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 Vermathen, P.
Right arrow Articles by Weiner, M. W.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Vermathen, P.
Right arrow Articles by Weiner, M. W.
(Radiology. 2000;214:403-410.)
© RSNA, 2000

Neuroradiology

Hippocampal Structures: Anteroposterior N-acetylaspartate Differences in Patients with Epilepsy and Control Subjects as Shown with Proton MR Spectroscopic Imaging1

Peter Vermathen, Dr rer nat, Kenneth D. Laxer, MD, Gerald B. Matson, PhD and Michael W. Weiner, MD

1 From the Magnetic Resonance Spectroscopy Unit, 114 M, Department of Veterans Affairs Medical Center, 4150 Clement St, San Francisco, CA 94121 (P.V., G.B.M., M.W.W.), and the Departments of Neurology (K.D.L.), Radiology (P.V., M.W.W.), Pharmaceutical Chemistry (G.B.M.), and Medicine (M.W.W.), University of California, San Francisco. Received August 26, 1998; revision requested October 5; revision received June 9, 1999; accepted July 22. P.V. supported in part by Deutsche Forschungsgemeinschaft Forschungsstipendium VE 190/1-1, and K.D.L., G.B.M., and M.W.W. supported in part by National Institutes of Health grant ROI-NS31966. Address reprint requests to P.V. (e-mail: vermath@itsa.ucsf.edu).


   Abstract
TOP
 Abstract
Introduction
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
PURPOSE: To determine the distribution of proton metabolites along the long axis of the hippocampus.

MATERIALS AND METHODS: Proton magnetic resonance (MR) spectroscopic imaging measurements were performed in the hippocampi of 14 control subjects and nine patients with unilateral mesial temporal lobe epilepsy.

RESULTS: Control subjects showed significantly lower ratios of N-acetylaspartate (NAA) to choline-containing compounds (Ch) and creatine plus phosphocreatine (CR) (NAA/[Cr + Ch]) in the anterior as compared with the posterior part of the hippocampus. Furthermore, a similar anteroposterior (AP) difference in NAA/(Cr + Ch) values was found in both ipsilateral and contralateral hippocampi of patients. In the patients compared with the control subjects, ipsilateral NAA/(Cr + Ch) levels were reduced in every part of hippocampal tissue with an average reduction of 17%, and contralateral NAA/(Cr + Ch) was reduced by about 10%. In the patients compared with the control subjects, the proportional reduction in ipsilateral NAA/(Cr + Ch) was greatest in voxels from anterior hippocampal regions.

CONCLUSION: AP differences could be a result of fewer neurons in the anterior compared with the posterior hippocampus or of the increasing thickness of the hippocampus from posterior to anterior, which leads to different contributions from adjacent tissue. Measurements of T2 showed that T2 differences are probably not responsible for these changes.

Index terms: Brain, atrophy, 1341.83 • Brain, metabolism, 1341.99 • Brain, MR, 1341.121411, 1341.12145 • Epilepsy, 1341.83 • Magnetic resonance (MR), spectroscopy, 1341.12145


   Introduction
TOP
 Abstract
 Introduction
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
In vivo proton MR spectroscopy of the hippocampus has been used in a number of diseases including epilepsy (1–7) and Alzheimer disease (8,9) to detect either metabolite differences between left and right hippocampi or differences in comparison with control subjects. In patients with mesial temporal lobe epilepsy (TLE), most studies detected reduced N-acetylaspartate (NAA) ipsilateral to the seizure focus either in absolute units or as compared with choline-containing compounds (Ch) and/or creatine plus phosphocreatine (Cr). NAA was reduced ipsilateral to the seizure focus when compared with the contralateral side or when compared with results in control subjects. In addition, many patients with mesial TLE also had reduced NAA in the contralateral hippocampus (3–7).

To our knowledge, however, little attention has been paid to anteroposterior (AP) metabolite differences in the hippocampus. Most previous studies did not pinpoint where in the hippocampus the data were derived. Since the AP extent of the hippocampus is approximately 3–4 cm, the possibility exists that there may be differences in the metabolites depending on the location of the study. There is some evidence that the distribution of NAA, which is located exclusively in neurons, might not be uniform in the hippocampus. In a study of the number of neuronal cells in the hippocampus of healthy control subjects, Mouritzen Dam (10) found an increase in the number of cells from anterior to posterior in the pyramidal cells in all subfields investigated. In addition, the number of granule cells varied by position in the hippocampus, but in this case the number decreased from anterior to posterior. In a similar study, Babb et al (11) did not find statistically significant differences in the number of cells between the posterior and anterior sections. In most subfields, however, the number of neuronal cells were greater posteriorly than anteriorly.

Another potential source of metabolite differences in hippocampal tissue is a result of tissue contamination. In most MR spectroscopic studies, the voxel size exceeded the thickness of the hippocampus. Thus, since the hippocampus is thicker anteriorly than posteriorly, different contributions from adjacent tissue could produce metabolite differences depending on the position of the voxel in the AP direction. Furthermore, in an MR imaging study of the hippocampus, the transverse relaxation time of the water signal was found to increase from the posterior to anterior hippocampus by approximately 4% (12). Therefore, the first goal of this study was to determine if there are statistically significant differences in metabolites and metabolite ratios within different regions of the hippocampus in control subjects.

