June 2000, VOLUME 174
NUMBER 6

Recommend & Share

June 2000, Volume 174, Number 6

Neuroradiology

Quantitative Comparison of Intrabrain Diffusion in Adults and Preterm and Term Neonates and Infants

+ Affiliations:
1Department of Medical Physics and Centre of Medical Imaging Research, University of Leeds, The Wellcome Wing, Leeds General Infirmary, Gt. George St., Leeds LS1 3EX, United Kingdom.

2Centre for Reproduction, Growth and Development, University of Leeds, D Fl. Clarendon Wing, Leeds General Infirmary, Leeds LS2 9NS, United Kingdom.

3Department of Paediatric Radiology, University of Leeds, Clarendon Wing, Leeds General Infirmary, Leeds LS2 9NS, United Kingdom.

Citation: American Journal of Roentgenology. 2000;174: 1643-1649. 10.2214/ajr.174.6.1741643

ABSTRACT
Next section

OBJECTIVE. Quantitative measurements of mean water diffusivity (Dav) were made in human neonates, infants, and adults to assess changes in brain tissue that occur with maturation.

SUBJECTS AND METHODS. Values of Dav were obtained by calculating the average of the diffusion measurements made with diffusion-sensitizing gradients placed along three orthogonal directions. The mean diffusivity, a rotationally invariant determination of apparent diffusion coefficient, was measured in five healthy prematurely born neonates and infants, in 10 healthy term neonates and infants, and in five adults.

RESULTS. Values of Dav were found to decrease with maturation in most parts of the brain. In prematurely born neonates and infants with a postmenstrual age (postgestastional age + postnatal age) under 36 weeks, the average value of Dav in frontal white matter was 1.90 × 10-3 mm2 sec-1. The corresponding value was measured as 1.62 × 10-3 mm2 sec-1 in neonates and infants born at term with a postnatal age of no more than 43 days and 0.79 × 10-3 mm2 sec-1 in the adult brain.

CONCLUSION. Values of Dav are known to decrease in neonates and young infants in the period immediately after ischemic insult. This decrease and the associated increase in signal intensity seen on diffusion-weighted imaging have been used to monitor ischemic brain injury in neonates and infants. Therefore, the decrease in Dav that occurs with maturation, which we report in this study, must be considered if quantitative diffusion measurements are used to assess ischemic neonatal brain injury.

Introduction
Previous sectionNext section

MR measurements of tissue water diffusion in the brain have been used to study adult stroke [1,2,3,4] and have shown that values of the apparent diffusion coefficient (ADC) acutely fall to subnormal levels before increasing to supranormal levels in the chronic period. Changes in diffusion occurring after ischemic insult in neonates and infants are consistent with the studies of adult stroke and ischemic insult in animal models in which the initial changes are thought to be related to cytotoxic edema (cell swelling), a decrease in extracellular space, or changes in membrane permeability [3]. Diffusion measurements have also been made in neonates and infants to monitor brain maturation [5,6,7,8,9,10]. The results were initially interpreted in terms of the development of myelin, which acts as a barrier to diffusion. Later studies have suggested that the observed decrease in ADC is a result of changes in the properties of the membranes or from a reduction in the water content of brain tissue [9, 10].

Water diffusion is a three-dimensional process that is usually anisotropic in brain parenchyma. The measurement of anisotropy can reveal subtle white matter features in the brain that are not seen on conventional MR images [11]. Diffusion anisotropy cannot be consistently quantified by applying diffusion-sensitizing magnetic-field gradients along only a single direction because measured values will depend on the relative orientation of the subject and these gradients [4]. Most previous studies of maturation in neonates and infants involving the measurement of ADC have been made by applying sensitizing gradients along only one or two directions. Consequently, it is impossible to make comparative studies using such measurements. Discrepancies as high as 52% have been reported between equivalent ADC values obtained from different studies in infants [7]. In our study, we have acquired average values of ADC (mean diffusivity - Dav). The Dav values are not dependent on the relative orientation of the head; therefore, they allow direct comparison with results obtained from other studies involving similar rotationally invariant quantitative diffusion measurements.

The aims of this study were to assess brain maturation in healthy human neonates and infants by measuring diffusion and to compare the results with those obtained in adults. The measured Dav values will be useful in assessing the changes that are known to occur after ischemic insult in neonates and infants [12,13,14]. A clinical imaging protocol was implemented that acquired T1- and T2-weighted MR images and ADC values in an acceptable examination time. Diffusion-weighted images were obtained in neonates and infants of different ages to compare the images with those of previous studies and to facilitate the identification of anatomic features within the diffusion data.

