusion studies were performed with veterans selected from the 24th U.S. Naval Reserve Mobile Construction Battalion (CB), a sample not statistically representative of all Gulf War veterans. To verify these previous findings, subjects were recruited from a representative national survey of over 8,000 veterans for arterial spin labeling hippocampus perfusion studies similar to those of the previous CB veteran study. The preliminary results reported here are for the sickest of the three major Gulf War illness variants, Syndrome 2. Materials and Methods: Subjects were selected from a cohort of 8,020 Gulf War veterans statistically representative of all active for the 1991 Gulf War, and classified by factor analysis (3) into veterans with Gulf War Syndromes 1, 2, 3, and deployed and non-deployed healthy controls. The preliminary results reported here are from Gulf War Syndrome 2 (Syn 2) (3,4) and healthy non-deployed veterans, as described in Table 1. All subjects were screened and gave written informed consent according to a study protocol approved by the local Institutional Review Board. A two-session perfusion study, double-blinded by group and single-blinded by infusate, was performed during the afternoon for each subject. Saline was infused during the first session and physostigmine (~0.6 mg), a short-acting cholinesterase inhibitor, was infused during the second session, each at 130 mL/hour for 30 minutes prior to imaging. Glycopyrrolate (0.3 mg) was injected prior to physostigmine infusion to combat nausea; a placebo saline injection was given prior to the saline infusion in the first session. During the imaging, subjects were awake with eyes closed. Studies were performed on a 3T Siemens TIM Trio whole-body MR scanner. The body coil was used for RF transmission; a Siemens 12-channel phased array head coil was used for signal reception. OPTIMAL FAIR (6) was used, with 3.5 mm thick oblique coronal slices. Two of the 20 slices were anterior to the hippocampus head. Imaging slices were reproducibly placed using Auto-Align, and 3D PACE (7), a real-time prospective motion correction technique, was used to help reduce motion artifacts. Hippocampus perfusion imaging parameters were: TR/TE = 3000/14 ms, FOV = 128 x 128 mm, matrix size = 64 x 64, in-plane resolution = 2 x 2 mm, slice thickness/distance factor = 3.5 mm/20%, number of imaging slices = 20, left to right phase encoding with 30% oversampling, 6/8 partial Fourier, iPAT GRAPPA factor = 2, A-P slice acquisition order, advanced 3D shimming, selective/spatiallyconfined inversion slab =148/328 mm, temporal bolus width (TI1)/ post-bolus delay (TI2) = 600/1000 ms, inferior saturation pulse number/size/ repetition interval = 40/20 mm/25 ms, and 150 labeling and control image pairs. Proton density images (M0) were acquired using the same EPI imaging parameters but 8 s TR. The hippocampus was segmented using the FIRST tool of FSL software. Hippocampus ROIs and corresponding anatomic images were co-registered to ASL time series. Image pre-processing was performed with SPM. CBF quantification used a single blood compartment model (8). Hippocampus CBF percentage changes were calculated as ∆CBF(%) = (CBF (physostigmine) – CBF (saline))/CBF (saline) x 100. Two tailed t tests were performed to detect significant differences (p<0.05) of measured perfusion across two sessions (paired t tests) and between groups (unpaired t tests) in left, right and bilateral regions of the hippocampus. Results and Discussion: Figure 1 shows co-registered anatomic, proton density and perfusion-weighted images of one representative healthy veteran, from the saline infusion session. Physostigmine challenge decreased hippocampal perfusion in the healthy non-deployed veterans, but increased hippocampal perfusion in veterans ill with Syndrome 2 (Figure 2), which is similar to previous findings (2). These hippocampus perfusion responses to physostigmine infusion are significantly opposite between the Syn2 and non-deployed control veteran groups in left, right and bilateral hippocampus (Figure 3). These preliminary results from a representative sampling of Gulf War veterans confirm the abnormal hippocampus perfusion observed in Gulf War veterans ill with Syndrome 2 in the previous study of subjects in the CB cohort (2). Acknowledgements: This study was supported by IDIQ contract VA549-P-0027, awarded and administered by the Department of Veterans Affairs Medical Center, Dallas, TX, , by DoD grant DAMD 17-01-1-0741, and by NIH (NCRR) Grant Number UL1RR024982. The content does not necessarily reflect the position or the policy of the Federal government or the sponsoring agencies, and no official endorsement should be inferred. References: 1. Haley et al. Psychiatry Res 2009; 171(3): 207-220. 2. Li et al. Proc ISMRM, 2010; 1978. 3. Haley et al. JAMA 1997; 277(3): 215-222. 4. Haley et al. Psychiatry Res 2001; 102(2): 175-200. 5. Menon et al. Brain Res 1009(1-2): 189-94. 6. Li et al. Proc ISMRM, 2010; 1756. 7. Thesen et al. MRM 2000; 44:457-465. 8. Buxton et al. JMRI 2005;22(6): 723-726. Fig. 1 Co-registered anatomic (top), proton density (middle) and perfusion-weighted (bottom) images of one healthy veteran from saline session. Eleven slices covering the hippocampus are displayed.