Findings in several studies have shown that the neuronal loss in patients with epilepsy is greater in the anterior than in the posterior part of the hippocampus (11,13,14). The second goal of this study was therefore to investigate whether AP differences were also detectable in patients with epilepsy and whether greater neuronal loss anteriorly provided a better separation between patients and control subjects. Potentially, AP differences might partially account for the measured differences between patients and control subjects.


   MATERIALS AND METHODS
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 References
 
Subjects
Fourteen healthy control subjects (10 men and four women; age range, 24–50 years; mean age, 33.9 years ± 7.4 [SD]) and nine patients (eight men and one woman; age range, 26–51 years; mean age, 36.2 years ± 9.3) with mesial TLE that was unilateral and medically refractory (on the basis of seizure symptoms and ictal electroencephalographic recordings) were included in this study (Table 1). All patients were evaluated at the Northern California Comprehensive Epilepsy Center (San Francisco). For patients with epilepsy, the ictal electroencephalographic findings were used as the standard for localization and lateralization of the seizure focus (ie, ipsilateral side). We had institutional review board approval, and informed consent was obtained from all subjects prior to MR spectroscopic imaging studies.


View this table:
[in this window]
[in a new window]
  		TABLE 1. Characteristics of Patients with Mesial TLE
 
MR Imaging and MR Spectroscopic Imaging Methods
Proton MR spectroscopic imaging was performed with a 1.5-T imager (Magnetom Vision; Siemens, Erlangen, Germany) with use of a standard circularly polarized head coil. Two-dimensional fast low-angle shot, or FLASH, images were acquired in coronal and sagittal orientations. Transverse oblique T1-weighted images (repetition time msec/echo time msec = 500/14; section thickness, 3 mm) were acquired with angulation parallel to the long axis of the hippocampus. With point-resolved spatially localized spectroscopy (PRESS), volume preselection including both hippocampi was performed parallel to the oblique transverse images, with typical transverse dimensions of 15 mm, left to right 105 mm, and anterior to posterior 70 mm (Fig 1). After global shimming (MAPSHIM; Siemens), we used the same PRESS sequence but without phase-encoding gradients for a localized shim. The PRESS box for shimming was smaller than the actual dimensions for the spectroscopic imaging measurement, and it was placed mainly in the anterior hippocampus. Our experience was that this local shimming improved the spectral quality from the anterior hippocampus but still maintained good spectral quality in posterior regions. Circular k-space encoding with a 24-point diameter (2) was used with a repetition time of 1.8 seconds and echo time of 135 msec, which resulted in a measurement time of 13 minutes.



View larger version (96K):
[in this window]
[in a new window]
  		Figure 1. Sagittal (left) and transverse (right) T1-weighted images (500/14; section thickness, 3 mm) with superimposed outline that indicates the typical PRESS volume.

 
Two different protocols were used. In all patients and in 10 control subjects, the 90° pulse of the PRESS sequence was applied perpendicular to the sagittal orientation (standard sequence). In addition in three of the 10 control subjects, a T2 measurement was performed with the same sequence with use of five different echo times (135, 202, 270, 405, and 540 msec). In four additional control subjects, the 90° pulse was applied perpendicular to the coronal direction to exclude the influence of an imperfect 180° pulse profile on metabolites in the AP direction.

Postprocessing
Postprocessing included zero filling to 32 x 32 and apodization in the spatial domain, which resulted in an effective voxel size of 2 cm3. The time domain data were zero filled to 1,024 data points and apodized with a Gaussian function, which corresponded to 2-Hz line broadening in the frequency domain. Spectral phasing and a polynomial baseline correction were also performed. The signals of NAA, Cr, and Ch were quantitated by means of curve fitting. The data were analyzed with standard spectroscopic software (LUISE; Siemens). To compare the metabolite signals between subjects, the signals were corrected for coil loading by multiplying the signal by the transmitter reference value (15).

Measurements with a homogeneous water phantom were performed (P.V.) with identical measurement parameters to determine experimentally the pulse excitation profiles of the 180° and 90° pulses, and the intensity of the water signal was used to correct for differences in metabolite areas and metabolite ratios due to chemical-shift differences. The data reported herein were corrected for these differences in metabolite areas and chemical-shift differences. The 90° pulse excited a section with a trapezoidal profile that was close to rectangular. Therefore, no correction for the section profile was applied for the control subjects measured with the 90° pulse in the AP direction.

Statistical Analysis
To analyze the significance of the measured slopes in the AP direction, P values were calculated by means of a one-population two-tailed t test. For comparisons between patients with mesial TLE and control subjects, two-population two-tailed t tests were performed. Differences with a P value of .05 were considered statistically significant.


   RESULTS
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 References
 
Figure 1 shows sagittal and transverse T1-weighted images (500/14; section thickness, 3 mm) with a superimposed outline that indicates the typical PRESS volume. The anterior border of the hippocampus was used as a landmark for positioning the PRESS box.