Subjects and Methods
Previous sectionNext section
Theory

Values of ADC were determined from measurements of the signal intensities obtained with (S) and without (S0) the diffusion-sensitizing gradients according to the following formula: 1 where x, y, and z are indexes that define the directions of the laboratory reference frame in which the measurement is performed and Dii is an element of the diffusion tensor . The diffusion weighting factor (b) has been described by Spielman et al. [4] and is dependent on the magnitude, duration, and separation of the diffusion-sensitizing gradients. To fully characterize diffusion, all components of the diffusion tensor need to be determined. This requires a relatively long measurement time, especially in a multislice examination involving the whole head. Therefore, we have measured the mean diffusivity (Dav), which provides a scalar measure of ADC comparatively quickly, and allows the diffusion measurement to be incorporated into routine clinical examinations. Dav is related to the trace of the diffusion tensor : 2

Dav can be determined by calculating the average of three orthogonal diffusion measurements, provided that there are no interactions between the diffusion-sensitizing gradient(s) and the gradients required for imaging [15]. The calculated Dav value is a rotationally invariant parameter that is independent of the relative orientation of the subject and the laboratory reference frame. The measurement of Dav allows comparison between different studies. Diffusion measurements made in tissue by applying the diffusion-sensitizing gradients along only a single axis can give results that are dependent on orientation, making comparative studies impossible. Uluğ et al. [1] have shown that even if orientation-independent or average diffusion-weighted images are acquired, such images can still be contaminated by the influence of the tissue T1, T2, or spin density.

MR Imaging Protocol

All diffusion measurements were made in the axial plane using the head coil of a 1.5-T Gyroscan ACS NT MR scanner equipped with POWERTRACK 6000 gradients (Philips Medical Systems, Best, The Netherlands). The main magnetic field (B0) was shimmed using an automatic procedure on the scanner just before the measurement was taken. The diffusion data were obtained with a multishot spin-echo echoplanar imaging sequence. ECG triggering and a navigator echo were used to reduce the motion artifacts, which compromise the accuracy of measured Dav values [16]. Gradients used to generate the images were refocused immediately after application; therefore, their contribution to the echo attenuation (S/S0) was minimal [17]. Imaging parameters included a 6-mm slice width with a 1-mm gap between the acquired slices, a field of view of 180 × 180 mm, a matrix size of 128 × 128, and a TE of 105 msec. The Dav values were calculated from three measurements with diffusion-sensitizing gradients placed along the x, y, and z axes, in turn; each measurement was completed in approximately 90 sec. Values of Dxx, Dyy, Dzz, and Dav were then calculated for regions of interest and on a pixel-by-pixel basis using equations 1 and 2.

The clinical imaging protocol implemented to study infants and neonates required the acquisition of T1- and T2-weighted images in addition to the measurement of diffusion data from almost all of the brain tissue. The concomitant time constraints led to the implementation of an imaging protocol in which Dav values were obtained from nine slices and from data acquired using only two values of b (0 and 1000 sec mm-2). The accuracy of the calculated Dav values was assessed in a preliminary study in which values obtained from a water-containing phantom and from the adult brain were compared with equivalent values in the literature. The effect of using only two values of b was assessed in this preliminary study by comparing the Dav values with those obtained from a separate and more time-consuming measurement involving five values of diffusion weighting (b = 0, 63, 250, 563, and 1000 sec mm-2).

Subjects

Neonates and infants were examined, after obtaining informed parental consent, using imaging protocols approved by the research ethics committee at our hospital. A pediatrician was present at all times during both the MR study and the transfer of subjects to and from the scanner. The neonates and infants showed no clinical signs of neural abnormality and had normal findings on sonography. They were continuously monitored during the MR examination by two separate ECG measurements and a device that measured oxygenation transcutaneously. To reduce head movement and, thus, motion artifacts, the subjects' heads were placed in a deflatable pillow containing small polystyrene balls; the pillow molded itself to the shape of the head on the removal of air. The infants and neonates were well wrapped in blankets and cotton wool was placed in their ears to diminish the effects of gradient-generated noise. The MR data was acquired from sleeping infants and neonates who were scanned after receiving food. Measurements made on subjects who were either awake or crying were rejected because these images contained significant artifacts that compromised the accuracy of Dav determinations and made image interpretation difficult. Data from all subjects examined were included in the data analysis provided the data were free of gross motion artifacts. Values of Dav were calculated both on a pixel-by-pixel basis to produce Dav maps and from regions of interest. Each region of interest was chosen on the basis of T1- and T2-weighted MR images. A region of interest was positioned in the frontal white matter, gray matter, cerebrospinal fluid, and in the splenium of the corpus callosum. The region of interest was defined using the Auto-trace tool of the image analysis program, Analyze AVW (version 1.0, Biomedical Imaging Resource, Mayo Foundation, Rochester, MN) [18]. This tool identified voxels located around a selected position that had gray-scale values corresponding to the anatomic region of interest. Each region of interest had an irregular shape confined to a particular region of anatomy and consisted of 30-100 voxels.

