Q10

Ultrasound in Med. & Biol., Vol. -, No. -, pp. 1–12, 2013Copyright � 2013 World Federation for Ultrasound in Medicine & Biology

Printed in the USA. All rights reserved0301-5629/$ - see front matter

/j.ultrasmedbio.2013.09.006

123456789

1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162

6364

http://dx.doi.org/10.1016

6566676869707172737475767778798081828384

d Original Contribution

3-D TRANSCRANIAL ULTRASOUND IMAGING WITH BILATERAL PHASEABERRATION CORRECTION OF MULTIPLE ISOPLANATIC PATCHES: A PILOT

HUMAN STUDY WITH MICROBUBBLE CONTRAST ENHANCEMENT

BROOKS D. LINDSEY,*y HEATHER A. NICOLETTO,z ELLEN R. BENNETT,z DANIEL T. LASKOWITZ,z

and STEPHEN W. SMITH**Department of Biomedical Engineering, Duke University, Durham, North Carolina, USA; yJoint Department of Biomedical

Engineering, University of North Carolina and North Carolina State University, Chapel Hill, North Carolina, USA; andzDivision of Neurology, Duke University Medical Center, Durham, North Carolina, USA

(Received 22 May 2013; revised 20 August 2013; in final form 1 September 2013)

AHall,brooks

858687888990919293949596979899

100

Abstract—With stroke currently the second-leading cause of death globally, and 87% of all strokes classified asischemic, the development of a fast, accessible, cost-effective approach for imaging occlusive stroke could havea significant impact on health care outcomes and costs. Although clinical examination and standard computedtomography alone do not provide adequate information for understanding the complex temporal events that occurduring an ischemic stroke, ultrasound imaging is well suited to the task of examining blood flow dynamics in realtime andmay allow for localization of a clot. A prototype bilateral 3-D ultrasound imaging system using twomatrixarray probes on either side of the head allows for correction of skull-induced aberration throughout two entirephased array imaging volumes. We investigated the feasibility of applying this custom correction technique infive healthy volunteers with Definity microbubble contrast enhancement. Subjects were scanned simultaneouslyvia both temporal acoustic windows in 3-D color flow mode. The number of color flow voxels above a commonthreshold increased as a result of aberration correction in five of five subjects, with a mean increase of 33.9%.The percentage of large arteries visualized by 3-D color Doppler imaging increased from 46% without aberrationcorrection to 60% with aberration correction. (E-mail: brooksdlindsey@gmail.com) � 2013 World Federationfor Ultrasound in Medicine & Biology.

Key Words: Transcranial, Phase aberration, 3-D ultrasound, Contrast enhanced.

101102103104105106107108109110111112113114115116117118119

INTRODUCTION

Stroke and cerebrovascular disease represent the second-leading cause of death globally (World HealthOrganization 2011) and constitute 17% of overall nationalhealth spending in the United States (Heidenreich et al.2011). Approximately 87% of strokes are ischemicstrokes (Roger et al. 2011), in which patients’ functionaloutcomes may be improved and recovery times reducedby intravenous administration of a thrombolytic agentsuch as recombinant tissue plasminogen activator (tPA)(Hacke et al. 1995).

Ischemic strokes may be treated with thrombolyticagents only within the first 4.5 h of symptom onset,with earlier times associated with more favorableoutcomes (Davis and Donnan 2009; Hacke et al. 2008).

ddress correspondence to: Brooks D. Lindsey, Room 331 TaylorCB #7575, Chapel Hill, NC 27599, USA. E-mail:dlindsey@gmail.com

1

FLA 5.2.0 DTD � UMB9643_proof �

120121122123124

Although computed tomography (CT) scans arecommonly used to exclude patients with hemorrhagesfrom thrombolytic treatment, clinical examination andstandard CT alone do not provide adequate informationfor understanding the complex temporal events thatoccur during an ischemic stroke (Sims et al. 2005). Ultra-sound imaging is well suited to the task of examiningblood flow dynamics in real time and may allow for local-ization of the ischemic region in a large-artery stroke(Kern et al. 2011).