Figure 2a depicts a sketch of the hippocampi and midbrain from an MR image in a control subject. An overlaid grid indicates the measurement matrix of the MR spectroscopic imaging measurement after zero filling and the positions from which the spectra were derived. A maximum of 12 spectra were analyzed for each hippocampus (six rows, two spectra in each row). Results for the two adjacent spectra in each row were averaged. The actual voxel size was 2 mL and therefore was bigger than the grid indicates. Thus, there is some information overlap between adjacent voxels. Figure 2a shows the approximate actual voxel size (shown as circles) as calculated from the point spread function (16). Spectra from the different locations within the hippocampal area are shown in Figure 2b. The regions from which each spectrum originated are indicated. Figure 2b clearly demonstrates a lower NAA anteriorly than posteriorly both as an absolute measure and relative to Ch and Cr.



View larger version (105K):
[in this window]
[in a new window]
  		Figure 2a. (a) Sketch of the hippocampi and midbrain from an MR image obtained in a control subject. Overlaid grid indicates the matrix of the MR spectroscopic imaging measurement after zero filling and the positions from which the spectra in b were derived. The circles indicate actual voxel size. The numbers 1-6 label the rows within the hippocampal area. (b) Spectra from rows 1-6 in a. The peaks are labeled.

 


View larger version (22K):
[in this window]
[in a new window]
  		Figure 2b. (a) Sketch of the hippocampi and midbrain from an MR image obtained in a control subject. Overlaid grid indicates the matrix of the MR spectroscopic imaging measurement after zero filling and the positions from which the spectra in b were derived. The circles indicate actual voxel size. The numbers 1-6 label the rows within the hippocampal area. (b) Spectra from rows 1-6 in a. The peaks are labeled.

 
Results with the Standard PRESS Sequence
All patients and most control subjects underwent MR spectroscopic imaging with a PRESS sequence with the 90° pulse perpendicular to the sagittal orientation (see Materials and Methods). A linear relationship between metabolite areas derived from the six positions in the AP direction was arbitrarily assumed so the slope for each control subject and each patient, in ipsi- and contralateral hippocampal tissue, could be calculated. The slopes were calculated for the metabolites and NAA/(Cr + Ch) by starting from the most anterior position and proceeding in the posterior direction. Figure 3 shows the slopes for all patients and control subjects. Mean slopes and P values for a one-population two-tailed t test are displayed in Table 2.



View larger version (13K):
[in this window]
[in a new window]
  		Figure 3. AP slopes of metabolite areas and NAA/(Cr + Ch) for individual participants, calculated by assuming a linear relationship between metabolite areas, were derived from the six positions in the AP direction in hippocampal tissue (1-6 in Fig 2a). contra = contralateral, ctrl = control, ipsi = ipsilateral.

 

View this table:
[in this window]
[in a new window]
  		TABLE 2. Average Slopes in the AP Direction in Hippocampal Tissue
 
Figure 3 shows that in all control subjects and patients, in ipsi- and contralateral hippocampal tissue, NAA/(Cr + Ch) exhibited a positive slope from anterior to posterior, that is, the anterior NAA/(Cr + Ch) value was lower than the posterior value in all participants (P < .001). Furthermore, lower NAA was found anteriorly in both hippocampi in eight of 10 control subjects (P < .02); similarly, lower NAA was found anteriorly in the ipsilateral hippocampus in eight of nine patients (P = .05) and also in the contralateral hippocampus in seven of nine patients (P = .06). In contrast, most control subjects and patients had higher Ch in the anterior hippocampal tissue than in the posterior tissue. The Ch slope was negative from anterior to posterior in nine of 10 control subjects (P < .001), in eight of nine patients in ipsilateral hippocampal tissue (P < .02), and in seven of nine patients in contralateral tissue (P < .03). For Cr, there was no consistent trend for control subjects or patients; the slope was negative in nine of 10 control subjects (P < .01) and in seven of nine patients in ipsilateral hippocampal tissue (P < .04) but in only four of nine patients in contralateral tissue (P = .94). The slopes for the three metabolites and NAA/(Cr + Ch) were not significantly different between control subjects and patients.

We calculated the mean metabolite signal and metabolite ratio from each position in hippocampal tissue for control subjects and separately for patients from ipsi- and contralateral hippocampal regions (Table 3). Figure 4 shows plots for control subjects from the voxel position within hippocampal tissue versus mean values of NAA/(Cr + Ch) (Fig 4a) and NAA (Fig 4b). NAA/(Cr + Ch) values increased from anterior to posterior hippocampal tissue (R = 0.98 with linear regression). The difference between the anterior part (voxels 1, 2) and the posterior part (voxels 4–6) is greater than 25%. The NAA signal in hippocampal tissue of control subjects also shows an increase from anterior to posterior (R = 0.56). The difference between anterior and posterior is about 12%. Similar to findings in control subjects, the ipsi- and contralateral hippocampi in patients also had lower NAA/(Cr + Ch) values anteriorly than posteriorly (Table 3). The AP slope (assuming a linear relationship) for patients was similar to that for control subjects. At every position in hippocampal tissue, however, ipsilateral and also contralateral NAA and NAA/(Cr + Ch) values were lower than the control values at the same position. Compared with findings in control subjects, the average reductions in patients were about 17% for ipsilateral and about 10% for contralateral NAA and NAA/(Cr + Ch) values.