Diffusion data, essentially free of artifacts caused by bulk-tissue motion were acquired from four groups of healthy individuals. Group I consisted of five prematurely born neonates and infants with a postmenstrual age (postgestational age + postnatal age) of 30-36 weeks (mean, 32 weeks) and a mean postnatal age of 18 days (range, 4-44 days). Group II consisted of 10 neonates and infants born at term with a mean postnatal age of 13 days (range, 1-43 days). Group III consisted of five adult volunteers with a mean age of 26 years (range, 20-30 years). Diffusion data were also acquired from two infants who were 4 and 8 months old.

Results
Previous sectionNext section
Diffusion-Weighted Images

Diffusion-weighted images of the brain obtained from healthy neonates and infants showed different signal intensity, depending on maturation and the direction of the applied diffusion-sensitizing gradient (Figs. 1A,1B,1C,2A,2B,2C,3A,3B,3C). Although small variations between individuals of a similar age were observed, the image features we describe were found to be representative for preterm neonates (with a postmenstrual age of more than 32 weeks), for term neonates, and for infants less than 4 months old.

figure
View larger version (164K)
Fig. 1A. —Diffusion-weighted MR images of 4-week-old male neonate with postmenstrual age of 33 weeks. Images acquired with diffusion sensitization (b = 1000 sec mm-2) applied in left-right (A), anterior-posterior (B), and through-plane (C) directions.

figure
View larger version (133K)
Fig. 1B. —Diffusion-weighted MR images of 4-week-old male neonate with postmenstrual age of 33 weeks. Images acquired with diffusion sensitization (b = 1000 sec mm-2) applied in left-right (A), anterior-posterior (B), and through-plane (C) directions.

figure
View larger version (160K)
Fig. 1C. —Diffusion-weighted MR images of 4-week-old male neonate with postmenstrual age of 33 weeks. Images acquired with diffusion sensitization (b = 1000 sec mm-2) applied in left-right (A), anterior-posterior (B), and through-plane (C) directions.

figure
View larger version (184K)
Fig. 2A. —Diffusion-weighted MR images of 1-day-old female neonate born at term. Images acquired with diffusion sensitization (b = 1000 sec mm-2) applied in left-right (A), anterior-posterior (B), and through-plane (C) directions.

figure
View larger version (155K)
Fig. 2B. —Diffusion-weighted MR images of 1-day-old female neonate born at term. Images acquired with diffusion sensitization (b = 1000 sec mm-2) applied in left-right (A), anterior-posterior (B), and through-plane (C) directions.

figure
View larger version (179K)
Fig. 2C. —Diffusion-weighted MR images of 1-day-old female neonate born at term. Images acquired with diffusion sensitization (b = 1000 sec mm-2) applied in left-right (A), anterior-posterior (B), and through-plane (C) directions.

figure
View larger version (179K)
Fig. 3A. —Diffusion-weighted MR images of 4-month-old male infant. Images acquired with diffusion sensitization (b = 1000 sec mm-2) applied in left-right (A), anterior-posterior (B), and through-plane (C) directions.

figure
View larger version (188K)
Fig. 3B. —Diffusion-weighted MR images of 4-month-old male infant. Images acquired with diffusion sensitization (b = 1000 sec mm-2) applied in left-right (A), anterior-posterior (B), and through-plane (C) directions.

figure
View larger version (179K)
Fig. 3C. —Diffusion-weighted MR images of 4-month-old male infant. Images acquired with diffusion sensitization (b = 1000 sec mm-2) applied in left-right (A), anterior-posterior (B), and through-plane (C) directions.