Real-time 2-D ultrasound imaging has been found tobe a portable technology for diagnosing stroke in the field(Bowman 2010; Chenaitia et al. 2011; Holscher 2012;Holscher et al. 2008; Hoyer et al. 2010; Schlachetzkiet al. 2012; Wilson et al. 2011). In the clinical setting,investigators have worked to standardize scanningprocedures for transcranial ultrasound with color flow(Nedelmann et al. 2009). Three-dimensional ultrasoundhas been cited as a potentially faster and clinically accept-able alternative to X-ray for examining blood flow in

4 October 2013 � 2:14 am � ce

print&

web4C=FPO

Fig. 1. A sound wave exiting the skull propagates through thebrain with speed cB and through the skull with speed cS. Onexiting the skull, the wavefront is incident on the elements ofa matrix array (dotted outlines). The delays at each elementDt(x, y) are a function of the spatial variation of the skull’s

thickness, Dz(x, y).

2 Ultrasound in Medicine and Biology Volume -, Number -, 2013

125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186

187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227

intracranial collateral arteries (Wessels et al. 2004).Ultrasound may also have value as a monitoring tool toavoid repeated CT scans, each of which requires transpor-tation of a critically ill patient, exposure to ionizing radi-ation and increased cost relative to ultrasound(Alexandrov et al. 2004; Becker et al. 2012).

The frame rates and spatial sampling necessaryfor real-time 3-D ultrasound imaging are enabled by theuse of a matrix array probe and parallel receive processing(Smith et al. 1991; von Ramm et al. 1991). This approachcarries inherent reductions in system performance relativeto 2-D ultrasound: (i) Sensitivity is reduced because ofelectrical impedance mismatch between the element ofa 2-D array and the cable it drives. (ii) Spatial resolutionand sensitivity are reduced because of weak focusing ofa transmit beam from which multiple receive beams areformed simultaneously.

In addition to these technical challenges associatedwith 3-D ultrasound imaging, the effects of the skull alsolimit transcranial ultrasound imaging. Aaslid et al. (1982)first reported the presence of a region of the humantemporal bone having decreased attenuation. Within thisacoustic window, attenuation of sound waves in the skullis approximately 2.8 dB/cm/MHz (White and Stevenson1978), whereas outside the window, it may vary from 30to 70 dB/cm/MHz (White and Stevenson 1978). In addi-tion, recent studies have examined other deleterious effectsof ultrasound propagation through the skull, includingmode conversion and total internal reflection (Clementet al. 2004; Maciak et al. 2009; Vignon et al. 2010;White et al. 2006), non-linear propagation (Pinton et al.2011) and refraction (Smith et al. 1986). Researchershave also presented promising results using microbubble-based parametric imaging approaches to localize regionsof abnormal perfusion (Holscher et al. 2005; Kern et al.2011; Meyer-Wiethe et al. 2009; Seidel et al. 2000).

228229230231232233234235236237238239240241242243244245246247248

Phase aberration correctionPerhaps the most commonly addressed skull-

induced effect is aberration, which reduces spatial andcontrast resolution caused by a mismatch between thevelocity of a longitudinal wave in the spatially inhomoge-neous skull (cSz 2327–2650 m/s) (Fry and Barger 1978;Pichardo et al. 2011) and the soft tissue velocity assumedby the scanner in beamforming (cB 5 1540 m/s). In thenear-field phase screen model (Fig. 1), the wavefrontaberration in the skull is modeled as a set of time delayadjustments at each element

Dtðx; yÞ5Dzðx; yÞ�

1

CS

21

CB

�(1)

whereDz is the thickness variation in x and y, andDt is theresulting arrival time variation.