View this table:
[in this window]
[in a new window]
  		TABLE 3. Average Metabolic Signals for Each Position in Hippocampal Tissue
 


View larger version (12K):
[in this window]
[in a new window]
  		Figure 4a. Plots measured with the standard PRESS sequence of mean ({block}) (a) NAA/(Cr + Ch) and (b) NAA (in arbitrary units) for each position in the hippocampal tissue (1-6 in Fig 2a) of all control subjects. Values for NAA/(Cr + Ch) and, to a lesser extent, NAA increase from anterior to posterior. Error bars indicate SD.

 


View larger version (11K):
[in this window]
[in a new window]
  		Figure 4b. Plots measured with the standard PRESS sequence of mean ({block}) (a) NAA/(Cr + Ch) and (b) NAA (in arbitrary units) for each position in the hippocampal tissue (1-6 in Fig 2a) of all control subjects. Values for NAA/(Cr + Ch) and, to a lesser extent, NAA increase from anterior to posterior. Error bars indicate SD.

 
In the very anterior part of the hippocampus in all patients, ipsilateral mean NAA/(Cr + Ch) was reduced more than 1 SD below the control mean, and ipsilateral NAA values were lower than contralateral values in all patients. In a few patients, however, the spectral quality was poor in the very anterior part of the hippocampus, and the values could not be obtained in this region. Also, in posterior parts of the hippocampus, ipsilateral NAA/(Cr + Ch) values were reduced in most patients compared with contralateral values and with values in control subjects (eg, at positions 4 and 5, NAA/[Cr + Ch] values in eight of nine patients were reduced more than 1 SD below the control mean).

In a comparison of ipsilateral values in patients with those in control subjects at all positions from anterior to posterior, we found NAA/(Cr + Ch) values were significantly reduced in patients (P values from <.03 to <.001, two-tailed t test) at all positions except the most posterior voxel.

Figure 5 shows NAA/(Cr + Ch) values on a logarithmic scale at different positions in hippocampal tissue for patients and control subjects, with a linear regression of the data. The plot shows a greater proportional reduction anteriorly than posteriorly of ipsilateral values in patients compared with contralateral tissue and values in control subjects. This suggests that findings in the anterior portion of the hippocampus might provide a better separation between patients and control subjects. Therefore, we calculated the effect size (difference between means divided by mean SD) for each position in hippocampal tissue between control subjects and ipsilateral tissue in patients (Table 3). The greatest effect sizes for NAA (2.9) and NAA/(Cr + Ch) (3.0) were found in the very anterior part of the hippocampal tissue (position 1). At position 2, however, the effect sizes for NAA and NAA/(Cr + Ch) were reduced to 1.5 and 1.1, respectively. From position 2 to position 5, the effect sizes for NAA and NAA/(Cr + Ch) increased owing to smaller SD. At position 6, the effect sizes were very small (0.5 and 0.6 for NAA and NAA/[Cr + Ch], respectively). Except for the effect size of Ch in the very anterior part (1.2), the effect sizes of Cr and Ch were below 0.6 at all other positions.



View larger version (16K):
[in this window]
[in a new window]
  		Figure 5. Semilogarithmic plot of NAA/(Cr + Ch) values versus different positions in hippocampal tissue (1-6 in Fig 2a) for control subjects and patients (contra = contralateral, ipsi = ipsilateral) with a linear regression of the data. At every position, ipsilateral and contralateral NAA/(Cr + Ch) values in patients are lower than those in control subjects. In comparison with the contralateral values and those in control subjects, the proportional reduction in ipsilateral values is greater anteriorly than posteriorly.

 
Results with PRESS with 90° Pulse in Coronal Orientation (four control subjects)
Similar to measurements with the standard sequence, a linear regression of NAA and NAA/(Cr + Ch) versus the position in hippocampal tissue for each patient showed a positive slope from anterior to posterior: NAA, 8.0 ± 6.0, P = .075; NAA/(Cr + Ch), 0.064 ± 0.013, P < .003 (Table 2). The Ch slope was negative from anterior to posterior for all four control subjects (-6.6 ± 8.6, P = .22). For Cr, the slope was negative in two of four control subjects (0.2 ± 4.2, P = .92). The NAA slope was greater, but not significantly, compared with the slope in control subjects measured with the standard sequence. The difference might be due to an imperfect correction for the pulse profile used for the regular sequence close to the edges of the PRESS box (voxels 1 and 6), where the deviation from an ideal rectangular shape is greatest. Omission of these two voxels from the slope determination with the regular sequence leads to a slope similar to that determined with the sequence with the 90° pulse applied in the coronal orientation. However, the NAA/(Cr + Ch) slope was similar to that determined with the standard sequence.

Figure 6 shows the mean NAA and NAA/(Cr + Ch) values calculated for each position in hippocampal tissue for the four control subjects. The difference between the anterior and posterior parts in hippocampal tissue was greater than 25% for both NAA and NAA/(Cr + Ch).



View larger version (12K):
[in this window]
[in a new window]
  		Figure 6a. Plots measured with the PRESS sequence with the 90° pulse in the AP direction of mean ({block}) (a) NAA/(Cr + Ch) and (b) NAA (in arbitrary units) for each position within hippocampal tissue (1-6 in Fig 2a) in four control subjects. As in Figure 4, the NAA/(Cr + Ch) and the NAA values increase from anterior to posterior. Error bars indicate SD.

