# Effects of number of diffusion gradient directions on derived diffusion tensor imaging indices in human brain.

**ABSTRACT** The effects of a number of diffusion-encoding gradient directions (NDGD) on diffusion tensor imaging (DTI) indices have been studied previously with theoretic analysis and numeric simulations. In this study, we made in vivo measurements in the human brain to compare different clinical scan protocols and to evaluate their effects on the calculated DTI indices.

Fifteen healthy volunteers were scanned with a 1.5T MR scanner. Single-shot DTI images were acquired using 3 protocols different in NDGD and number of excitations (NEX) for each direction (NDGD/NEX = 6/10, 21/3, 31/2). Means and standard error of mean (SEM) were calculated and compared in 6 regions of interest (ROIs) for mean diffusivity (D), fractional anisotropy (FA), diffusion tensor eigenvalues (lambda(1), lambda(2), and lambda(3)), and correlation coefficients (r) of these indices among the 3 DTI protocols.

At the ROI level, no significant differences were found for the mean and SEM of D and FA among protocols (P > .05). The 6-NDGD protocol, however, yielded higher values for lambda(1) and lambda(2) and lower values for lambda(3) in most ROIs (P < .05) compared with the other protocols. At the voxel level, the correlation between the protocols r(21-31) were higher than r(6-21) and r(6-31) in most ROIs. The correlation of FA among 3 protocols also increased with increasing anisotropy.

For ROI analyses, different NDGDs lead to similar values of FA and D but different eigenvalues. However, different NDGDs at the voxel level provide varying values. The selection of the NDGD, therefore, should depend on the focus of different DTI applications.

**0**Bookmarks

**·**

**88**Views

- [Show abstract] [Hide abstract]

**ABSTRACT:**Diffusion tensor imaging (DTI) is very useful for investigating white matter integrity in ageing and neurological disorders; thus, evaluating its reproducibility under different acquisition protocols and analysis methods may assist in the design of clinical studies. To measure the reproducibility of DTI in normal subjects, this study include (1) depicting the reproducibility of DTI measurements in commonly used regions-of-interest analysis by intraclass correlation coefficient (ICC) and coefficient of variation (CV), (2) evaluating and comparing inter and intrasession test-retest reproducibility, and (3) illustrating the effect of the number of diffusion-encoding directions (NDED) and registration algorithms on measurement reproducibility. DTI measurements exhibit high reproducibility, with overall (430/480) ICC ≥ 0.70, (478/480) within-subject CV (CVws) ≤10.00 % and between-subject CV (CVbs) ranging from 1.32 to 13.63 %. Repeated measures ANOVAs and paired t tests were conducted to compare inter and intrasession reproducibility with different diffusion sampling schemes and registration algorithms. Our results also confirmed that increasing the NDED could improve the accuracy and reproducibility of DTI measurements. In addition, we compared reproducibility indices that were derived using different registration algorithms, and a tensor-based deformable registration yielded the most reproducible results. Finally, we found that increasing the NDED could reduce the difference between the reproducibility of measurement derived using different registration algorithms and between the reproducibility of intersession and intrasession. Our results suggest that the choice of DTI acquisition protocol and post-processing methods can influence the accurate estimation and reproducibility of DTI measurements and should be considered carefully for clinical applications.Neuroradiology 03/2014; · 2.37 Impact Factor - SourceAvailable from: Georg GrönHans-Peter Müller, Jan Kassubek, Georg Grön, Reiner Sprengelmeyer, Albert C Ludolph, Stefan Klöppel, Nicola Z Hobbs, Raymund Ac Roos, Alexandra Duerr, Sarah J Tabrizi, Michael Orth, Sigurd D Süssmuth, Georg Bernhard Landwehrmeyer[Show abstract] [Hide abstract]

**ABSTRACT:**Corrupted gradient directions (GD) in diffusion weighted images may seriously affect reliability of diffusion tensor imaging (DTI)-based comparisons at the group level. In the present study we employed a quality control (QC) algorithm to eliminate corrupted gradient directions from DTI data. We then assessed effects of this procedure on comparisons between Huntington disease (HD) subjects and controls at the group level.BioMedical Engineering OnLine 09/2014; 13(1):128. · 1.75 Impact Factor - SourceAvailable from: Jean-Marc PeyratStelios Angeli, Nicholas Befera, Jean-Marc Peyrat, Evan Calabrese, George Allan Johnson, Christakis Constantinides[Show abstract] [Hide abstract]

**ABSTRACT:**The complex cardiac fiber structural organization and spatial arrangement of cardiomyocytes in laminar sheetlets contributes greatly to cardiac functional and contractile ejection patterns. This study presents the first comprehensive, ultra-high resolution, fully quantitative statistical tensor map of the fixed murine heart at isotropic resolution of 43 μm using diffusion tensor (DT) cardiovascular magnetic resonance (CMR).Journal of Cardiovascular Magnetic Resonance 10/2014; 16(1):77. · 4.44 Impact Factor

Page 1

ORIGINAL

RESEARCH

Effects of Number of Diffusion Gradient Directions

on Derived Diffusion Tensor Imaging Indices in

Human Brain

H. Ni

V. Kavcic

T. Zhu

S. Ekholm

J. Zhong

BACKGROUND AND PURPOSE: The effects of a number of diffusion-encoding gradient directions

(NDGD) on diffusion tensor imaging (DTI) indices have been studied previously with theoretic analysis

and numeric simulations. In this study, we made in vivo measurements in the human brain to compare

different clinical scan protocols and to evaluate their effects on the calculated DTI indices.