The diffusion-weighted MR images of the brain in preterm neonates showed relatively high signal intensity in the cerebral cortex compared with that in white matter (Fig. 1A,1B,1C). Further, relatively high signal intensity was seen in the posterior limb of the internal capsule when the diffusion-sensitizing gradients were positioned either in the left-right (Fig. 1A) or in the anterior-posterior direction (Fig. 1B), whereas the corpus callosum returned high signal intensity when these gradients were placed through-plane (Fig. 1C). When compared with term neonates or infants, we found that it was relatively more difficult to acquire diffusion-weighted images free of motion artifacts when examining the brains of nonsedated premature infants.

For infants less than 44 days old who were born at term and also neonates born at term, diffusion-weighted images showed a relative increase in white matter signal intensity compared with that observed in premature neonates, resulting in apparent decrease in contrast between white matter and cerebral cortex (Fig. 2A,2B,2C). In addition, more regions of relative hyperintensity were observed in these images. Thus, from a postmenstrual age of approximately 40 weeks, the corona radiata and the posterior limb of the internal capsule showed high signal intensity when the diffusion-sensitizing gradients were placed along the left-right direction (Fig. 2A).

MR diffusion-weighted images acquired from the brains of older infants (Fig. 3A,3B,3C) showed relative hyperintensity in the posterior limb of the internal capsule, in the corona radiata, and in the optic radiation when the diffusion-sensitizing gradient was placed along the left-right direction (Fig. 3A). The internal capsule and some regions running along the left-right direction showed high signal intensity when the sensitizing gradient was applied anterior-posterior (Fig. 3B), whereas the corpus callosum again showed high signal intensity when the gradient was applied in the through-plane direction (Fig. 3C).

When diffusion-weighted images were acquired from slices around the petrous bone, some infants and neonates had diffusion-weighted images containing areas of high signal intensity corresponding to susceptibility artifact [19] (data not shown). The images in Figures 2A,2B,2C and 3A,3B,3C resemble those of Sakuma et al. [6], who acquired their data using smaller values of diffusion weighting (b value).

Apparent Diffusion Coefficient (Mean diffusivity - Dav)

Values of Dav measured in a water-containing phantom are shown in Table 1. There is a less than 5% difference between measured values of Dav obtained by placing the diffusion-sensitizing gradient along each of the three orthogonal laboratory axes, in turn. Further, there is good agreement both between measurements made using two and five values of the diffusion weighting gradients and with values of Dav of water obtained from the literature [15, 20].

TABLE 1 Measured Values of Diffusion in Water-Containing Phantom at 22°C

Table 2 contains average values of Dav obtained from selected regions of interest within the brains brains of term and preterm neonates and infants and in the adult brain (groups I-III). The values shown for each age group are average values obtained from five premature infants, five adults, and 10 term neonates, respectively. This table also shows average Dav values calculated from measurements involving the use of five b values. This data gave very good straight lines when ln S/S0 was plotted against b (data not shown). This indicated that the diffusion data followed the expected relationship (for b values up to 1000 sec mm-2) and that there was a small error in the fit to the data. Values of Dav in the adult brain that were obtained from the literature [17, 21] are included in this table.

TABLE 2 Dav Values (× 10-3 mm2 sec-1) in Selected Regions of Interest in Healthy Brain

Representative color images mapping Dav values calculated on a pixel-by-pixel basis in four different brains are shown in Figures 4,5,6,7. Figure 4 was obtained from a premature neonate and shows variability of Dav in the white matter. There is a relatively large area in the frontal white matter in this neonate with comparatively high values of Dav, and in the posterior white matter there are areas with lower average values of mean diffusivity. As the brain matures, values in white matter decrease, particularly in the frontal region. The images obtained from adults show little contrast between gray and white matter, an observation consistent with the region of interest measurements made both by our study and by others [17, 21] and shown in Table 2.

figure
View larger version (209K)
Fig. 4. —MR image mapping quantitative values of apparent diffusion coefficient (Dav) in brain of 12-day-old female neonate with postmenstrual age of 35 weeks. Values of nonlinear scale should be multiplied by 10-3 mm2 s-1 to obtain value of Dav.

figure
View larger version (206K)
Fig. 5. —MR image mapping quantitative values of apparent diffusion coefficient (Dav) in brain of 1-day-old female neonate born at term. Values of nonlinear scale should be multiplied by 10-3 mm2 s-1 to obtain value of Dav.

figure
View larger version (221K)
Fig. 6. —MR image mapping quantitative values of apparent diffusion coefficient (Dav) in brain of 8-month-old female infant. Values of nonlinear scale should be multiplied by 10-3 mm2 s-1 to obtain value of Dav.

figure
View larger version (208K)
Fig. 7. —MR image mapping quantitative values of apparent diffusion coefficient (Dav) in brain of 20-year-old woman. Values of nonlinear scale should be multiplied by 10-3 mm2 s-1 to obtain value of Dav.