 


View larger version (12K):
[in this window]
[in a new window]
  		Figure 6b. Plots measured with the PRESS sequence with the 90° pulse in the AP direction of mean ({block}) (a) NAA/(Cr + Ch) and (b) NAA (in arbitrary units) for each position within hippocampal tissue (1-6 in Fig 2a) in four control subjects. As in Figure 4, the NAA/(Cr + Ch) and the NAA values increase from anterior to posterior. Error bars indicate SD.

 
T2 Determination (three control subjects)
To examine the possibility that differences in metabolite T2 values might be responsible for the AP differences, we measured individual T2 values for NAA, Cr, and Ch in three control subjects in each position in hippocampal tissue (Fig 7). Table 4 lists the mean metabolite T2 values for each position. The mean T2 values over the entire region were the following: NAA, 355 msec ± 22; Cr, 277 msec ± 30; Ch, 370 msec ± 58. The T2 values for Ch were scattered between 250 and 540 msec, whereas the T2 values for Cr and NAA showed less variation. The T2 values for all three metabolites showed a trend for higher values in posterior compared with anterior hippocampal tissue; however, the trend was significant for only the T2 values of NAA (P < .02).



View larger version (11K):
[in this window]
[in a new window]
  		Figure 7. Plots indicate T2 values for three control subjects ({block}, {triangleup}, x) for each position in hippocampal tissue (1-6 in Fig 2a), with linear regression.

 

View this table:
[in this window]
[in a new window]
  		TABLE 4. Metabolite T2 Values for Each Position in Hippocampal Tissue
 
By arbitrarily assuming a linear relationship between the T2 values and the position in hippocampal tissue, the slopes for the three metabolite T2 values were calculated. The slopes of the three metabolites were of comparable magnitude: NAA, 10.9 ± 4.0; Ch, 13.1 ± 11.3; Cr, 7.8 ± 6.7. The T2 difference for NAA between the anterior and posterior parts of hippocampal tissue was about 13%. With use of these slopes to calculate AP metabolite differences, T2 differences may account for approximately 6% of the AP differences in NAA. Since the T2 differences for all three metabolites point in the same direction, however, T2 differences can account for only about 1% of the AP differences in NAA/(Cr + Ch).


   DISCUSSION
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
References
 
Our results show that in control subjects and patients, NAA/(Cr + Ch) values in hippocampal tissue were significantly higher posteriorly than anteriorly. This increase was due to an increase in NAA and a decrease in Cr and Ch from anterior to posterior. The extent of the AP changes exceeded 25% and was therefore greater than the difference found in some studies (1–7) between patients and control subjects or between ipsilateral and contralateral hippocampal tissue in patients with TLE. In patients with mesial TLE, NAA and NAA/(Cr + Ch) values were lower in each row in ipsilateral tissue (Table 3) than they were in control subjects, with an average reduction in NAA and NAA/(Cr + Ch) of 17% in ipsilateral tissue and about 10% in contralateral tissue. The proportional reduction was greatest in the anterior part of hippocampal tissue. This demonstrates that exact voxel selection in the hippocampus is of paramount importance in a comparison of hippocampi results from different hemispheres or of patients with control subjects. This may be a particular problem for single-volume MR spectroscopic studies. The use of MR spectroscopic imaging is advantageous in that it allows postacquisition positioning.

In previous studies, MR spectroscopy has been shown to lateralize the seizure focus in many patients with TLE (1–7). However, there remained a considerable number of patients in whom no spectral abnormalities could be found or they were falsely lateralized. We found that the proportional reduction in NAA/(Cr + Ch) in ipsilateral hippocampal tissue of patients with mesial TLE was greatest in the very anterior part, providing the greatest effect size between control subjects and patients in ipsilateral tissue. This is in agreement with histologic findings that revealed greatest neuronal loss in anterior portions of the hippocampus in patients with mesial TLE (11,13,14). This finding would suggest the use of spectra mainly from the anterior regions to detect the seizure focus. However, smaller SDs in the posterior hippocampus due to a better spectral quality with better shimming also caused relatively high effect sizes posteriorly. To use the greater reductions found anteriorly for better separation between patients and control subjects, improvements would be necessary in the spectral quality (eg, by means of better shimming).

The differences between control subjects and patients in ipsi- and contralateral tissue, as well as the effect sizes, were smallest in the most posterior portion of the hippocampus, which indicates that this region is least affected by the disease or the contributions from adjacent tissue are greatest. The results emphasize the importance of exact voxel selection when right and left hippocampal spectra are compared or when patients are compared with control subjects. Exact voxel selection reduces the variability of the results and also prevents spurious differences between patients and control subjects.

Technical Artifacts as an Explanation for AP Differences
Radio-frequency pulse profiles.—Pulse profiles with a shape different from trapezoidal directly affect signal intensity and to a smaller extent metabolite ratios as a result of chemical-shift displacements. Phantom measurements showed that the 180° pulse used in the PRESS sequence was not trapezoidal. However, we corrected the signal areas measured for the pulse profile. Furthermore, results in the four control subjects who underwent MR spectroscopic imaging with the 90° pulse in the AP direction were equivalent to the corrected results with the regular sequence without the necessity for profile corrections. Therefore, the AP differences cannot be due to imperfect pulse profiles.