METHODS: Fifteen healthy volunteers were scanned with a 1.5T MR scanner. Single-shot DTI images

were acquired using 3 protocols different in NDGD and number of excitations (NEX) for each direction

(NDGD/NEX ? 6/10, 21/3, 31/2). Means and standard error of mean (SEM) were calculated and

compared in 6 regions of interest (ROIs) for mean diffusivity (?D?), fractional anisotropy (FA), diffusion

tensor eigenvalues (?1, ?2, and ?3), and correlation coefficients (r) of these indices among the 3 DTI

protocols.

RESULTS: At the ROI level, no significant differences were found for the mean and SEM of ?D? and FA

among protocols (P ? .05). The 6-NDGD protocol, however, yielded higher values for ?1and ?2and

lower values for ?3in most ROIs (P ? .05) compared with the other protocols. At the voxel level, the

correlation between the protocols r21–31were higher than r6–21and r6–31in most ROIs. The correlation

of FA among 3 protocols also increased with increasing anisotropy.

CONCLUSION: For ROI analyses, different NDGDs lead to similar values of FA and ?D? but different

eigenvalues. However, different NDGDs at the voxel level provide varying values. The selection of the

NDGD, therefore, should depend on the focus of different DTI applications.

W

(WM) changes that are not normally seen on conventional

MR imaging can be detected. DTI has been applied in various

diseases, such as Alzheimer disease (AD),1-4multiple sclerosis

(MS),5-15and HIV infections,16-18to monitor and assess WM

changes.

There are some basic requirements in clinical applications

of DTI. For example, the total scan time cannot be too long,

relatively thin sections are required for accurate depiction of

structures, and a sufficient number of sections is needed to

cover the entire brain. Optimization of the DTI acquisition

protocols is needed with regard to the above limitations and

requirements. One of the most important factors in DTI ac-

quisition is the number of diffusion-encoding gradient direc-

tions(NDGD).Atleast6diffusion-weighted(DW)imagesfor

every section (ie, NDGD ? 6) are needed to calculate the dif-

fusion tensor (D), and all DTI indices are calculated from D.

As NDGD increases, more DW images are used for the calcu-

lation of D, resulting in more accurate D estimation. Alterna-

tively, more averaging of each DW image also results in a

higher signal-to-noise ratio (SNR) and improved estimation

of D. However, both methods require a longer scan time. In

ith MR diffusion tensor imaging (DTI), diffusion an-

isotropy can be quantified, and subtle white matter

mostclinicalreports(e.g.,onADstudies),diffusion-encoding

gradients were applied in only 6 directions.1-3As the scanner

hardware has improved rapidly in recent years, use of more

DW directions has become more popular. For example, in 1

study, 30 directions were used.4

There is still no clear conclusion about different schemes

for selection of NDGD and number of excitations (NEX) for

theevaluationofDTIindices.Someresearchers19,20claimthat

using more than 6 DW gradient directions provides better

measures of the D than the conventional 6 directions. In 1

study,noadvantageintheuseofmorethan6samplingorien-

tations was shown as long as the selected orientations point to

the vertices of an icosahedron.21Another study22determined

that the minimum number of unique encoding directions re-

quired for robust anisotropy estimation is between 18 and 21.

ArecentstudywithMonteCarlosimulations23concludedthat

at least 20 unique sampling orientations were necessary for a

robust estimation of anisotropy, whereas at least 30 unique

samplingorientationswererequiredforarobustestimationof

tensor orientation and mean diffusivity. The error propaga-

tion on effects of NDGD and b value on fractional anisotropy

(FA)wererecentlyinvestigatedbytheoreticanalysis,24andthe

results suggested an increase in error propagated to calculate

FA as NDGD decreased. To our knowledge, no experimental

studieshaveinvestigatedtheeffectsofvariousNDGDsonDTI

measurements in vivo.

The purpose of this work was to compare DTI protocols

with combinations of various NDGD and various numbers of

images. The effects of the number of diffusion gradient direc-

tions on the calculated DTI indices were analyzed under typi-

cal clinical conditions, and suggestions were made for more

reliable DTI protocol designs. The main goal of our study was

to test, in humans, previous results from simulation studies,

Received July 19, 2005; accepted after revision December 16.

From the Departments of Radiology (H.N., S.E., J.Z.), Neurology (V.K.), Biomedical Engi-

neering (T.Z., J.Z.), University of Rochester Medical Center, Rochester, NY.

This work was supported by National Institutes of Health grant NS32024, the Schmitt

Foundation, and the American Alzheimer’s Association.

