Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
J Magn Reson Imaging. Author manuscript; available in PMC 2010 March 4.
Published in final edited form as:
PMCID: PMC2832220

Fast 3D 1H MRSI of the Corticospinal Tract in Pediatric Brain



To develop a 1H magnetic resonance spectroscopic imaging (MRSI) sequence that can be used to image infants/children at 3T and by combining it with diffusion tensor imaging (DTI) tractography, extract relevant metabolic information corresponding to the corticospinal tract (CST).

Materials and Methods:

A fast 3D MRSI sequence was developed for pediatric neuroimaging at 3T using spiral k-space readout and dual band RF pulses (32 × 32 × 8 cm field of view [FOV], 1 cc iso-resolution, TR/TE = 1500/130, 6:24 minute scan). Using DTI tractography to identify the motor tracts, spectra were extracted from the CSTs and quantified. Initial data from infants/children with suspected motor delay (n = 5) and age-matched controls (n = 3) were collected and N-acetylaspartate (NAA) ratios were quantified.


The average signal-to-noise ratio of the NAA peak from the studies was ≈22. Metabolite profiles were successfully acquired from the CST by using DTI tractography. Decreased NAA ratios in those with motor delay compared to controls of ≈10% at the CST were observed.


A fast and robust 3D MRSI technique targeted for pediatric neuroimaging has been developed. By combining with DTI tractography, metabolic information from the CSTs can be retrieved and estimated. By combining DTI and 3D MRSI, spectral information from various tracts can be obtained and processed.

Keywords: volumetric MRSI, fast MRSI, CST tractography

Single voxel spectroscopy has been used to access various diseases in the pediatric brain. The use of a localizing technique via PRESS (1) or STEAM (2) enables collecting information from a rectangular region of interest (ROI). For many cases this information suffices in determining the desired relevant metabolic information if the ROI can be well defined. However, there is need to expand the ROI to two dimensions (2D) or even three dimensions (3D) for certain applications. A wide scope of studies falls into this category where the spatial information can be an important factor. In addition, another realistic reason for multidimensional application can be that the region to acquire the spectroscopic information from might not be well characterized from the scout or other anatomical images. For example, when the targeted ROI is the corticospinal tract (CST), this region might not be readily visible in the conventional scout T1- or T2-weighted images when imaging infants.

Multivoxel spectroscopy, 2D or 3D, can be used to overcome these limitations while increasing the spatial coverage. Conventional phase-encoded multivoxel spectroscopy is a robust and simple method but it accompanies an increase in scan time that is proportional to the number of voxels resolved. For 3D applications this can amount to a significant scan time, limiting its use for clinical applications. For pediatric exams the long scan time can be crucial due to the fact that they are more vulnerable to move during the longer scans. For this reason there is a strong desire for a faster scanning method for this population. Sedation is often used which decreases susceptibility to motion, but a faster scan would still be desirable to reduce the overall sedation time.

The purpose of this study was to develop a 1H magnetic resonance spectroscopic imaging (MRSI) sequence that can be used to image infants/children at 3T and by combining it with diffusion tensor imaging (DTI) tractography, extract relevant metabolic information corresponding to the CST. By linking spectroscopic imaging with DTI to extract voxels corresponding to the CST, comprehensive metabolic assessment of the CST could be obtained. To demonstrate the potential usefulness of this approach, data were acquired from infant/children with developmental motor delay and controls and N-acetylaspartate (NAA) level ratios from the CSTs were compared. Previous studies have indicated the rapid metabolic alterations during development and its differences from abnormal developmental conditions (3-5). Polynomial, logarithmic, or exponential fits have been used to describe the changes for normal development. Specifically, these studies have noted the increase in NAA levels as normal development occurs.