Artifacts due to section displacement.—The applied MR spectroscopic imaging sequence excited a section that included both hippocampi. Because of the chemical-shift difference between NAA and Cr or Ch of about 1 ppm, the application of a section-selective pulse causes the Cr and Ch signals and the NAA signal to arise from slightly different section locations and might, thus, produce artifacts. However, we used a gradient strength of 3 mT/m to be applied with the section-selective pulse. At a field strength of 1.5 T, this results in a difference in the section position for NAA and Cr or Ch of approximately 0.5 mm. With a section thickness of 15 mm, we consider this difference too small to have a significant effect on the results.

Artifacts due to B0 inhomogeneity.—In the presence of B0 inhomogeneity, a frequency-selective pulse excites a section that can be warped (eg, have a wavy-appearing surface) and in the case of the hippocampus could bend increasingly at one end compared with the other. In our measurements, however, the frequency shift across the section was always below 10 Hz and did not show a consistent pattern for different subjects. With a gradient strength of 3 mT/m for the section-selective pulse, this results in only very small differences in the section position. Furthermore, although the shimming was usually better in posterior regions of the hippocampus than in the anterior regions, the difference in line width between spectra from anterior and posterior regions was not great enough to substantially change the thickness of the section.

Artifacts due to the spatial response function.—Owing to the breadth of the point spread function, the metabolite response is really an average over a significant volume of tissue (Fig 2) (16). Conceivably, this could lead to an inaccurate representation of metabolite levels. As long as the metabolite levels are locally uniform and change only slowly in the AP direction, however, the MR spectroscopic imaging result provides an accurate mapping of these levels. The applied spatial apodization was enough to markedly reduce the ripples of the point spread function so that only nearby tissue contributed significantly to the voxel intensity.

T2 Changes as an Explanation for AP Differences
Mean T2 values obtained in this study for NAA and Ch were comparable to those published for other brain regions (17–23) (Table 4). Mean T2 obtained for Cr was slightly longer than published values. Differences in T2 between the anterior and posterior regions of hippocampal tissue were only moderate, though significant for NAA. The shorter T2 of NAA in anterior hippocampal tissue, however, can account for only 6% of the differences found in NAA and for only 1% of the differences found in NAA/(Cr + Ch). Therefore, although only three T2 measurements were performed, it is unlikely that T2 differences account for the AP differences found in patients and control subjects.

Hippocampal Anatomy as an Explanation for the AP Differences
The thickness of the hippocampus is only several millimeters and therefore is smaller than the section thickness in this and to our knowledge all other published MR spectroscopic imaging studies of hippocampal tissue (1–7). Since the hippocampus is thicker anteriorly than posteriorly, contributions from adjacent tissue could produce the detected metabolite differences depending on the position of the voxel in the AP direction. Recently, a segmentation study of hippocampal tissue in control subjects was performed in our laboratory (Schuff NS, unpublished data, 1999). Findings in that study revealed posteroanterior differences in the voxel composition of white matter and gray matter at MR spectroscopic imaging. Voxels selected from the pes hippocampi had a gray matter contribution about 10% higher than that in voxels from posterior regions, presumably owing to differences in the thickness of the hippocampus. Furthermore, differences in metabolite concentrations and metabolite ratios between gray matter and white matter were found in several studies (24–28). The reported differences between NAA and NAA relative to Cr, Ch, or both in pure gray matter and in pure white matter were quite different, depending on the method used and the brain region investigated. However, the ratios of metabolites NAA to Cr, Ch, or both in pure gray matter to those in pure white matter were in the range of 0.7–1.5. This metabolite difference between gray matter and white matter, together with an estimated difference in voxel composition of about 10% for gray matter and white matter between anterior and posterior hippocampal portions, could account for AP metabolite differences of as much as 5%. This estimation is speculative, however, because metabolite concentrations in hippocampal pure white matter and pure gray matter are not known. Differences in voxel composition between gray matter and white matter could possibly account for some of the AP metabolite differences found in this study. However, the differences in white matter and gray matter contribution to the voxel composition in hippocampal tissue are only moderate and, therefore, probably not solely responsible for an AP difference that exceeded 25%.

Number of Neuronal Cells as an Explanation for AP Differences
In a histologic study, Mouritzen Dam (10) found an increase from anterior to posterior in the number of pyramidal cells in all subfields of the hippocampus, whereas the number of granule cells decreased in the AP direction. Babb et al (11) found that in most subfields, the number of neuronal cells was greater posteriorly than anteriorly, although these results were not statistically significant. NAA is located exclusively in neurons and therefore can serve as a neuronal marker. Although a greater number of neuronal cells is not necessarily accompanied by higher NAA levels, the lower NAA level detected in anterior hippocampal tissue might be due to fewer neuronal cells anteriorly as suggested by findings in the histologic studies.

In conclusion, we found an AP gradient of NAA/(Cr + Ch) in hippocampal tissue of control subjects and patients with mesial TLE, with lower values anteriorly than posteriorly, that was most likely due to a combination of greater numbers of neurons in the posterior hippocampus and partial volume effects. The differences between control subjects and patients with mesial TLE were greatest in the anterior hippocampus. Thus, this finding may lead to improved methods for seizure focus localization in mesial TLE. These results demonstrate the importance of positional control for exact voxel positioning and selection in MR spectroscopic studies.