This work was presented in part at the 90th Scientific Assembly and Annual Meeting of

the Radiological Society of North America as an oral Scientific Presentation; Nov 28–Dec

3, 2004; Chicago, Ill.

Address correspondence to Jianhui Zhong, PhD, 601 Elmwood Ave, Box 648, University of

RochesterMedical Center, Rochester,

rochester.edu

NY14642-8648; e-mail:jianhui.zhong@

1776

Ni ? AJNR 27 ? Sep 2006 ? www.ajnr.org

Page 2

primarily based on the study of Jones,23which determined

“cutoffs”ofDTIreliabilityatspecificNDGDsof20and30for

FA and Trace, respectively. Consequently, NDGD ? 6 (a very

commonly used protocol) and NDGD ? 21 and 31 were used

in different acquisitions in the present study.

In recent clinical applications, in addition to the com-

pound indices (such as mean diffusivity [?D?] and FA, which

are the 2 most widely used DTI indices), the individual eigen-

values (?1, ?2, and ?3) of the diffusion tensor were also used

because they may provide additional information.13-14,25-28

Therefore,weanalyzedinthisstudy5DTIindices:FA,?D?,?1,

?2, and ?3.

Materials and Methods

Subjects

Fifteen healthy adult volunteers (aged 27–61 years; mean age, 38.8

years;9menand6women)participatedinthestudy.Noneofpartic-

ipants had any history of neurologic disorder or brain injury. The

study was approved by the institutional review board. Informed con-

sent was obtained from all subjects after the nature of the study had

been thoroughly described.

MR Imaging Protocols

All MR images were acquired on a GE Signa (Excite 11.0, GE Health-

care, Milwaukee, Wis) 1.5 T MR scanner with a standard quadrature

head coil. In addition to conventional images (3D fast-spoiled gradi-

ent and fluid-attenuated inversion recovery) for whole brain anat-

omy, DTI images with a single-shot pulsed-gradient spin-echo echo-

planar sequence in coronal orientation were obtained. For each

subject,DTIimageswereacquiredbyusing3protocolswithdifferent

combinations of NDGD and NEX for each direction. The 3 DTI ac-

quisitionswereperformedwith6,21,and31noncollinearNDGDand

wereaveraged(imageswereaveragedonthefly)10,3,and2timesfor

each image (ie, NDGD/NEX ? 6/10, 21/3, 31/2). The 3 protocols

resulted in almost the same total number of DW images for each

sectionbeforesignalintensityaveraging(60,63,and62,respectively),

andsimilartotalscantime(9:36,9:04,and8:48minutesandseconds,

respectively). The order of the 3 protocols during acquisitions was

randomized across the subjects to reduce bias in the data, in a single

sessionforeachsubject.Onereference(b?0)imageforeachsection

was acquired for all 3 protocols with same NEX averages. DTI was

acquiredinaninterleavedfashion;ie,1b?0imageforeverysections

was acquired before all diffusion-encoded images in every repeat set.

OtherparametersforDTIwere:repetitiontime/echotime?8000/85

ms,matrix?128?128,FOV?24cm,sectionthickness/gap?3.8/0

mm, and b-factor ? 1000 s/mm2, 28 sections with the center of cor-

pus callosum (CC) as the middle of coverage.

Definitions of DTI Indices

From all diffusion-weighted images, the general diffusion tensor was

first diagonalized, and the yielded scalar invariants of the tensor, in-

cluding diffusion eigenvalues ?1, ?2, and ?3, were derived for each

image pixel. ?1, ?2, and ?3were used to calculate ?D? and FA, which

are defined as

?D? ?1

3??1? ?2? ?3?,

FA ??

2

3

???1? ?D??2? ??2? ?D??2? ??3? ?D??2

??1

2? ?2

2? ?3

2

.

?D?, FA, ?1, ?2, and ?3 were used as DTI indices to compare the 3

protocols. All DTI indices were calculated and corresponding maps

were created with the use of custom software.

Image Postprocessing

Before tensor calculation, images were corrected for motion artifact

and eddy current distortion for each subject with the use of an algo-

rithmproposedbyAnderssonandSkare29thatcorrectsinterprotocol

motion artifacts and eddy current artifacts simultaneously. Image

coregistrationswereperformedamongthe3DTIdatasetswiththeuse

of AIR 5.0 (http://bishopw.loni.ucla.edu/AIR5) to minimize the bias

caused by subject motion during scanning. For each subject, images

without diffusion weighting (b ? 0) in 1 of the 3 protocol datasets

were randomly selected as reference, and images (b ? 0) from the

other 2 protocols were coregistered to this reference. The generated

transformation matrix was then applied to all DW images within

same protocol. After coregistration, regions of interest (ROIs) were

drawn manually on images from 1 protocol selected randomly, and

these ROIs were then translated to the other 2 protocols for calcula-

tionofall5DTIindices.Figure1Aisanexampleofthecoregistration

results among the images of the 3 protocols, with ROI definition for

the posterior portion of CC.