K. Gopinath, W. Ringe, L. Ouyang, K. Carter, B. Thapa-Chhetry, L. Butler, A. Goyal, P. Gandhi, Y. Fang, S. Ganji, L. Jiang, S. Vaidya, R. Briggs, and R. Haley Department of Radiology, UT Southwestern Medical Center, Dallas, TX, United States, Department of Internal Medicine, UT Southwestern Medical Center, Dallas, TX, United States, Department of Psychiatry, UT Southwestern Medical Center, Dallas, TX, United States

K. Gopinath, L. Butler, B. Thapa-Chhetry, A. Goyal, P. Gandhi, Y. Fang, L. Ouyang, S. Ganji, L. Jiang, S. Vaidya, D. Buhner, W. Ringe, R. Briggs, and R. Haley Department of Radiology, UT Southwestern Medical Center, Dallas, TX, United States, Department of Internal Medicine, UT Southwestern Medical Center, Dallas, TX, United States, Department of Psychiatry, UT Southwestern Medical Center, Dallas, TX, United States

Introduction: Central pain is a widespread symptom of ill Gulf War veterans [1,2]. A previous study [3] reported a two-fold increase in cooling detection threshold during Quantitative Sensory Testing (QST) of right foot in GW Illness veterans. Higher cooling thresholds (in all extremities) and higher warming thresholds (in hands) in veterans suffering from Gulf War Illness were also found in another previous study [unpublished]. In this study, brain activation in response to warm sensation stimuli and painfully hot stimuli was measured with a QST fMRI paradigm, and differences between four groups of Gulf War veterans with Syndromes 1 (Syn1), 2 (Syn2), 3 (Syn3) [2], and control group were assessed. Methods: Fifty-three right–handed male Gulf-War Veterans: 11 Syn1 (ages 40-60 yrs; mean 51.3 yrs), 16 Syn2 (ages 53-73 yrs; mean 62.8 yrs), 12 Syn3 (ages 47-66 yrs; mean 56.7 yrs), and 14 controls (ages 51-76 yrs; mean 60.5 yrs), were studied with a QST fMRI paradigm. Written informed consent was obtained from all subjects. Warm sensation and hot-pain temperature thresholds for all subjects were first determined outside the scanner with a Medoc Pathway with ATS thermode (Medoc, Ramat Yishai, Israel), using the method of limits [4]. The thermode was placed on the right inner forearm of the subjects. MR scans were performed with a Siemens 3T Tim Trio scanner using a 12-channel array receive-only head coil. During each of the six fMRI scans (3 each for warm sensation and hot-pain sensation) 10 thermal stimuli were applied. For each stimulus, the temperature ramped up to the threshold temperature at a rate of 8°C/sec and stayed at that temperature for 3 sec before ramping down to the baseline temperature of 32°C/sec. ISIs of 14, 16 and 18 seconds in duration were pseudo-randomized. FMRI scans were obtained with a whole-brain sagittal gradient echo EPI sequence (TR/TE = 2000/24 ms, FA = 90, in-plane resolution = 3 mm x 3 mm, 40 slices with thickness 3.5-4 mm). A highresolution T1-weighted anatomical scan using a MPRAGE sequence was also acquired. The 3 hot-pain temperature functional scans were separated in time (by 4 – 5 min) by doing the anatomical scans between them to avoid any confounds that might arise because of subject sensitization to repeated high temperature stimuli. The voxel time series data from each condition were motion corrected, smoothed with a FWHM = 5 mm isotropic gaussian filter and concatenated. Hemodynamic response (HDR) to the warm and pain stimuli were estimated with GLM-based deconvolution analysis. The estimated HDR maps were spatially normalized to the Talairach template. Individual group activation maps and between-group differences in warm sensation and hot pain activation were assessed with Student’s t-tests on the HDR amplitudes. These t-maps were clustered and significance of cluster-level activation was assessed with Monte-Carlo modeling [5]. Results and Discussion: No significant differences were noticed in the warm sensation threshold (p > 0.5) assessed outside the scanner between the 4 groups, and in the hot pain threshold (p > 0.5) between control group and Syn2 & Syn3. Syn1 had significantly higher hot pain threshold (p < 0.05) than the other groups. FMRI activation to warm sensation in the control group was similar to what is seen in age-matched controls [6]. The control group showed significantly higher activation (p < 0.01) to warm sensation compared to Syn1 and Syn2 (Figure 1, Table 1). No significant difference between the warm sensation activation patterns of the control groups and Syn3 was noted.