   Footnotes
 
Abbreviations: AP = anteroposterior Ch = choline-containing compounds Cr = creatine plus phosphocreatine NAA = N-acetylaspartate NAA/(Cr + Ch) = ratio of NAA to Cr   plus Ch PRESS = point-resolved spatially   localized spectroscopy TLE = temporal lobe epilepsy

Author contributions: Guarantors of integrity of entire study, K.D.L., M.W.W.; study concepts and design, P.V., G.B.M., K.D.L., M.W.W.; definition of intellectual content, P.V., G.B.M., K.D.L., M.W.W.; literature research, P.V., K.D.L.; clinical studies, P.V., G.B.M., K.D.L.; experimental studies, P.V.; data acquisition, P.V., G.B.M.; data analysis, P.V.; statistical analysis, P.V.; manuscript preparation and editing, P.V.; manuscript review, K.D.L., M.W.W., G.B.M.


   References
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 

  1. Spencer DD, ed. Workshop on magnetic resonance techniques and epilepsy research. Magn Reson Imaging 1995; 13:1171-1211.[Medline]
  2. Hugg JW, Laxer KD, Matson GB, Maudsley AA, Weiner MW. Neuron loss localizes human temporal lobe epilepsy by in vivo proton magnetic resonance imaging. Ann Neurol 1993; 34:788-794.[Medline]
  3. Ende GR, Laxer KD, Knowlton RC, et al. Temporal lobe epilepsy: bilateral hippocampal metabolite changes revealed at proton MR spectroscopic imaging. Radiology 1997; 202:809-817.[Abstract/Free Full Text]
  4. Vermathen P, Ende G, Laxer KD, Knowlton RC, Matson GB, Weiner MW. Hippocampal NAA in neocortical epilepsy and mesial temporal lobe epilepsy. Ann Neurol 1997; 42:194-199.[Medline]
  5. Cendes F, Caramanos Z, Andermann F, Dubeau F, Arnold DL. Proton magnetic resonance spectroscopic imaging and magnetic resonance imaging volumetry in the lateralization of temporal lobe epilepsy: a series of 100 patients. Ann Neurol 1997; 42:737-746.[Medline]
  6. Connelly A, Jackson GD, Duncan JS, King MD, Gadian DG. Magnetic resonance spectroscopy in temporal lobe epilepsy. Neurology 1994; 44:1411-1417.[Abstract/Free Full Text]
  7. Ng TC, Comair YG, Xue M, et al. Temporal lobe epilepsy: presurgical localization with proton chemical-shift imaging. Radiology 1994; 193:465-472.[Abstract/Free Full Text]
  8. Schuff N, Amend D, Ezekiel F, et al. Changes of hippocampal N-acetyl-aspartate and volume in Alzheimer's disease: a proton MR spectroscopic imaging and MRI study. Neurology 1997; 49:1513-1521.[Abstract/Free Full Text]
  9. Block W, Treaber F, Kuhl CK, et al. 1H-MR spectroscopic imaging in patients with clinically diagnosed Alzheimer's disease. Rofo Fortschr Geb Roentgenstr Neuen Bildgeb Verfahr 1995; 163:230-237.[Medline]
  10. Mouritzen Dam A. The density of neurons in the human hippocampus. Neuropathol Appl Neurobiol 1979; 5:249-264.[Medline]
  11. Babb TL, Brown WJ, Pretorius J, Davenport C, Lieb JP, Crandall PH. Temporal lobe volumetric cell densities in temporal lobe epilepsy. Epilepsia 1984; 25:729-740.[Medline]
  12. Woermann FG, Symms MR, Birnie K, Barker GJ, Duncan JS. Measurement of hippocampal T2 relaxation time: a new fast FLAIR dual echo technique (abstr). Proceedings of the Fifth Meeting of the International Society for Magnetic Resonance in Medicine. Berkeley, Calif: International Society for Magnetic Resonance in Medicine, 1997; 212.
  13. Mouritzen Dam A. Epilepsy and neuron loss in the hippocampus. Epilepsia 1980; 21:617-629.[Medline]
  14. O'Connor WM, Masukawa L, Freese A, Sperling MR, French JA, O'Connor MJ. Hippocampal cell distributions in temporal lobe epilepsy: a comparison between patients with and without an early risk factor. Epilepsia 1996; 37:440-449.[Medline]
  15. Michaelis T, Merboldt KD, Bruhn H, Hänicke W, Frahm J. Absolute concentration of metabolites in the adult human brain in vivo: quantification of localized proton MR spectra. Radiology 1993; 187:219-227.[Abstract/Free Full Text]
  16. Hugg JW, Maudsley AA, Weiner MW, Matson GB. Comparison of k-space sampling schemes for multidimensional MR spectroscopic imaging. Magn Reson Med 1996; 36:469-473.[Medline]
  17. Hennig J, Pfister H, Ernst T, Ott D. Direct absolute quantification of metabolites in the human brain with in vivo localized proton spectroscopy. NMR Biomed 1992; 5:193-199.[Medline]
  18. Barker PB, Soher BJ, Blackband SJ, Chatham JC, Mathews VP, Bryan RN. Quantitation of proton NMR spectra of the human brain using tissue water as an internal concentration reference. NMR Biomed 1993; 6:89-94.[Medline]
  19. Christiansen P, Henriksen O, Stubgaard M, Gideon P, Larsson HB. In vivo quantification of brain metabolites by 1H-MRS using water as an internal standard. Magn Reson Imaging 1993; 11:107-118.[Medline]
  20. Kreis R, Ernst T, Ross BD. Development of the human brain: in vivo quantification of metabolite and water content with proton magnetic resonance spectroscopy. Magn Reson Med 1993; 30:424-437.[Medline]
  21. Danielsen ER, Henriksen O. Absolute quantitative proton NMR spectroscopy based on the amplitude of the local water suppression pulse: quantification of brain water and metabolites. NMR Biomed 1994; 7:311-318.[Medline]
  22. Hagberg G, Seelig J. Localized proton spectroscopy of the human brain: comparison of different methods for absolute quantitation (abstr). Proceedings of the 12th Annual Scientific Meeting of the Society of Magnetic Resonance in Medicine. Berkeley, Calif: Society of Magnetic Resonance in Medicine, 1993; 979.
  23. Hänicke W, Michaelis T, Merboldt KD, Frahm J. On the use of fully automated data analysis method for in vivo MRS (abstr) In: Proceedings of the 12th Annual Scientific Meeting of the Society of Magnetic Resonance in Medicine. Berkeley, Calif: Society of Magnetic Resonance in Medicine, 1993; 977.
  24. Lim KO, Spielman DM. Estimating NAA in cortical gray matter with applications for measuring changes due to aging. Magn Reson Med 1997; 37:372-377.[Medline]
  25. Doyle TJ, Bedell BJ, Narayana PA. Relative concentrations of proton MR visible neurochemicals in gray and white matter in human brain. Magn Reson Med 1995; 33:755-759.[Medline]
  26. Hetherington HP, Pan JW, Mason GF, et al. Quantitative 1H spectroscopic imaging of human brain at 4.1 T using image segmentation. Magn Reson Med 1996; 36:21-29.[Medline]
  27. Schuff N, Soher BJ, Young K, et al. Regression and histogram analysis of 1H MR spectroscopic imaging data (abstr) In: Proceedings of the Sixth Meeting of the International Society for Magnetic Resonance in Medicine. Berkeley, Calif: International Society for Magnetic Resonance in Medicine, 1998; 28.
  28. Schuff N, Amend DL, Capizzano A, et al. NAA reductions in parietal and frontal cortex of Alzheimer's disease (abstr). Proceedings of the Sixth Meeting of the International Society for Magnetic Resonance in Medicine. Berkeley, Calif: International Society for Magnetic Resonance in Medicine, 1998; 729.