ROI Selection

We selected 6 different ROIs, primarily encompassing white matter

structures with considerably varying anisotropy. These ROIs, which

arecommonlyusedinmanyclinicalDTIstudies,werealsowellvisible

and distinguishable to be easily delineated in colored FA maps. The

different protocols were compared for each of the following 6 white

matter ROIs: callosal fibers, including anterior genu (CCA), middle

body (CCM), and posterior splenium (CCP); association fibers, bi-

lateral superior longitudinal fasciculus (SLF); limbic system fibers,

bilateralcingulum(CIN);andprojectionfibers,internalcapsule(IC).

ROIs were defined with respect to the CC; ie, the ROIs were posi-

tionedat3locations(Fig1B,A,B,C):thecenterofgenuofCC(forCCA),

the center of CC (for CCM, middle CIN, middle SLF, and IC), and the

center of splenium (for CCP, posterior CIN, and posterior SLF). Three

adjacent sections were included for every ROI. The CIN and SLF were

combinedwithbilateralROIsinthemiddleandposteriorlocations,and

theICwascombinedwithbilateralROIsinthemiddleCClocation.We

usedtheAtlasofHumanWhiteMatterAnatomy30asanadditionaltool

fordefiningROIs.EachindividualROIwasmanuallydelineatedbyusing

color-coded FA maps with average numbers of voxels of 748 for CCA,

410forCCM,1336forCCP,371forCIN,1002forSLF,and1275forIC.

AnexampleofthetracingofROIsisshowninFig1B.Manualdelineation

of the ROIs was performed independently by 2 of the authors, and no

significant differences were found between their measurements. Voxels

contaminated with CSF were eliminated with filters for FA ? 0.01, and

?D?1.70?10?3mm2/s.

Data Analyses

Two levels of analyses were performed to test the effects of different

NDGDs from 3 protocols.

ROILevel.Meanvaluesandtheirstandarderrorofmeans(SEMs)

for FA, ?D?, ?1, ?2, and ?3from each ROI were separately analyzed in

a 1-way repeated measures analysis of variance (ANOVA), with

NDGD as a within-subject categoric variable. Greenhouse-Geisser

adjustment31fordegreesoffreedomwasappliedtotheNDGDfactor

because of the inherent violation of the repeated measures assump-

tion of sphericity. Where appropriate, post hoc analyses were con-

BRAIN

ORIGINAL RESEARCH

AJNR Am J Neuroradiol 27:1776–81 ? Sep 2006 ? www.ajnr.org

1777

Page 3

ducted using the Tukey Honestly Significant Difference tests32with a

family-wise error rate of .05.

Voxel Level. We evaluated the similarity among the 3 protocols

with different NDGDs by comparing pair-wise correlation coeffi-

cients (r) for FA, ?D?, ?1, ?2, and ?3values on a voxel-by-voxel basis

for each ROI. For each DTI index and each ROI, we computed 3

correlation coefficients: r21–31(between 21-NDGD and 31-NDGD

protocols), r6–21(between 6-NDGD and 21-NDGD protocols), and

r6–31(between 6-NDGD and 31-NDGD protocols). r6–21, r6–31, and

r21–31were compared in 1-way repeated measures ANOVA for all

ROIs. Finally, we evaluated the correlation coefficients of FA as a

function of mean anisotropy in the ROIs.

Results

For ANOVA analysis at ROI level, the mean values of FA, ?D?,

?1,?2,and?3forthe3protocolsaresummarizedinFig2forall

ROIs. Neither FA nor ?D? showed significant differences

among the 3 protocols (P ? .05) (Fig 2A, -B). However, there

were significant effects of NDGD on the 3 eigenvalues. Post

hoc analyses showed that ?1of the 6-NDGD protocol was

higher than ?1of the 21-NDGD and 31-NDGD protocols in 5

of6ROIs(Fig2C):CCA,CCP,CIN,SLF,andIC(P?.002).?2

of the 6-NDGD protocol was also higher than ?2of the 21-

NDGD and 31-NDGD protocols in 4 of 6 ROIs (Fig 2D):

CCM, SLF, and CIN (P ? .0001), but in CCP only between

6-NDGD and 21-NDGD protocols (P ? .04). In contrast, ?3

of the 6 NDGD protocol was lower than those of the 21-

NDGD and 31-NDGD protocols in 2 of 6 ROIs (Fig 2e): SLF

and CIN (P ? .0001), whereas in CCA and CCP, ?3of the

6-NDGDwassignificantlylowerthanin31-NDGD(P?.02).

Analyses of SEMs in ROIs for FA, ?D?,

?1, ?2, and ?3measures showed no sig-

nificant differences for any of the DTI

indices in any of the ROIs among the 3

protocols.

For ANOVA analyses at the voxel

level,weusedvoxelastherandomvari-

able. Because of the large number of

voxels, ANOVA analyses possessed

such a power that even miniscule, em-

piricallyinsignificant

among the 3 protocols (eg, 1% change

in FA), became significant. Thus, we

evaluated the differences among the 3

protocolsbycorrelatingtheDTIindices

obtained at voxel level for each ROI.