FLA 5.2.0 DTD � UMB9643_proof �

If the spatial variations of the skull can be measuredusing either backscattered echoes or a transmittedbeacon wave, then the scanner’s transmit and receivetime delays might be corrected by adding a–Dt to cancelthe appropriate Dt at each (x, y) position in eqn (1). Thistightens the focus that had been broadened by propaga-tion through the skull. Miller-Jones (1980) measuredarrival time using a contralateral active source asa coherent, high-signal-to-noise ratio (SNR) correctionbeacon, an approach that Vignon et al. combined withinverse filtering to compensate for the variousfrequency-dependent effects of the skull and allow forthe formation of a tight focus within the skull withoutthe presence of a beacon inside the skull (Aubry et al.2001; Tanter et al. 2001; Vignon et al. 2006). Our owngroup reported 3-D correction of in vivo aberration ona 2-D array using a multi-lag cross-correlation tech-nique (Ivancevich et al. 2008, 2009). All of theseadaptive imaging techniques suffer from limitedspatial stability of the aberrator, in which correctingfor a single aberrator restores spatial resolution andcontrast only over a small sub-region of the image,

4 October 2013 � 2:14 am � ce

print&web4C=FPO

Fig. 2. In the multi-isoplanatic patch phase aberration correc-tion technique, unfocused waves are steered off axis to investi-gate several unique propagation paths through the near-fieldskull aberrator, indicated by differently colored rays. Each colorcorresponds to a subset of image lines within a 3-D volume. Inthis work, five isoplanatic patches are estimated and corrected.The central update region (red) is 32� 3 32�; outer regions are

16� 3 32� or 32� 3 16�.

3-D transcranial US with aberration correctionQ1 d B. D. LINDSEY et al. 3

249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310

311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372

known as the isoplanatic patch (Dahl et al. 2005; Vignonet al. 2006).

Aberration correction techniques have been estab-lished in other fields including astronomy (Muller andBuffington 1974), electromagnetics (Steinberg 1981)and seismology (Taner et al. 1974), which have informedaberration correction in medical ultrasound. Approachesin astronomy, ultrasound and microscopy have usedbeacon signals similar to those in the current method(Andersen et al. 1997; Azucena et al. 2010; Flax andO’Donnell 1988; Haworth et al. 2008). However, thecurrent technique is unique among techniques in othermodalities in its use of reciprocity between twoindependent 3-D scans. One proposed solution for the iso-planatic patch in other fields is the use of multi-conjugateadaptive optics, in which multiple deformable mirrors areused to characterize multiple aberrating layers, where thenumber of deformable mirrors is equal to the number ofthin aberrating layers (Johnston and Welsh 1994). Thisapproach has been implemented to correct lens aberra-tions in optical imaging of the human retina (Bedggoodet al. 2006; Dubinin et al. 2008) and atmosphericaberrations in the case of large ground-based telescopes(Beckers 1988; Marchetti et al. 2007). Such an approachis physically similar to the matrix representation ofaberration developed in spatiotemporal inverse filteringtechniques (Aubry et al. 2001; Tanter et al. 2001), inwhich the entire propagation operator must be acquired,with each row of this matrix corresponding to a differentlayer of distortion.

Despite the sensitivity challenges posed by 3-Dultrasound and the presence of the skull, our group has re-ported moderate success in 3-D transcranial imagingstudies. Although we have reported the ability to imageblood flow transcranially in three dimensions in someindividuals (Lindsey and Smith 2013; Lindsey et al.2011), we require microbubble contrast agent toachieve sensitivity levels that are diagnostically usefulin a general population.

Recently, our group reported a technique for phaseaberration correction using two 2-D arrays placed overopposing temporal acoustic windows capable of correct-ing multiple isoplanatic patches in a 3-D phased arrayscan (Lindsey and Smith 2013). The presence of twoarrays allows us to sequentially transmit multiple steeredwavefronts and receive them on the opposite side of thehead after they have traveled unique propagation pathsthrough the aberrating media (Fig. 2). These measure-ments are then used to apply unique delay update mapsto different sub-regions of a 3-D phased array scan,correcting multiple isoplanatic patches and extendingthe benefits of phase aberration correction to an entirevolume (Fig. 3). This technique builds on the investiga-tions of Waag et al. into propagation path dependence of

FLA 5.2.0 DTD � UMB9643_proof �

aberration, in which measurements made using multiplesets of transmit-receive aperture pairs are used to correctfor path-dependent effects that cause the near-field phasescreen model to break down (Astheimer et al. 2006; Tillettet al. 2010; Waag and Astheimer 2005).