The main objective of this study was not to derive a clinical conclusion but rather to show the potential clinical utility of this approach. In demonstrating this feature we follow an approach similar to that of a clinical study. It should be noted that the limited number of subjects for this study does not validate any clinical suggestions. In the next section we illustrate the implementation of a fast 3D MRSI pulse sequence for use in pediatric patients. We apply the pulse sequence to pediatric patients with delayed motor development.


Fast 3D 1H MRSI Pulse Sequence

A fast 3D 1H MRSI sequence using PRESS excitation was developed for pediatric brain imaging at 3T (Fig. 1). Two identical dual band spectral spatial spin echo pulses, designed to excite 1% of the water signal and suppress lipids while passing the NAA, creatine (Cr), and choline (Cho) resonances was optimized for 3T brain imaging and used instead of the conventional PRESS 180° pulses (6).

Figure 1
Pulse sequence diagram. A dual band spectral spatial spin echo pulse is used. Spiral readout gradients are played for fast data acquisition and full coverage. VSS (very selective saturation) pulses are used to further suppress signals coming from the ...

In-plane (kx and ky) spatial information was encoded using spiral k-space readout trajectories while chemical shift information was encoded by repeating the spiral trajectories. The spiral trajectory was designed to cover a 32 × 32 matrix over a 32 × 32 cm field of view (FOV) (kxmax = kymax = 0.5 cm−1). The relatively large FOV was chosen so that ringing artifacts, mostly arising from the residual lipid components, do not alias into the ROI. The resulting spiral trajectory had a length of 1.024 msec, thereby enabling a spectral bandwidth of 976 Hz using a 4 μs sampling receiver. During a single readout, 512 spirals were continuously played out, which required a readout length of 524 msec. To satisfy the required imaging parameter of a 32 × 32 spatial coverage, 16 spatial interleaving of the spirals were used. A detailed depiction of spiral-based k-space readouts for spectroscopic imaging has been previously described (7). After data acquisition, data were reconstructed using gridding. A 4 Hz Lorentzian apodization function was used in the spectral domain to increase the spectral signal-to-noise ratio (SNR). No windowing was performed in the spatial frequency domain. Since no zero-filling was used the nominal spatial resolution, as measured by the full-width half-maximum value of the point spread function, should be 1 cc with iso-resolution. To verify this, simulation of the point spread function was performed by assuming an impulsive object in the image domain and reconstructing with the method described above to validate this.

Eight phase encodes (kz) were also used to acquire 3D information with nominal resolution of 1 cm. Spatial saturation pulses were used to exclude excitation of the subcutaneous lipid regions. The PRESS selection box was chosen as large as possible within brain tissue. This was to reduce lipid signals from being excited while collecting necessary metabolic information. Although spectral spatial refocusing pulses and spatial saturation pulses were both used, additional suppression of the lipid signal was typically required. The PRESS selection box size was normally 10 × 10 × 6 cm located within the pediatric brain. Data were collected using an eight-channel receiver coil providing increased SNR. The imaging parameters were as follows: 32 × 32 × 8 matrix over a 32 × 32 × 8 cm FOV, 1 cc iso-resolution, TR/TE = 1500/130, 980 Hz spectral bandwidth, 2 signal averages, 6:24 minute total scan time. Data were collected on a GE 3T (General Electric Healthcare Technologies, Milwaukee, WI) scanner located at our institution.

Diffusion and Standard Imaging Sequences

The diffusion tensor imaging data is acquired in 3.1 minutes using a multirepetition, single-shot echo planar sequence with 15 gradient directions, b = 0 and 1000 s/mm2, TR = 11.5 seconds, TE = 61.8 msec, 1 repetition, FOV = 28 × 14 cm, matrix = 256 × 128, slice thickness = 2.2 mm with no gap, 167 kHz readout bandwidth, and no ramp sampling. A sense factor of two was used to reduce the geometric distortions. The resulting in-plane resolution was 1.1 mm. Interleaved axial slices were acquired for full brain coverage.