This article has been cited by other articles:


Home page
 RadiologyHome page
K. G. King, L. Glodzik, S. Liu, J. S. Babb, M. J. de Leon, and O. Gonen
Anteroposterior Hippocampal Metabolic Heterogeneity: Three-dimensional Multivoxel Proton 1H MR Spectroscopic Imaging--Initial Findings
Radiology, October 1, 2008; 249(1): 242 - 250.
[Abstract] [Full Text] [PDF]

Home page
 NeurologyHome page
H. P. Hetherington, R. I. Kuzniecky, K. Vives, O. Devinsky, S. Pacia, D. Luciano, B. Vasquez, S. Haut, D. D. Spencer, and J. W. Pan
A subcortical network of dysfunction in TLE measured by magnetic resonance spectroscopy
Neurology, December 11, 2007; 69(24): 2256 - 2265.
[Abstract] [Full Text] [PDF]

Home page
 RadiologyHome page
P. Vermathen, K. D. Laxer, N. Schuff, G. B. Matson, and M. W. Weiner
Evidence of Neuronal Injury Outside the Medial Temporal Lobe in Temporal Lobe Epilepsy: N-Acetylaspartate Concentration Reductions Detected with Multisection Proton MR Spectroscopic Imaging--Initial Experience
Radiology, January 1, 2003; 226(1): 195 - 202.
[Abstract] [Full Text] [PDF]

Home page
 Arch NeurolHome page
M. Feichtinger, E. Pauli, I. Schafer, K. W. Eberhardt, B. Tomandl, J. Huk, and H. Stefan
Ictal Fear in Temporal Lobe Epilepsy: Surgical Outcome and Focal Hippocampal Changes Revealed by Proton Magnetic Resonance Spectroscopy Imaging
Arch Neurol, May 1, 2001; 58(5): 771 - 777.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Roentgenol.Home page
Y.-Y. Hsu, M.-C. Chen, K.-E. Lim, and C. Chang
Reproducibility of Hippocampal Single-Voxel Proton MR Spectroscopy and Chemical Shift Imaging
Am. J. Roentgenol., February 1, 2001; 176(2): 529 - 536.
[Abstract] [Full Text] [PDF]

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 Vermathen, P.
Right arrow Articles by Weiner, M. W.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Vermathen, P.
Right arrow Articles by Weiner, M. W.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
RADIOLOGY RADIOGRAPHICS RSNA JOURNALS ONLINE