In this article, we describe the first results of thehuman application of this phase aberration correctiontechnique in a pilot study of five healthy volunteerswith microbubble contrast enhancement.

METHODS

Scanning systemThe study was performed using the Volumetrics

Model 1 scanner (Volumetrics Medical Imaging,Durham, NC, USA), a real-time 3-D ultrasound systemwith 16:1 parallel receive processing (Smith et al. 1991;von Ramm et al. 1991). This system has 256 transmitchannels and 256 shared transmit/receive channels. A2-D phased array transducer on this system typicallytransmits 256 broadened beams in succession. For eachtransmit beam, 16 receive beams are formed from echoesarranged in a 43 4 pattern centered on the transmit beam.The transmit beam spacing is 4� in a typical 64� 3 64�

scan, enabling frame rates of up to 30 volumes/s.To enable simultaneous real-time 3-D imaging with

two 2-D arrays, the scanner’s channels and its 4096 imagelines were split equally between two matrix arrays,enabling simultaneous acquisition of two 64� 3 64� pyra-midal volumes (Lindsey et al. 2011; Smith et al. 2009).Each matrix array has 128 transmit elements and 128shared transmit/receive elements. For each transmitbeam, 16 receive beams are formed from one of thetwo transducers. However, transmit beam separation has

4 October 2013 � 2:14 am � ce

Q2

print&web4C=FPO

Fig. 3. Percentage brightness increase with lateral distance intranscranial aberration correction in a single human subjectfor a traditional phase aberration correction (black) anda multiple isoplanatic patch correction (blue). Although thebrightness improvement decreases laterally when a single iso-planatic (IP) patch is assumed, brightness gains are maintainedwhen multiple corrections are performed, as described in the

caption to Figure 2.

print&web4C=FPO

Fig. 4. (a) Typical volume interrogated during a transtemporalexamination. Two such volumes are acquired simultaneouslyand are then registered and fused offline into a single visualiza-tion. (b) Head frame used to position two arrays over bilateral

temporal bone windows.

4 Ultrasound in Medicine and Biology Volume -, Number -, 2013

373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434

435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496

now doubled in the elevation direction, so transmit beamsin each 64� 3 64� volume are spaced at 4� in azimuth and8� in elevation.

During in vivo scanning, four slices from the twoprobes are displayed simultaneously in real time: one azi-muth and one elevation slice from each transducer, corre-sponding to coronal and transverse planes intranstemporal imaging (Fig. 4a). The operator usesa trackball control to select any four slices in eithervolume for display. All experiments were performedusing commercial (Volumetrics) sparse 2-D arrays havinginter-element separation of 0.35 3 0.35 mm and remap-ped by a custom printed circuit board to allow for dualsimultaneous 3-D imaging (Lindsey et al. 2011). Eachsparse matrix array has 128 elements with an aperturediameter of 6.6 mm, producing a –6-dB pulse-echo beam-width of 5.5 mm at a depth of 70 mm (Smith et al. 2009);studies were performed with 3-D echo at 2.5 MHz and3-D color Doppler at 1.8 MHz to maximize sensitivityfor the system and transducer.

Human scanning protocolFive healthy volunteers were scanned after

providing informed consent according to a protocolapproved by the Duke University institutional reviewboard. The two matrix array probes were placed ina head frame (Fig. 4b) (Mark III, Spencer Technologies,

FLA 5.2.0 DTD � UMB9643_proof �

Seattle, WA, USA), allowing an experienced sonographer(H.N.) to manipulate them and locate the subject’sacoustic window. Once the desired field of view was at-tained, the probes were locked in place using customattachments on the head frame. The subject was in-structed to remain still for the duration of the study.

For each probe, five sequential pitch-catch measure-ments were made using transmitted beacon waves steeredin five different directions: (0�, –15�), (0�, 0�), (0�, 15�),(15�, 0�) and (15, 0�). Update region (isoplanatic patch)size was assumed to be 32� in both azimuth and elevation,in line with published values for the skull isoplanaticpatch (Ivancevich et al. 2009; Vignon et al. 2006). Thisassumption yields five update regions: a central32� 3 32� patch and four side patches measuring either32� 3 16� or 16� 3 32� (Fig. 2). This aberration correc-tion approach has been previously described in greaterdetail (Lindsey and Smith 2013).