Figure 3 shows correlation coefficients

across the 3 protocols for the 5 DTI in-

differences

dices in all ROIs. Overall, r21–31was always higher than r6–21

and r6–31in all ROIs for FA and ?1,in almost all ROIs for ?2

and ?3(r21–31was only lower than r6–21in IC for ?2and ?3),

but not for ?D?. However, ANOVA and post hoc analyses

showedsignificantlyhigherr21–31thanr6–21andr6–31forFAin

CCA and SLF (P ? .05) (Fig 3A), for ?1in CCA and SLF (P ?

.05) (Fig 3C), and for ?3in CCA and CCM (P ? .03) (Fig 3E).

Analyses also showed significantly higher r21–31than r6–31in

CINforFAandinCCAfor?2(Fig3D)(P?.03).However,for

?D? r6–31was higher than r21–31in CCM and r6–21was higher

than r21–31in IC (P ? .03).

TherelationshipbetweencorrelationcoefficientsofFAand

the mean anisotropy in different ROIs is illustrated in Fig 4.

Linear fitting of r with FA revealed that there was a positive

trend of r21–31, r6–21, and r6–31with increasing anisotropy in

ROIs.Therewasasignificantlinearrelationshipforr21–31and

for r6–31(P ? .05), but not for r6–21(P ? .05).

Discussion

Calculation of diffusion tensor D is based on apparent diffu-

sion coefficient values from each diffusion-weighting direc-

tion. Both increasing NDGD and more averaging of DW im-

ages along each diffusion direction (ie, larger NEX) may

improve estimation of D. So far, theoretic analysis and com-

putersimulations19-24havebeenusedtoinvestigatetheeffects

ofSNRinimagesanddifferentNDGDonquantificationofFA

and?D?.Ourinvivohumanbrainstudyprovidedareal-world

case test for these simulation and numeric studies. In our

study, different combinations of NDGD and NEX are com-

pared,withasimilartotalscanningtime.Fortheprotocolwith

larger NDGD, more DW images are acquired but with less

Fig 1. A, An example of coregistration across 3 protocols.

Images on the left, middle, and right are obtained with the

6-NDGD, 21-NDGD, and 31-NDGD protocols, respectively.

A region of interest (ROI) of the splenium of CC is initially

drawn on only 1 image from 1 of the 3 protocols and then

translated to the corresponding images from the other 2

protocols.

B, Spatial definition of ROIs: all ROIs are positioned

relative to CC in the anterior (A, CCA), middle [B, CCM,

cingulum (D), SLF (E), and IC (F)], and posterior [C, CCP,

cingulum (D), and SLF (E)] locations.

1778

Ni ? AJNR 27 ? Sep 2006 ? www.ajnr.org

Page 4

averaging (ie, a lower SNR) for each DW image; for protocols

with fewer NDGD, there are fewer DW images, but more av-

eraging or a higher SNR for each DW image. First, we found

that at the ROI level, FA and ?D? showed no significant differ-

ence due to the number of diffusion gradient directions. Sec-

ond, we found that at the voxel level, the 3 protocols did not

provide consistent measures, but 21 and 31 diffusion gradient

directions provided more similar measures of FA than the

6-direction protocol.

It is known that noise in DW images introduces errors in

calculated diffusion tensor that propagates through diagonal-

ization into final calculations of eigenvalues, FA, and diffusiv-

ity. It is also generally believed that the larger the NDGD, the

lesser the error. Our ROI-based analysis, however, resulted in

comparable values of FA and ?D? for 3 levels of NDGD, sug-

gesting that 6 or larger NDGDs will generate similar FA and

?D?. A simulation study23found decreasing variation in esti-

mates of FA and trace as the number of sampling directions

increases, but the effect diminishes as SNR for images in-

creases. Another study22showed that the bias of mean FA val-

uescanbereducedmorebyincreasingSNRthanbyincreasing

NDGD; with the SNR0(SNR for b ? 0 image) in the range of

10–100, the relationship between the FA and number of DW

directions was independent. In our study the SNR0was in the

range of 27–60 (lowest for the 31-NDGD protocol with 2 av-

eragesofb?0images,andhighestwiththe6-NDGDprotocol

with 10 averages of b ? 0 images). According to studies that

haveinvestigatedtheissueofeffectsofSNRonDTI,19,22,23,33,34

theSNRsusedinthepresentstudyareintherangewherenoise

variation will have minimal effects on these DTI indices such

as FA and ?D?. With higher SNR, less

variance is expected; however, acquisi-

tion time will be longer and this is

generally beyond what is accepted for

practical clinical use. The protocol pa-

rametersweusedhereweredesignedto

get high SNR with a clinically accept-

able acquisition time.