Additional standard imaging sequences included: 1) T1-weighted sagittal and axial spin echo images (4 mm thickness) using repetition time (TR) = 500 msec, echo time (TE) = 11 msec, 1 excitation, and 192 × 256 acquisition matrix; and 2) axial 3D fast spin echo (FSE) images using TR = 4000 msec, TE = 85 msec, echo train length (ETL) = 16, bandwidth (BW) = 15, 1.5 mm slice thickness, 6 locations/slab, and 192 × 256 acquisition matrix; 3) Coronal volumetric 3D gradient echo images with radiofrequency spoiling (SPGR) images (1 mm thickness) with TR = 36 msec, TR = 9 msec, flip angle = 35°, NEX = 1.

Data Processing and Analysis

Spectroscopic data reconstruction was performed for each coil by apodizing using a 4-Hz Gaussian filter followed by gridding and inverse fast Fourier transform (FFT). Using residual water signal from the dual band sequence, zero and first-order phasing was applied to every voxel spectrum. Spectra combination from the individual coils was accomplished by weighting according to the phased water signal amplitude to optimize for SNR.

Data quantification was performed following the reconstruction. Using in-house software for DTI fiber tracking (8), the CSTs were identified. Voxels from the spectroscopic image corresponding to the CSTs were extracted. Voxel shifting was performed prior to the extraction to center the spectra on the CST whenever possible to reduce partial volume errors. Spectral quantification of the extracted voxels was performed using LCModel fitting (9). These quantitative values corresponding to the location of the CSTs were summed and averaged. Cramer–Rao bounds of the LCModel fit beyond 10% were excluded from the analysis.

For data analysis, NAA ratios (NAA/Cr and NAA/Cho) from the infants with motor delay were compared with those from the controls. For controls a logarithmic function was fit to the spectroscopic data as a function of age to emulate maturation.

Data Acquisition

Patients and controls were recruited from a group of patients already scheduled for MRI at our outpatient facility. Prior to the scan, all parents of the subjects were approached and, if agreeable to participation, were consented for the protocol, which was approved by our Institutional Review Board (IRB). The five subject infants and children were being studied for motor delay, whereas the three controls, of similar ages, had normal motor function and were being studied for esotropia, language delay, or evaluation of a facial hemangioma. A more complete description of the patient population is listed in Table 1. Spectroscopic data were acquired using the newly developed 3D MRSI pulse sequence after the routine imaging sequences had been completed.

Table 1
Patient Population Information for 1H MRSI Study of Motor Delay Infants


By using DTI tractography to identify the CST we were able to extract and analyze spectroscopic information on most of the supratentorial CST from the 3D MRSI dataset and test whether NAA ratios are reduced in the CST compared to normal age-matched controls.

Figure 2 shows representative image slices from a motor delayed child and its corresponding spectra prior to any voxel shifting. Spectral regions surrounding the CSTs are shown. Good spectral quality data can be seen using the proposed sequence. The average SNR of the NAA metabolite from the gray matter, measured by dividing the signal amplitude by the standard deviation of the noise, was ≈22 ± 2.5. This was slightly higher than the SNRs we normally achieved from adult studies, which were ≈20 ± 2.9. The smaller size of the head of infants accounts for less noise seen by the receiver coils, thereby increasing the SNR. Estimated standard deviation from the LCModel fit typically resulted in ≈5–7% (NAA: 6.04 ± 2, Cr: 6.35 ± 1.66, Cho: 6.06 ± 1.79) and below 15% for the metabolites of interest. Simulation of the point spread function resulted in a 1 cc resolution (data not shown).

Figure 2
Representative slices (b = 0 images from the DTI acquisitions) and spectra from a motor-delayed child. The left CST is identified via DTI tractography and the regions, marked in white spots, are indicated by the arrows in the image figures. Spectra are ...