Discussion
Previous sectionNext section
Diffusion-Weighted Imaging

Some previous MR studies measuring water diffusion within brain tissue in healthy infants have presented data in the form of diffusion-weighted images [5, 6, 8]. The diffusion-weighted images shown in Figures 1A,1B,1C,2A,2B,2C,3A,3B,3C were acquired using an echo planar imaging sequence and a navigator echo technique [16] and were almost free of motion artifacts. Few of the previous diffusion-weighted studies show axial images acquired from the healthy neonatal brain.

Each diffusion-weighted image in Figures 1A,1B,1C,2A,2B,2C,3A,3B,3C was obtained by applying diffusion-sensitizing gradients along one of three orthogonal directions. This data cannot be used as a quantitative characterization of diffusion anisotropy, but it can indicate that diffusion anisotropy is present. Moseley et al. [22] have shown that the diffusivity and hypointensity in diffusion-weighted images are greatest when the diffusion gradients and the fiber tracts are aligned parallel to each other and smallest when they are perpendicular. The images in Figures 1A,1B,1C,2A,2B,2C,3A,3B,3C are consistent with this observation; the splenium, for example, is relatively hypointense when the diffusion gradients are placed in the left-right direction (Figs. 1A, 2A, and 3A) and relatively hyperintense when the gradients are placed through-plane (Figs. 1C, 2C, and 3C). The images in Figure 1A,1B,1C show that marked diffusion anisotropy is already present in the posterior limb of the internal capsule and in the corpus callosum at a postmenstrual age of 33 weeks. As the brain matures, it is apparent that diffusion becomes markedly anisotropic in the optic radiation and the corona radiata in the term neonates. These age-related changes are indicative of greater restriction of the diffusing water molecules within the developing white matter. Wimberger et al. [23] have shown that these changes in diffusion anisotropy cannot be considered to be indicative of the onset of myelination because anisotropic diffusion is observed in the corpus callosum of the neonates and autopsy studies indicate that the corpus callosum does not myelinate until the 13th postconceptional month [24].

ADC Measurements (Dav)

The determination of accurate Dav values is not trivial, especially when the subjects are uncooperative and examination time is limited. Our measurement protocol involved the use of ECG gating, a head restraint, a navigator echo [16], and a fast multishot echo planar imaging sequence. These procedures reduce artifacts in the acquired data that arise from subject motion and compromise the accuracy of calculated Dav values. The effects of bulk motion are particularly evident in the background signal outside the head in the Dav map acquired from a premature neonate (Fig. 4). These maps were calculated from the ratio in equation 1 and, therefore, show motion artifacts particularly clearly. Despite the care taken to minimize motion in our patients, some neonates and infants did appear to move during their scans and the resulting artifacts prevented the determination of accurate Dav values. This was a marked problem in sleeping (nonsedated) premature neonates. We are not certain of the reason, but the smaller head size and the longer values of T1 in brain tissue in this age group results in lower signal levels, making quantification of diffusion difficult.

The difference between ADC values in water measured by placing the diffusion-sensitizing gradient along each of the three orthogonal laboratory frame axes, in turn, is less than 5% (Table 1), indicating that the effects of the imaging gradients on the accuracy of measured Dav values is small. The data obtained from both the phantom and the brain shows that there is good agreement between the measurements made using two b values (two values of diffusion weighting) and the lengthier measurement using five b values. Furthermore, values of ADC in water and in the adult brain are in good agreement with corresponding data in the literature [17, 21]. Therefore, we are confident that we have a reasonable trade-off in terms of speed of measurement and accuracy in our determinations of ADC (Dav). A recent study by Shimony et al. [11] has shown that it is possible to make quantitative measurements of diffusion in an acceptably short period and that the data can be used to generate images that map the diffusion anisotropy.