In contrast with the ROI analyses,

our correlation analyses at the voxel

level showed that r21-31was higher than

r6-21and r6–31for FA, suggesting that

measurements with 21-NDGD are

closer to 31-NDGD than to 6-NDGD

protocols. Some previous simulation

studies23,24have shown that when the

NDGD is larger, there is less uncer-

taintyanderrorpropagation.Thework

by Jones23specifically suggests that at

least 20 unique sampling orientations

are necessary for a robust estimate of

anisotropy. Our results are consistent

with these conclusions suggesting that

increasing NDGD beyond 21 has little

effect on FA, and 21-NDGD is probably sufficient for in vivo

human study of FA. However, for ?D?, the r21–31, r6–21, and

r6–31display random variations in different ROIs, which also

probablysuggeststhatNDGDlessthan31hasnotreachedthe

stable measure for ?D?. This result also supports the view that

at least 30 unique sampling orientations are required for a

robust estimate of mean diffusivity.23

Increased correlations of FA between the 3 protocols with

increase in anisotropy (Fig 4) indicate that the higher the an-

isotropy,thecloserthesimilarityamongprotocolswithdiffer-

ent NDGDs. Furthermore, these findings suggest that DTI

protocols are more reliable in FA estimations from ROIs with

high anisotropy. This finding is again in accordance with the

report23showing that absolute uncertainty of FA decreases

with the increase in anisotropy.

Inourstudy,ROI-basedanalysisshowedsignificantdiffer-

ences between 3 protocols for mean ?1, ?2, and ?3, but not for

FA and ?D?. It is not clear at the moment whether this is be-

cause DTI eigenvalues are more sensitive than ?D? or FA to

biologic variations, or for some other reasons. These differ-

encesamongeigenvaluesmayalsobepartlyrelatedtoSNRs.A

Monte Carlo simulation34showed that the accuracy of the

computed individual eigenvalues is more influenced by noise

contaminationthancompoundindicessuchasFAand?D?.In

the present study, it is likely that the SNRs were not high

enough for ?1, ?2, and ?3calculations, which resulted in de-

tectable effects from NDGD, but were sufficiently high for FA

and ?D?, leading to a minimal effect of NDGD. Intuitively,

calculations of FA and ?D? by combining ?1, ?2, and ?3may

cancelatleastpartiallythevariationofeacheigenvalueamong

Fig 2. Mean value in 6 ROIs for the 3 protocols; A for FA,

B for ?D?, C for ?1, D for ?2, and E for ?3. ?D?, ?1, ?2, and

?3are measured in 10?3mm2/s. White bars for 6-NDGD,

gray bars for 21-NDGD, and black bars for 31-NDGD

protocols. *, P ? .05. Error bars are for 1 SEM.

AJNR Am J Neuroradiol 27:1776–81 ? Sep 2006 ? www.ajnr.org

1779

Page 5

the3protocols.Theresultsonlyshowsignificantdifferencesin

mean ?1, ?2, and ?3between the protocol with 6-NDGD and

the other 2 protocols, but not between protocols with 21-

NDGDand31-NDGD.Inaddition,correlationanalysesatthe

voxel level showed that r21–31was higher than r6–21and r6–31

for ?1, ?2, and ?3in most ROIs. Following the reasoning from

the simulation studies,23,24these results may imply that in-

creasing NDGD beyond 21 has little effect on ?1, ?2, and ?3.

Because we did not use protocols with NDGD between 6 and

21, we cannot provide exact information about the lowest

NDGD necessary for a robust estimate of eigenvalues.

Furthermore,itisnecessarytoaddresstheeffectofmotion

thatmightaffecttheoutcomeofinvivo

studies such as ours. Calculation of re-

liable DTI indices requires combina-

tion of raw images from different ac-

quisitionsin sequential

which in turn makes DTI susceptible

for motion artifacts. In our study, be-

fore calculations of diffusion tensor el-

ements, all non–diffusion-weighted

images and all DW images in different

directionswerecoregisteredwithinand

among the 3 protocols (ie, motion arti-

facts were corrected in both “interpro-

tocol” and “intraprotocol” modes).

This step minimized motion effects

during scanning. However, motion

contaminationsamongdifferentacqui-

sitions in the same gradient direction

foraveragingwerenotcompletelyelim-

inated, because images were output af-

ter simple signal intensity algebraic av-

eraging in a GE clinical scanner;

therefore, motion artifacts were cor-

rected after images were averaged. It is

scanning,

possible, therefore, that the motion effect is more prominent

for the protocol with 6-NDGD than the protocols with 21-

NDGDand31-NDGD,becauseimagecoregistrationwasper-

formedforevery10,3,and2images,respectively,ineachcase.

Futurestudyshouldtrytoseparatetherawdataforindividual

acquisitions before averaging and perform coregistration for

all non–diffusion-weighted images and all DW images in dif-

ferentdirectionsfordifferentacquisitionsin3protocols.Car-

diac gating may also be used to minimize pulsation effects,

another variant of motion artifact.

Thereareusuallylimitationsinthetotalnumberofimages

allowed in each series (on the GE scanner we used for this

study, the limit was 1024 images). Larger NDGDs results in

more images after average because in most scanners, images

areaveragedonthefly.Therefore,thereisatrade-offbetween

NDGDs and number of sections. Protocols with fewer

NDGDs (such as 6) allow for more sections or repetitions,

when the total number of images allowed is limited.