Briefly, single-channel radiofrequency data wereacquired using a digitizing board (25 MHz, PDA14, Sig-natec, Corona, CA, USA) on a PC (Dell, Round Rock,TX, USA). These signals were filtered axially (FIR band-pass filter, 60%BW) and laterally (Dahl and Feehan 2008)(FIR lowpass, cutoff at 75% of spatial Nyquist), beforeremoval of the steering component. Phase aberrationswere estimated using a multi-lag least-squares estimationtechnique (Gauss et al. 2001; Liu and Waag 1994). Thisestimation algorithm is performed by computing thenormalized cross-correlation between all element signalswithin a specified spatial lag (3 mm), and the resultingover-determined system was approximately solved usinglinear least squares, yielding five delay update maps, orversions of Dt(x,y), for each probe. These maps werepassed back to the scanner via Ethernet connection, andtransmit and receive delays were updated for all scanningmodes. The entire computation and delay update processrequired approximately 3 min. Although a faster imple-mentation could further improve this technique byminimizing effects of physiologic and patient motion,we successfully demonstrated correction (Lindsey andSmith 2013) because the transducer was stationary

4 October 2013 � 2:14 am � ce

3-D transcranial US with aberration correction d B. D. LINDSEY et al. 5

497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620

relative to the aberrator and the correction source wasa coherent, spatially stationary signal rather than a diffusesource or moving scatterers. After the human study, thedelay maps from each subject were analyzed to determinethe root-mean-square amplitude and full width at half-maximum of the autocorrelation of each measured aberra-tor, measurements of aberration strength and spatialfrequency content, respectively.

Once delays were updated, the subject was givenan intravenous bolus injection of 10 mL/kg activatedDefinity (Lantheus Medical Imaging, North Billerica,MA, USA) followed by a 10-mL saline flush accordingto the Definity prescribing information. Aberration-corrected images were acquired in color flow mode.After the flow enhancement caused by Definity haddissipated, the scanner’s delays were returned to theirdefault values (i.e., aberration correction was removed).A second bolus injection was administered per pre-scribing information, and control images were acquired.During microbubble enhancement, four corrected andfour control acquisitions were saved. Each acquisitionconsisted of approximately 10–12 volumes in time ata frame rate of approximately 3 volumes/s. Three-dimensional Doppler settings included a four-cyclepulse and an ensemble length of 7. The Doppler pulsesequence followed that of Jensen (1996), in whichecho and Doppler pulses are inter-leaved. The VMI on-screen display showed a mechanical index of 1.2 anda thermal index of 0.1. Because on-screen values assume0.3 dB/cm/MHz (American Institute of Ultrasound inMedicine [AIUM]/National Electrical ManufacturersAssociation [NEMA] 1998) and actual attenuation is ex-pected to be approximately 3 dB/cm/MHz for 3 mm ofbone inside the window and 1 dB/cm/MHz for 7 cm ofbrain tissue, maximum in vivo attenuated peak pressuresare expected to be approximately three times lower thanindicated, ensuring subject tolerability for this studywhile also providing potential room for increasing themechanical index in future studies. The real-time videodisplay from the entire study—in which two orthogonalslices from each transducer are displayed simulta-neously—was also recorded to a digital video recorderdisk for later review.

For quantitative comparison, the magnitude of the3-D color Doppler data of the control scans was analyzedoffline (MATLAB, The Mathworks, Natick, MA, USA)by manually selecting a threshold for 3-D renderingthat produced optimal display of signal and rejection ofnoise. This is analogous to adjusting the Doppler rejecton the scanner. This same threshold was then applied tothe corrected volumes and the change in the number ofvoxels above this threshold was recorded. This increasein Doppler magnitude is indicative of improvement insensitivity caused by aberration correction, because

FLA 5.2.0 DTD � UMB9643_proof �

when aberration is corrected in each isoplanatic patch,the radiofrequency echo data are summed coherently,increasing echo SNR and producing a correspondingincrease in the correlation coefficient (the Dopplermagnitude) between beamformed echo acquisitions atsuccessive pulse repetition intervals.