Figure 3 shows an example of a resulting DTI tractography and summed MRSI spectra extracted from the CSTs from an age-matched control (control #2) and a motor-delayed child (patient #2). Metabolite ratios show a relatively decreased NAA value for the motor-delayed infant compared to age-matched control. While the tracts identified (shown in yellow) represent the CST, the spectral analysis was performed within those tracts as indicated by the anatomic image representing the bottom-most plane and blue patch representing the upper-most plane. These regional constraints were due to the spectroscopic acquisition limitations rather than from limitations from the DTI analysis. Extending beyond this limit superiorly resulted in bad spectral quality due to heavy lipid contamination even with the dual band spectral spatial RF pulses. In the inferior direction, line broadening due to B0 inhomogeneity from the skull base and sinuses limited data quantification.

Figure 3
Combined DTI and MRSI pediatric studies of the CSTs. Center: Example DTI tractography identifying the CSTs (the blue cut-plane indicated the top of the PRESS box while the anatomic image represents the bottom plane included in the analysis). Left and ...

The NAA/Cho and NAA/Cr ratios obtained from the CSTs of all the study subjects are given in Fig. 4. The three control data are fit to a logarithmic curve to assume nominal growth curves. It can be seen that all of the motor-delayed children have NAA ratios below the normative data curve, although the amount of falloff varies. One patient (21 months old) was diagnosed as diplegic with a right-sided preference, and data points are plotted for both the combined CSTs (upper data point) and only the most affected side (lower data point). This is an added benefit of using volumetric imaging in that bilateral comparisons can be readily accomplished. Although the sample size is too small to assert this, the overall findings from our pilot study indicate that NAA ratios are reduced for those with motor delay compared to controls with an effect size of ≈10% (9.5 ± 6.4% for NAA/Cho and 6.2 ± 3.5% for NAA/Cr), similar to the findings reported by Filippi et al (10), Kulak et al (11), and Fayed et al (12).

Figure 4
NAA/Cho (left) and NAA/Cho (right) ratios computed from the CSTs of three controls and five patients with motor delay. A logarithmic function is fit to the data acquired from controls to assume normative maturation curves. Given this assumption, the overall ...


In this article we present our study on the development of a fast 3D MRSI acquisition technique targeted for pediatric applications. The pulse sequence obtained 1 cc spectroscopic data from a 3D volume region within a 6.5-minute scan time and provided good SNR (≈22) for data analysis. By using DTI tractography to accurately identify the CST, spectroscopic information corresponding to this region were extracted and analyzed. As an illustrative example, we used this technique to compare NAA ratios in infants with and without motor delay. Overall, the study including a T1-weighted imaging, DTI, and MRSI sequence can be performed within a 12-minute total scan time.

The use of a rapid spiral-based readout acquisition using PRESS modules allows volumetric coverage within a clinically acceptable scan time. The eight-channel receiver coils at 3T provide additional SNR increases. A unique feature of combining DTI and 3D MRSI is the ability to segment and extract specific pathways within the brain. Here we used DTI to acquire spectra along the CST through the use of a fiber tracking algorithm. Other pathways can likely be extracted for different studies of the brain using this or similar protocols. These characteristics are the strengths of this study.

One of the major limitations of this study is that even at a spatial resolution of 1 cc, which is relatively good for in vivo MRS, there are considerable partial volume effects with respect to the CSTs. Hence, metabolic information was obtained from both the tracts of interest and the immediately surrounding tissue. This is a fundamental limitation and, to the extent that the additional tissue is normal, we lose statistical power in detecting changes in NAA. In relation to our study population, the partial volume effects are confounded by the fact that several of our patients were globally delayed and, therefore, had other concurrent neurological deficits in addition to motor delay. By using voxel shifting of our spectroscopic data to center the spectra voxel with the CST voxel, this effect was reduced by a certain amount although the finite resolution of our acquisition still precludes a perfect CST spectra (13). A related concern comes from the spatial distortions inevitable when using EPI readouts for diffusion imaging. This feature, noticeable in the EPI images of Fig. 2, consequently is another source of inaccurate metabolic assessment of the CST.