Finally,itshouldbeemphasizedthatDTImeasurementsin

general are very sensitive to even minor differences in hard-

wareandsoftware,acquisitionparameters,andprocessingde-

tails.Theconclusionsfromthisstudywithalimitednumberof

acquisitions protocols on a single scanner should therefore be

taken with precaution. More data may be needed before these

results can be generalized. It is hard or impossible to obtain a

“gold standard” for human study. Without such a standard, it

is impossible to completely evaluate reliability or consistency

of acquisition protocols. More sophisticated and probably

more exhaustive studies may be needed to obtain ultimate

conclusions.

Fig 3. Correlation coefficient (r) in 6 ROIs between 2 of the

3 protocols. A for FA, B for ?D?, C for ?1, D for ?2, and E

for ?3. White bars are for r6–21, gray bars for r6–31, and

black bars for r21–31. *, P ? .05. Error bars are for 1 SEM.

Fig 4. The linear fitting between correlation coefficient of FA versus FA value of different

ROIs.

1780

Ni ? AJNR 27 ? Sep 2006 ? www.ajnr.org

Page 6

Conclusion

In summary, this study suggested that the 3 protocols with 6,

21, and 31 NDGDs generate significant differences for ROI-

based ?1, ?2, and ?3measurements but not for FA and ?D?

measurements. Voxel-based analyses, on the other hand,

showed significant differences of NDGDs for all DTI indices.

Therefore, it is likely that for applications in which only mean

values of FA and ?D? measurements within ROIs are needed

and detailed voxel-based effects can be ignored, the NDGDs

employed do not make much difference as long as the SNR is

above a certain minimum level, and a protocol with 6-NDGD

issufficient.Forapplicationsinwhich?1,?2,and?3measures

withinROIsareneeded,protocolswithmorethan21NDGDs

are necessary for estimation of these eigenvalues. For applica-

tions in which voxel-based information is desired, such as in

longitudinalstudiesofdiseaseprogressormonitoringoftreat-

menteffects,aprotocolwithatleast21NDGDshouldbeused

forestimationofFA,?1,?2,and?3,andaprotocolwithatleast

31 NDGD should be used for estimation of ?D?. Our results

showed that when similar acquisition time is maintained,

NDGD has greater effect than NEX at NDGD lower than 21

for eigenvalues analysis and voxel-based FA analysis and at

NDGD lower than 31 for voxel-based diffusivity analysis.

Clinical DTI protocols should be better designed to balance

thetrade-offsbetweenNDGDandNEX,togetherwithshorter

scantimeandmoresections.Inaddition,forapplicationswith

different ROIs, regions with lower anisotropy may need pro-

tocols with larger NDGDs for reliable FA estimation. Because

thestudyusedalimitednumberofacquisitionsprotocolsona

single scanner, precaution should be taken in generalizing the

above conclusions.

Acknowledgments

We appreciate Robert Ambrosini’s help in coregistration of

images.

References

1. Chepuri NB, Yen YF, Burdette JH, et al. Diffusion anisotropy in the corpus

callosum. AJNR Am J Neuroradiol 2002;23:803–08

2. Takahashi S, Yonezawa H, Takahashi J, et al. Selective reduction of diffusion

anisotropyinwhitematterofAlzheimerdiseasebrainsmeasuredby3.0Tesla

magnetic resonance imaging. Neurosci Lett 2002;332:45–48

3. Fellgiebel A, Wille P, Muller MJ, et al. Ultrastructural hippocampal and white

matter alterations in mild cognitive impairment: a diffusion tensor imaging

study. Dement Geriatr Cogn Disord 2004;18:101–08

4. ZhongJ,NiH,ZhuT,etal.MRdiffusiontensorimaging(DTI)andneuropsy-

chological testing for neuronal connectivity in Alzheimer’s disease (AD) pa-

tients. Proceedings of SPIE: Medical Imaging 2004—Physiology, Function, and

Structure from Medical Images 2004;5369:238–49

5. Coombs BD, Best A, Brown MS, et al. Multiple sclerosis pathology in the nor-

mal and abnormal appearing white matter of the corpus callosum by diffu-

sion tensor imaging. Mult Scler 2004;10:392–97

6. Kealey SM, Kim Y, Provenzale JM. Redefinition of multiple sclerosis plaque

size using diffusion tensor MRI. AJR Am J Roentgenol 2004;183:497–503

7. Cassol E, Ranjeva JP, Ibarrola D, et al. Diffusion tensor imaging in multiple

sclerosis: a tool for monitoring changes in normal-appearing white matter.