Mean variance of the Doppler data was also com-pared between the corrected and control cases. Varianceis expected to decrease with correction as coherentsignals overcome random noise (Ivancevich et al. 2008).

To display 3-D data, simultaneously acquiredvolumes were averaged in time and registered and fusedinto a single 3-D visualization as described previously(Lindsey et al. 2011, 2013). Video recordings of thereal-time display were reviewed by an author experienced(B.L.) in viewing this type of bilateral 3-D scan, anddetection of the following major vessels for each subjectwas evaluated: left and right middle cerebral arteries, leftand right internal carotid arteries, anterior cerebral artery,left and right posterior cerebral arteries.

RESULTS

The anatomy of the vessels appeared to be accurateand was not greatly affected by aberration correction.Three-dimensional color Doppler data characteristicswith and without aberration correction are summarizedin Table 1. The number of Doppler voxels abovea common threshold increased in all five subjects asa result of correcting aberration, with a mean increaseof 33.9 6 40.6%. The same rendering threshold wasused for the aberrated (control) and corrected cases withineach subject to allow direct comparison. However, thevariance of the Doppler data decreased as a result of aber-ration correction in only two of five subjects. This will beexamined in greater detail in the Discussion.

One subject exhibited a window failure, present in8%–29% of the population (Baumgartner et al. 1997;Gahn et al. 2000; Hashimoto et al. 1992; Krejza et al.2007; Lindsey et al. 2013; Marinoni et al. 1997; Postertet al. 1997; Seidel et al. 1995). In this subject, nolandmarks were visible on the B-mode echo image, andonly noise was visible with contrast enhancement, butwithout aberration correction (Fig. 5a). With aberrationcorrection, blood flow in a single middle cerebral arterywas visualized (Fig. 5b).

In evaluating the detectability of the major cerebralarteries (Table 2), contrast enhancement without aberra-tion correction allowed visualization of 16 of 35 possiblemajor arteries (45.7%). Contrast enhancement with aber-ration correction allowed visualization of 21 of 35possible major arteries (60.0%). Ability to image bloodflow in large cerebral arteries is an important indicatorof any imaging system for use in stroke given that 68%

4 October 2013 � 2:14 am � ce

Table 1. Percentage changes in Doppler magnitude andvariance resulting from aberration correction in five

healthy volunteers

Subject1

Subject2

Subject3

Subject4 (windowfailure)

Subject5

Increase in Dopplervoxels abovecommon threshold

105.4% 23.6% 16.6% 19.6% 4.29%

Decrease in Dopplervariance

–18.4% 10.9% –0.353% –0.652% 8.386%

print&

web4C=FPO

6 Ultrasound in Medicine and Biology Volume -, Number -, 2013

621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682

683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728

of ischemic strokes affect the anterior circulation—arteries supplied by the left and right internal carotidarteries—with 96% of these occlusions occurring in themiddle cerebral arteries (Bogousslavsky et al. 1988).

Results for two subjects in whom aberration correc-tion improved visualization are illustrated in Figures 6and 7. In Figure 6, aberration correction enables visuali-zation of both proximal middle cerebral arteries, whichhad not been possible without correction, and revealsthe S-shaped structure of the carotid siphon in threedimensions. In Figure 7, aberration correction ina different subject illuminates the left internal carotidartery and produces more continuous flow in the rightinternal carotid artery. These improvements are discussedfurther in the Discussion.

Finally, in Figure 8, the entire circle of Willis is visu-alized both before and after correction in a fourth subject.Although in this subject all major arteries were visiblewithout aberration correction, correction improved conti-nuity of flow imaging by raising the SNR for Dopplerimaging as described under Methods. In particular, incomparing Figure 8(c and d), correction can be seen to re-move a registration error involving the anterior cerebralartery and to improve imaging of continuous flow in bothmiddle cerebral arteries and one posterior cerebral artery.