The use of an interior PRESS box, in combination with outer-volume spatial selection saturation pulses, provides excellent reduction of subcutaneous lipid signals and high-quality spectroscopic data. However, this approach has the distinct disadvantage of prohibiting the acquisition of metabolic information from the entire CST as well as most of the cerebral cortex, as seen in Fig. 3. Further improvements to implement a truly volumetric 1H MRSI sequence capable of interrogating the entire brain would be attractive (14).

Although the main focus of this study was not in its clinical use for assessing developmental delay, a note on this subject is worthwhile. Developmental delay occurs in ≈5%–10% of the childhood population (15,16), with motor dysfunction occurring in 10%–20% of these cases (17,18). There is growing evidence that early identification of children with developmental delay is critical to treatment of, or intervention for, a disability and lessening its impact on the functioning of the child and family (19). MRI plays an important role in the comprehensive evaluation of children with many types of developmental delay. With respect to motor delay, because the CST play a critical role in normal motor function, it is reasonable to assume that at least a subset of motor dysfunction is associated with abnormalities in the CST. The ability to accurately assess in vivo the integrity of the CST has been limited with DTI perhaps playing the largest role to date (20,21).

There are several, albeit limited, reports in the literature concerning 1H MRS findings in developmental delay. While Martin et al (22) did not find any significant changes in brain metabolite ratios of children with global developmental delay, several other investigators reported detected abnormalities using 1H-MRS. Filippi et al (10), in a cohort of children over the age of 2 with mild developmental delay and normal MRI scans, reported significantly decreased NAA/Cr ratios and elevated Cho/Cr ratios in frontal and parieto-occipital subcortical white matter. In a two-slice 1H-MRSI (MRS imaging) study comparing autism, developmental delay, and normal control by Friedman et al (23), an 8% reduction of global NAA was found between the developmentally delayed and control children. In a more recent article, Kulak et al (11) reported NAA/Cr ratios in the basal ganglia are negatively correlated with learning disabilities in patients with spastic cerebral palsy. A study more closely related to our targeted population was published by Fayed et al (12). In that study of 12 children with isolated developmental delay (6 had isolated motor delay and 2 had motor delay in combination with language expressive delay) and 11 controls, 7%–15% decreases in NAA ratios were reported from single voxels located in the left centrum semiovale. In our study we found that, compared to age-matched controls, the NAA ratios were reduced in motor-delayed infants with an effect size of ≈10%. Additional data from both controls and motor delayed infants need to be collected to further confirm this difference.

In conclusion, we have developed a fast and robust 3D MRSI technique targeted for pediatric neuroimaging. In 6:24 minutes, 3D volumetric data with 1 cc resolution can be readily acquired. Necessary spectroscopic information was extracted by combining with DTI tractography. The ability to analyze spectra from the CST pathway (and other pathways) is unique to 3D MRSI.


Contract grant sponsor: National Institutes of Health; Contract grant numbers: NIH NS40117, RR09784; Contract grant sponsor: National Center for Research Resources, US Public Health Services; Contract grant number: 5 M01 RR-01271 to the the pediatric clinical research center, University of California, San Francisco.