Mult Scler 2004;10:188–96

8. Ciccarelli O, Werring DJ, Barker GJ, et al. A study of the mechanisms of nor-

mal-appearingwhitematterdamageinmultiplesclerosisusingdiffusionten-

sor imaging—evidence of Wallerian degeneration. J Neurol 2003;250:287–92

9. Bozzali M, Cercignani M, Sormani MP, et al. Quantification of brain gray

matter damage in different MS phenotypes by use of diffusion tensor MR

imaging. AJNR Am J Neuroradiol 2002;23:985–88

10. Guo AC, MacFall JR, Provenzale JM. Multiple sclerosis: diffusion tensor MR

imaging for evaluation of normal-appearing white matter. Radiology 2002;

222:729–36

11. GuoAC,JewellsVL,ProvenzaleJM.Analysisofnormal-appearingwhitemat-

ter in multiple sclerosis: comparison of diffusion tensor MR imaging and

magnetization transfer imaging. AJNR Am J Neuroradiol 2001;22:1893–900

12. Cercignani M, Inglese M, Pagani E, et al. Mean diffusivity and fractional an-

isotropy histograms of patients with multiple sclerosis. AJNR Am J Neurora-

diol 2001;22:952–58

13. Oh J, Henry RG, Genain C, et al. Mechanisms of normal appearing corpus

callosum injury related to pericallosal T1 lesions in multiple sclerosis using

directional diffusion tensor and 1H MRS imaging. J Neurol Neurosurg Psychi-

atry 2004;75:1281–86

14. Henry RG, Oh J, Nelson SJ, et al. Directional diffusion in relapsing-remitting

multiple sclerosis: a possible in vivo signature of Wallerian degeneration. J

Magn Reson Imaging 2003;18:420–26

15. Clark CA, Hedehus M, Moseley ME. Diffusion time dependence of the appar-

ent diffusion tensor in healthy human brain and white matter disease. Magn

Reson Med 2001;45:1126–29

16. Ragin AB, Storey P, Cohen BA, et al. Whole brain diffusion tensor imaging in

HIV-associated cognitive impairment. AJNR Am J Neuroradiol 2004;25:195–

200

17. Pomara N, Crandall DT, Choi SJ, et al. White matter abnormalities in HIV-1

infection: a diffusion tensor imaging study. Psychiatry Res 2001;106:15–24

18. Filippi CG, Ulug AM, Ryan E, et al. Diffusion tensor imaging of patients with

HIV and normal-appearing white matter on MR images of the brain. AJNR

Am J Neuroradiol 2001;22:277–83

19. Papadakis NG, Xing D, Houston GC, et al. A study of rotational invariant and

symmetric indices of diffusion anisotropy. Magn Reson Imaging 1999;17:

881–92

20. Skare S, Hedehus M, Moseley ME, et al. Condition number as a measure of

noise performance of diffusion tensor data acquisition schemes with MRI. J

Magn Reson 2000;147:340–52

21. Hasan KM, Parker DL, Alexander AL. Comparison of gradient encoding

schemes for diffusion-tensor MRI. J Magn Reson Imaging 2001;13:769–80

22. Papadakis NG, Murrills CD, Hall LD, et al. Minimal gradient encoding for

robust estimation of diffusion anisotropy. Magn Reson Imaging 2000;18:

671–79

23. JonesDK.Theeffectofgradientsamplingschemesonmeasuresderivedfrom

diffusion tensor MRI: a Monte Carlo study. Magn Reson Med 2004;51:807–15

24. Poonawalla AH, Zhou XJ. Analytical error propagation in diffusion anisot-

ropy calculations. J Magn Reson Imaging 2004;19:489–98

25. Song SK, Sun SW, Ramsbottom MJ, et al. Dysmyelination revealed through

MRIasincreasedradial(butunchangedaxial)diffusionofwater.NeuroImag-

ing 2002;17:1429–36

26. Peng SS, Tseng WY, Chien YH, et al. Diffusion tensor images in children with

early-treated, chronic, malignant phenylketonuric: correlation with intelli-

gence assessment. AJNR Am J Neuroradiol 2004;25:1569–74

27. ThomallaG,GlaucheV,KochMA,etal.Diffusiontensorimagingdetectsearly

Wallerian degeneration of the pyramidal tract after ischemic stroke. Neuro-

image 2004;22:1767–74

28. PengSS,LeeWT,WangYH,etal.Cerebraldiffusiontensorimagesinchildren

with tuberous sclerosis: a preliminary report. Pediatr Radiol 2004;34:387–92

29. AnderssonJH,SkareS.Amodel-basedmethodforretrospectivecorrectionof

geometric distortions in diffusion-weighted EPI. Neuroimage 2002;16:177–99

30. WakanaS,JiangH,Nagae-PoetscherLM,etal.Fibertract-basedatlasofhuman

white matter anatomy. Radiology 2004;230:77–87

31. Greenhouse SW, Geisser S. On methods in the analysis of profile data. Psy-

hometrika 1959;24:95–112

32. KeselmanHJ,RoganJC.TheTukeymultiplecomparisontest:1953–1976.Psy-

chological Bulletin 1977;84:1050–56

33. Anderson AW. Theoretical analysis of the effects of noise on diffusion tensor

imaging. Magn Reson Med 2001;46:1174–88

34. Pierpaoli C, Basser PJ. Toward a quantitative assessment of diffusion anisot-

ropy. Magn Reson Med 1996;36:893–906

AJNR Am J Neuroradiol 27:1776–81 ? Sep 2006 ? www.ajnr.org

1781

#### View other sources

#### Hide other sources

- Available from Sven Ekholm · May 22, 2014
- Available from ajnr.org