In Figure 9, characteristics of the measured aberra-tors are presented for all five subjects. For each subject,five unique propagation paths through the aberrating skull

Fig. 5. In a single subject with window failure, aberration corartery (MCA) and visualization of flow in the proximal porti

possible to distinguish betw

FLA 5.2.0 DTD � UMB9643_proof �

(i.e., five isoplanatic patches) were measured on each sideof the head. The strength of an aberrator is commonly ex-pressed as the root-mean-square (RMS) amplitude of theaberration map, whereas the spatial frequency content ofan aberrator is expressed as the full width at half-maximum (FWHM) of the autocorrelation of the aberra-tion map. In Figure 9(a and b), the measured mean andstandard deviation of the RMS aberration amplitudefrom five isoplanatic patch measurements are shown forthe left and right sides of the head for each subject. Alarger standard deviation indicates a greater variation inaberrator strength among different isoplanatic patcheswithin a single subject. Measured FWHM correlationlengths are also provided for the left and right sides ofthe head for each subject (Fig. 9c and d). A larger stan-dard deviation indicates greater variation in themaximum spatial frequency content among different iso-planatic patches within a single subject.

DISCUSSION

As stated in the Results, the aberration correctiontechnique employed resulted in a mean increase in thenumber of Doppler voxels above the display thresholdof 33.9%. Results in an initial human study with twopatients did not use microbubble contrast agent andproduced a mean increase in color flow voxels of 92%.The present study with contrast agent produced a Dopplersignal with higher pre-correction SNR, providing a moreaccurate evaluation of the aberration correction techniquedescribed. Although all five subjects had an increase invoxels as a result of aberration correction, the varianceof the increase in Doppler voxels was large because ofthe small number of patients scanned and the fact thatin one subject (subject 1), the improvement was substan-tial (Fig. 6). In the results presented, the mean variancedid not decrease in many cases, which could indicateeither that the aberration correction failed or that thismetric is unsuitable for measuring improvements causedby aberration correction in in vivo Doppler data. This

rection allowed identification of the left middle cerebralon of this vessel (b). Without correction (a), it was noteen vessel and noise.

4 October 2013 � 2:14 am � ce

729730731732733734735736737738739740741742743744

Table 2. Rates of detection of major cerebral arteries infive healthy volunteers

Without aberrationcorrection

With aberrationcorrection

Left ICA 2/5 4/5Right ICA 3/5 4/5Left MCA 3/5 5/5Right MCA 4/5 4/5ACA 2/5 2/5Left PCA 1/5 1/5Right PCA 1/5 1/5Echo landmarks 4/5 4/5

ACA 5 anterior cerebral artery; ICA 5 internal carotid artery;MCA 5 middle cerebral artery; PCA 5 posterior cerebral artery.

print&

web4C=FPO

3-D transcranial US with aberration correction d B. D. LINDSEY et al. 7

745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806

807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846

metric was used because it has previously been used(Ivancevich et al. 2006, 2008; Lindsey and Smith 2013)and because there is no standard metric for evaluatingthe effect of aberration correction on Doppler data.When this study and the previous study (Lindsey andSmith 2013) are combined, variance decreased in fourof seven subjects. Presented images suggest that aberra-tion correction did, in fact, improve image quality, soit is possible that correcting aberration may not alwaysyield a decrease in Doppler variance because of eitherincreasing amplitude of vessel wall echoes or anincreased sensitivity to complex flow dynamics aftercorrection. The effect of aberration correction on Dopplervariance requires further study.

Though microbubble contrast agent was used in thiswork, we would prefer to image without the use ofcontrast agents to ease logistic requirements of scanningin the field. However, previous studies investigating thetolerability of contrast-enhanced diagnostic transcranialultrasound imaging have found no indication of tissue

Fig. 6. In an anterior-posterior view of a single subject (subjectimal middle cerebral artery (MCA) bilaterally where these vessalso brings out the S-shaped curve of the carotid siphon (segm

FLA 5.2.0 DTD � UMB9643_proof �

damage or blood-brain barrier disruption for pressuresup to 2.69 MPa (Schlachetzki et al. 2002).