Average values of Dav calculated from selected regions of interest (Table 2) are in good agreement with those obtained recently by Neil et al. [9], who also used a rotationally invariant method to measure diffusion. The agreement was good even for values of Dav obtained in the cerebral cortex, where partial volume effects might affect measured values because of the close proximity of cerebrospinal fluid. Our values are similar to those obtained by Neil et al. both in the cortex and in the basal ganglia (where the partial volume effects are not expected to be significant). For the frontal white matter in term neonates, Neil et al. measured an average value of 1.45 ± 0.12 × 10-3 mm2 sec-1, which compares favorably with the value of 1.62 ± 0.16 × 10-3 mm2 sec-1 obtained in our study. Our Dav values are slightly higher than equivalent measurements made by Hüppi et al. [10] and somewhat lower than those made by Toft et al. [7]. In white matter in neonates, an interpolated value of 1.98 - 2.11 × 10-3 mm2 sec-1 can be inferred from the data of Toft et al. This relatively high value may be a consequence of a measurement technique in which the diffusion-sensitizing gradient was applied along only a single direction.

With the exception of cerebrospinal fluid, in the first year of life there is a trend of decreasing average value of Dav with increasing maturation in most areas of the brain. The average value of Dav in the frontal white matter is 1.90 × 10-3 mm2 sec-1 in premature neonates and infants, 1.62 × 10-3 mm2 sec-1 in term neonates, and 0.79 × 10-3 mm2 sec-1 in adults. The corresponding decrease in the gray matter of the cerebral cortex is less than in the frontal white matter: a value of 1.29 × 10-3 mm2 sec-1 in premature neonates drops to 0.87 × 10-3 mm2 sec-1 in adult gray matter. In the splenium, average Dav values decrease from 1.43 × 10-3 mm2 sec-1 in the premature neonate to 0.75 × 10-3 mm2 sec-1 in the adult. The trend of decreasing Dav values with increasing age has also been reported by other researchers [7,9,10].

Values of Dav have been calculated for each pixel in the acquired data set and then displayed in the form of images (Figs. 4,5,6,7). Differences in Dav between different brain structures and the changes occurring with maturation are apparent in these images. At birth, there is a relatively large difference in Dav values between gray and white matter. This difference generates contrast between gray and white matter in the Dav maps, which decreases with maturation. Thus, in the Dav maps of the neonates, there is a clear distinction between gray and white matter (Figs. 4 and 5), which is less evident in equivalent images of the older infant and adult brains (Figs. 6 and 7).

The diffusion-weighted images of prematurely born neonates showed a high level of contrast between the white matter (frontal, parietal, and occipital) and the cerebral cortex with the cortex showing comparatively high signal intensity (Fig. 1A,1B,1C). Although it is possible that this contrast might be a consequence of contamination caused by differences in spin density or T1 [1] between these tissue types, the observed differences in intensity are more likely to be caused by the relatively large values of Dav in white matter in the premature neonates.

The decrease in Dav values in frontal white matter occurring with maturation (Table 2) (Figs. 4,5,6,7) can be interpreted using results obtained from studies both in animal models [23, 25] and in humans [9,10]. Neil et al. [9] have suggested that the change seen in human neonates is a consequence of the water loss that occurs in the developing brain. These authors argue that the higher water content in the less mature brain results in a greater separation of membranes and other barriers to diffusion. The animal study by Baratti et al. [25] led those researchers to suggest that the decrease in Dav in maturing kittens was caused primarily by changes in diffusivity within the cellular compartment. These changes are thought to be caused by an increased concentration of macromolecules and by a greater membrane-surface to cell-volume ratio. A study of rats conducted by Wimberger et al. [23] has shown changes in diffusion anisotropy in some parts of the brain before myelination. Wimberger et al. suggested that changes in diffusion occur at a time when premyelinated axons interact with glial cells, resulting in an increase in axonal diameter, changes in axonal membranes, and an increase in microtubule-associated proteins, which is associated with an increase in the number of oligodendrocytes. These conclusions are consistent with results obtained by Hüppi et al. [10], who have detected anisotropic diffusion in white matter in the period immediately preceding myelination. These conclusions are also consistent both with our studies of magnetization transfer [26] and with the bandlike structures we have observed in white matter on T2-weighted images, corresponding to migrating glial cells [27]. The glial cells form in the persisting germinal matrix and are apparent on MR images as they migrate through the frontal white matter toward the cerebral cortex [27]. The mature oligodendroglia are implicated in myelination and the concomitant loss of water.

The further decrease in Dav values occurring in the frontal white matter in infants (Fig. 6) is probably partially caused by the formation of myelin. Neil et al. [9] have shown that in term neonates, Dav values in the more fully myelinated posterior limb of the internal capsule are lower when compared with the unmyelinated anterior limb of the internal capsule. We have also observed an age-related increase in magnetization transfer in neonates and young infants that we believe to be caused by the decreasing brain water content and the accumulation of myelin and its precursors [26].