1. Bottomley PA. Spatial localization in NMR spectroscopy in vivo. Ann N Y Acad Sci. 1987;508:333–348. [PubMed]
2. Frahm J, Bruhn H, Gyngell ML, Merboldt KD, Hanicke W, Sauter R. Localized high-resolution proton NMR spectroscopy using stimulated echoes: initial applications to human brain in vivo. Magn Reson Med. 1989;9:79–93. [PubMed]
3. 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. [PubMed]
4. Kimura H, Fujii Y, Itoh S, et al. Metabolic alterations in the neonate and infant brain during development: evaluation with proton MR spectroscopy. Radiology. 1995;194:483–489. [PubMed]
5. Horska A, Kaufmann WE, Brant LJ, Naidu S, Harris JC, Barker PB. In vivo quantitative proton MRSI study of brain development from childhood to adolescence. J Magn Reson Imaging. 2002;15:137–143. [PubMed]
6. Cunningham CH, Vigneron DB, Marjanska M, et al. Sequence design for magnetic resonance spectroscopic imaging of prostate cancer at 3 T. Magn Reson Med. 2005;53:1033–1039. [PubMed]
7. Adalsteinsson E, Irarrazabal P, Topp S, Meyer C, Macovski A, Spielman DM. Volumetric spectroscopic imaging with spiral-based k-space trajectories. Magn Reson Med. 1998;39:889–898. [PubMed]
8. Berman JI, Mukherjee P, Partridge SC, et al. Quantitative diffusion tensor MRI fiber tractography of sensorimotor white matter development in premature infants. Neuroimage. 2005;27:862–871. [PubMed]
9. Provencher SW. Automatic quantitation of localized in vivo 1H spectra with LCModel. NMR Biomed. 2001;14:260–264. [PubMed]
10. Filippi CG, Ulug AM, Deck MD, Zimmerman RD, Heier LA. Developmental delay in children: assessment with proton MR spectroscopy. AJNR Am J Neuroradiol. 2002;23:882–888. [PubMed]
11. Kulak W, Sobaniec W, Smigielska-Kuzia J, Kubas B, Walecki J. Metabolite profile in the basal ganglia of children with cerebral palsy: a proton magnetic resonance spectroscopy study. Dev Med Child Neurol. 2006;48:285–289. [PubMed]
12. Fayed N, Morales H, Modrego PJ, Munoz-Mingarro J. White matter proton MR spectroscopy in children with isolated developmental delay: does it mean delayed myelination? Acad Radiol. 2006;13:229–235. [PubMed]
13. Haacke E, Brown R, Thompson M. Magnetic resonance imaging. Wiley-Liss; New York: 1999.
14. Gu M, Spielman D. Robust lipid suppression using multiple frequency selective pulses for brain spectroscopic imaging at 3T; 14th Annual Meeting ISMRM; Seattle. 2006. p. 3213.
15. Drillien CM, Pickering RM, Drummond MB. Predictive value of screening for different areas of development. Dev Med Child Neurol. 1988;30:294–305. [PubMed]
16. Petersen MC, Kube DA, Palmer FB. Classification of developmental delays. Semin Pediatr Neurol. 1998;5:2–14. [PubMed]
17. Shevell M, Ashwal S, Donley D, et al. Practice parameter: evaluation of the child with global developmental delay: report of the Quality Standards Subcommittee of the American Academy of Neurology and the Practice Committee of the Child Neurology Society. Neurology. 2003;60:367–380. [PubMed]
18. Edwards SL, Sarwark JF. Infant and child motor development. Clin Orthop Relat Res. 2005:33–39. [PubMed]
19. Developmental surveillance and screening of infants and young children Pediatrics. 2001;108:192–196. [PubMed]
20. Filippi CG, Lin DD, Tsiouris AJ, et al. Diffusion-tensor MR imaging in children with developmental delay: preliminary findings. Radiology. 2003;229:44–50. [PubMed]
21. Glenn OA, Henry RG, Berman JI, et al. DTI-based three-dimensional tractography detects differences in the pyramidal tracts of infants and children with congenital hemiparesis. J Magn Reson Imaging. 2003;18:641–648. [PubMed]
22. Martin E, Keller M, Ritter S, Largo RH, Thiel T, Loenneker T. Contribution of proton magnetic resonance spectroscopy to the evaluation of children with unexplained developmental delay. Pediatr Res. 2005;58:754–760. [PubMed]
23. Friedman SD, Shaw DW, Artru AA, et al. Regional brain chemical alterations in young children with autism spectrum disorder. Neurology. 2003;60:100–107. [PubMed]