Introduction Abnormalities in ratios of brain metabolite peak areas were previously measured via magnetic resonance spectroscopy (MRS) at 1.5T in the pons and basal ganglia [1,2] and in bilateral hippocampus [3] of Gulf War Syndrome patients. The original veterans of the Seabees cohort studied in [1] recently participated in a follow-up study at 3T. A part of the protocol used single voxel spectroscopy (SVS) H MR to study metabolite concentrations in the left and right basal ganglia of normal controls and Syndrome II (“confusion-ataxia”) [4] patients. Personnel performing the data acquisition and initial analysis were blinded with respect to the group memberships. The blind was then partially lifted to split the subjects into two unidentified groups A and B. The group comparison indicates significantly lower N-acetylaspartate-tocreatine (NAA/Cr) ratio in group A compared to group B. We hypothesize that group A is Syndrome II and group B is normal controls. This pilot study helps to improve the understanding of the chemical changes in the brain in Gulf War Syndrome and possibly in other related neurodegenerative conditions. Methods Twenty Gulf War veterans (9 members of group A and 11 members of group B) were studied with the same SVS left and right basal ganglia protocol. A standard single voxel spectroscopy sequence (SVS PRESS) was used on a Siemens 3T Trio TIM with the following parameters: TR/TE/NS = 2500ms/30ms/96, voxel (centered in the basal ganglia, bilaterally) volume = 12.0 mL (20 mm x 30 mm x 20 mm), spectral width = 2000 Hz, water suppression bandwidth = 50 Hz, data points = 1024, acquisition time = 4:10 min. An unsuppressed water spectrum was also acquired for eddy current compensation and quantitation. Good quality high-resolution and high-contrast localizer images were used for proper voxel positioning to ensure its reproducibility. Reproducibility and quality of the single voxel spectroscopy (SVS) methods used was previously demonstrated [5]. Post-processing of the MRS data was performed using LCModel [6]. Group and hemispheric differences were studied with unbalanced 2-way Analysis of Variance (ANOVA). Results and Discussion Manual adjustment of the shim currents produced metabolite half-height line width of 9-14 Hz in most sessions. An example of the basal ganglia spectrum processed with LCModel is shown in Figure 1. Although a water reference spectrum was collected for metabolite quantitation, the NAA/Cr ratio is reported for comparison with the original 1997-8 study and because it is less sensitive to individual shim variations and possible metabolite relaxation time differences [7]. The NAA/Cr ratio in the basal ganglia was significantly lower in group A than in group B (group effect p=0.0125, unbalanced 2-way ANOVA). The group difference is greater in the right basal ganglia (10% difference, p=0.022) than in the left (8% difference, p=0.15). See also Table 1 and Figure 2. In the original study [1], Syndrome II patients had lower, highly statistically significant NAA/Cr in the right basal ganglia than controls (18% difference, p < 0.001) and lower, but less significant NAA/Cr in the left basal ganglia (9% difference, p < 0.09). Additionally, the NAA/Cr ratio is significantly higher in the left basal ganglia than the right in both groups (overall hemispheric effect significance of p=0.0001). Conclusion N-acetylaspartate is widely regarded as a marker of neuronal viability and any changes in disease are important. Our preliminary findings show differences in this ratio between controls and Gulf War syndrome II patients. However, to transition from metabolite ratios (or “institutional units”, if water reference is used) for metabolite concentrations to absolute metabolite concentrations, relaxation time measurements in vivo are needed [7,8] as well as determination of the compartmentation in the human brain [9]. Acknowledgments This study was supported by the VA IDIQ contract number VA549-P-0027 awarded and administered by the Dallas, TX VA Medical Center. 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. The authors would like to acknowledge Victoria Vescovo for help with the data acquisition and Drs. Kaundinya Gopinath and Mette Posamentier for useful discussions. References [1] Haley, R. et al., Radiology 2000; 215: 807-817. [2] Meyerhoff, D.J. et al., Proc Intl Soc Mag Reson Med 2001; 9: 994. [3] Menon et al., Brain Res. 2004; 1009: 189-194. [4] Haley R. et al., JAMA 1997; 277:215-222. [5] Cheshkov, S. et al, Proc. Intl. Soc. Mag. Res. Med., 15: 1357 (2007). [6] Provencher, S., Magn Reson Med 1993; 30: 672-679. [7] Traber, F. et al., J Magn Reson, 19: 537-545 (2004) [8] Chang, A. et al., Proc. Intl. Soc. Mag. Res. Med. 16: 1600 (2008). [9] Ernst, T. et al., J Magn Reson, B 1993, 102: 1-8.