In conclusion, we have implemented an imaging protocol that obtains a rotationally invariant scalar measure of ADC. The protocol covers almost all of the brain in a relatively short period, allowing the measurements to be part of a routine examination that also includes conventional MR imaging. Our study has shown that values of mean diffusivity (Dav) in brain tissue in neonates and infants decrease in the first months of life. Because Dav values are also known to decrease immediately after ischemic insult [3], the changes seen in healthy neonates and infants will require consideration when assessing measurements of ADC in neonates with suspected ischemic injury.

Supported in part by Leeds Teaching Hospital National Health Service Trust, which received funding from the National Health Service Executive.

The views expressed in this publication are those of those of the authors and not necessarily those of the National Health Service Executive.

Address correspondence to S. F. Tanner.

We thank Philips Medical Systems for providing the diffusion acquisition software.

References
Previous sectionNext section
1. Uluğ AM, Beauchamp N Jr, Bryan RN, van Zijl PCM. Absolute quantitation of diffusion constants in human stroke. Stroke 1997; 28:483-490 [Google Scholar]
2. Marks MP, de Crespigny A, Lentz D, Enzmann DR, Albers GW, Mosely ME. Acute and chronic stroke: navigated spin-echo diffusion-weighted MR imaging. Radiology 1996; 199:403-408 [Google Scholar]
3. Schlaug G, Siewert B, Benfield A, Edelman RR, Warach S. Time course of the apparent diffusion coefficient (ADC) abnormality in human stroke. Neurology 1997; 49:113-119 [Google Scholar]
4. Spielman DM, Butts K, de Crespigny A, Mosely M. Diffusion weighted imaging of clinical stroke. Int J Neuroradiol 1995; 1:44-55 [Google Scholar]
5. Takeda K, Nomura Y, Sakuma H, Tagami T, Okuda Y, Nakagama T. MR assessment of normal brain development in neonates and infants: comparative study of T1 and diffusion weighted images. J Comput Assist Tomogr 1997; 21:1-7 [Google Scholar]
6. Sakuma H, Nomura Y, Takeda K, et al. Adult and neonatal brain: diffusion anisotropy and myelination with diffusion-weighted MR imaging. Radiology 1991; 180:229-233 [Google Scholar]
7. Toft PB, Leth H, Peitersen B, Lou H, Thomsen C. The apparent diffusion coefficient of water in gray and white matter of the infant brain. J Comput Assist Tomogr 1996; 20:1006-1011 [Google Scholar]
8. Nomura Y, Sakuma H, Takeda K, Tagami T, Okuda Y, Nakagawa T. Diffusional anisotropy of the human brain assessed with diffusion-weighted MR: relation with normal brain development and aging. AJNR 1994; 15:231-238 [Google Scholar]
9. Neil JJ, Shiran SI, McKinstry RC, et al. Normal brain in human newborns: apparent diffusion coefficient and diffusion anisotropy measured by using diffusion tensor MR imaging. Radiology 1998; 209:57-66 [Google Scholar]
10. Hüppi PS, Maier SE, Peled S, et al. Microstructural development of human newborn cerebral white matter assessed in vivo by diffusion tensor magnetic resonance imaging. Pediatric Research 1998; 44:585-590 [Google Scholar]
11. Shimony JS, McKinstry RC, Erbil A, et al. Quantitative diffusion-tensor anisotropy brain MR imaging: normative human data and anatomic analysis. Radiology 1999; 212:770-784 [Google Scholar]
12. Robertson RL, Maier SE, Robson CD, Mulkern RV, Kara PM, Barnes PD. MR line scan diffusion imaging of the brain in children. AJNR 1999; 20:419-425 [Google Scholar]
13. Pennock JM, Cowan FM, Schweiso JE, et al. Clinical role of diffusion-weighted imaging: neonatal studies. MAGMA 1994; 2:273-278 [Google Scholar]
14. Cowan FM, Pennock JM, Hanrahan JD, Manji KP, Edwards AD. Early detection of cerebral infarction and hypoxic ischemic encephalopathy in neonates using diffusion-weighted magnetic resonance imaging. Neuropediatrics 1994; 25:172-175 [Google Scholar]
15. Conturo TE, McKinstry RC, Akbudak E, Robinson BH. Encoding of anisotropic diffusion with tetrahedral gradients: a general mathematical duffusion formalism and experimental results. Magn Reson Med 1996; 35:399-412 [Google Scholar]
16. Ordridge RJ, Helpern JA, Qing ZX, Knight RA, Nagesh V. Correction of motional artifacts in diffusion-weighted MR images using navigator echoes. Magn Reson Imag 1994; 12:455-460 [Google Scholar]
17. Pierpaoli C, Jezzard P, Basser PJ, Barnett A, Di Chiro G. Diffusion tensor MR imaging of the human brain. Radiology 1996; 201:637-648 [Google Scholar]
18. Robb RA. A software system for interactive and quantitative analysis of biomedical images. In: Hohne KH, Fuchs H, Pizer SM, eds. 3D imaging in medicine. Berlin: Springer Verlag, 1990: 333-361 [Google Scholar]
19. Butts K, Pauly J, de Crespigny A, Mosely M. Isotropic diffusion-weighted and spiral-navigated interleaved EPI for routine imaging of acute stroke. Magn Reson Med 1997; 38:741-749 [Google Scholar]
20. Mills R. Self diffusion in normal and heavy water in the range 1-45°. J Phys Chem 1973; 77:685-689 [Google Scholar]
21. Jakob PM, Chen Q, Lovblad KO, Griswold M, Warach S, Edelman RR. Single-shot diffusion weighted imaging of the brain with HASTE. In: Oudkerk M, Edelman R, eds. High-power gradient MR-imaging advances in MRI. II. London: Blackwell Science, 1997:177-181 [Google Scholar]
22. Moseley ME, Cohen Y, Kucharczyk J, et al. Diffusion-weighted MR imaging of anisotropic water diffusion in cat central nervous system. Radiology 1990; 176:439-445 [Google Scholar]
23. Wimberger DM, Roberts TP, Barkovich AJ, Prayer LM, Mosely ME, Kucharczyk J. Identification of “premyelination” by diffusion-weighted MRI. J Comput Assist Tomogr 1995; 19:28-33 [Google Scholar]
24. Brody BA, Kinney HC, Kloman AS, Gilles FH. Sequence of central nervous system myelination in human infancy. 1. An autopsy study of myelination. J Neuropathol Exp Neurol 1987; 46:283-301 [Google Scholar]
25. Baratti C, Barnett AS, Pierpaoli C. Comparative MR imaging study of brain maturation in kittens with T1, T2, and the trace of the diffusion tensor. Radiology 1999; 210:133-142 [Google Scholar]
26. Tanner S, Berry E, Childs A-M, et al. Age dependence of quantitative magnetisation transfer measurements in the infant brain. In: Proceedings of the sixth meeting of the International Society for magnetic resonance in medicine. Berkely, CA: International Society for Magnetic Resonance in Medicine, 1998:92 [Google Scholar]
27. Childs A-M, Ramenghi LA, Evans DJ, et al. MR features of developing periventricular white matter in preterm infants: evidence of glial cell migration. AJNR 1998; 19:971-976 [Google Scholar]

Recommended Articles

Quantitative Comparison of Intrabrain Diffusion in Adults and Preterm and Term Neonates and Infants

Full Access, , ,
American Journal of Roentgenology. 2000;174:1723-1728. 10.2214/ajr.174.6.1741723
Abstract | Full Text | PDF (3770 KB) | PDF Plus (3881 KB) 
Full Access, ,
American Journal of Roentgenology. 2002;179:1515-1522. 10.2214/ajr.179.6.1791515
Abstract | Full Text | PDF (25160 KB) | PDF Plus (25278 KB) 
Full Access
American Journal of Roentgenology. 2000;174:1523-1528. 10.2214/ajr.174.6.1741523
Citation | Full Text | PDF (1166 KB) | PDF Plus (1208 KB) 
Full Access, , , , ,
American Journal of Roentgenology. 2000;174:1717-1722. 10.2214/ajr.174.6.1741717
Abstract | Full Text | PDF (8453 KB) | PDF Plus (946 KB) 
Full Access, , ,
American Journal of Roentgenology. 2007;189:476-486. 10.2214/AJR.07.2132
Abstract | Full Text | PDF (645 KB) | PDF Plus (495 KB) 
Full Access, , , ,
American Journal of Roentgenology. 2003;180:851-859. 10.2214/ajr.180.3.1800851
Citation | Full Text | PDF (16495 KB) | PDF Plus (16573 KB)