Although the magnitude of the Doppler data—usedclinically in power Doppler and to weight or thresholdcolor Doppler velocity data—exhibited the improvementexpected with aberration correction, the variance inDoppler data decreased in only two of five subjects.Although variance was expected to decrease throughoutthe entire volume, it should be noted that within a singleisoplanatic patch, correction quality is non-uniform andis expected to decrease moving from the center to theedge of an isoplanatic patch in the same way thatimprovement caused by correction diminished withlateral distance when only a single patch was assumedthroughout an entire 3-D volume (Fig. 3). Doppler vari-ance is also expected to vary temporally throughout thecardiac cycle. From a scanning system perspective, theability to measure these spatial and temporal dependen-cies of the variance of the Doppler data is limitedspatially by line separation and range gates and tempo-rally by ensemble lengths. These factors cause ultra-sound color flow data to be coarsely sampled relativeto B-mode data, making color flow statistics moresusceptible to and less able to measure variations athigh spatial or temporal frequencies. Perhaps relativelylow frame rates contribute to overall high variances inthis particular imaging system. Further investigationwith larger numbers of both inter- and intra-subjectacquisitions could better elucidate this issue.

A scanner with higher parallelization (greater than16:1) may also address these frame rate considerations.A more modern scanning system could also improveimage quality and increase spatial sampling of aberrators,as somemodern 3-D scanners have as many as 9000 activeelements, whereas the scanner used in this study had only

1), aberration correction enables visualization of the prox-els could not be visualized before correction. Correctionents C2–C3) portion of the internal carotid artery (ICA).

4 October 2013 � 2:14 am � ce

847848849850851852853854855856857858859860861862863864865866867868

Q3

Q4

print&web4C=FPO

Fig. 7. Visualization of the left internal carotid artery (ICA) is greatly improved in the aberration-corrected case (b) rela-tive to the aberrated case (a) in an anterior-posterior view in single subject (subject 3). In oblique views of the same data,the vessel morphology is visible in the corrected rendering (d). The S-shaped carotid siphon is particularly visible on the

left side.

8 Ultrasound in Medicine and Biology Volume -, Number -, 2013

869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914915916917918919920921922923924925926927928929930

931932933934935936937938939940941942943944945946947948949950951952953954955956957958959960961962963964965966967968969970971972973974975976977978979980981982983984985986987988989990991992

256 receive channels. Although access to channel datawould still be required—a difficult proposition givenpartial beamforming in probe handles (Cochrane et al.2010; Savord and Solomon 2003)—such a scanner wouldalso allow computation to be performed online and muchmore quickly.

In examining the aberration characteristics illus-trated in Figure 9, we contribute to the small number ofexisting in vivo measurements of aberrator characteris-tics (Ivancevich et al. 2009; Lindsey and Smith 2013;Vignon et al. 2010). The RMS aberrator amplitudespresented here are within 35 to 60 ns, in closeagreement with previously published values. The lowmean FWHM correlation length for subject 4, thesubject exhibiting window failure, may be explained bythe fact that in the lower-SNR case associated witha poor acoustic window, uncorrelated noise produceda rapid decline in the autocorrelation of the aberrationmap. The relatively large error bars in Figure 9(cand d) suggest that the measurement of five differentpropagation paths through the skull yields measurements

FLA 5.2.0 DTD � UMB9643_proof �

of aberrators with different spatial frequency content,further suggesting that unique aberrators (i.e., differentisoplanatic patches) are being measured, which is thegoal in this technique.

In comparing Table 1 and Figure 9, RMS aberrationamplitude is consistent across all five subjects, so there isnot a strong relationship between the strength of the aber-rator and the improvement resulting from correction inthe patients scanned. However, aberration correctionyielded the smallest improvement in subject 5, the subjectexhibiting the smallest RMS delay and the greatestdiscrepancy in FWHM correlation length between thetwo sides of the head. The presence of a high-spatial-frequency aberrator adjacent to the left transducer mayhave been difficult to overcome, especially given thatthe correction beacon for this side had to pass throughthe widely varying FWHM correlation lengths measuredon the right side in this subject.

Another important effect of ultrasound propagationthrough the skull is amplitude distortion. Althoughprevious studies have reported the benefit of correcting

4 October 2013 � 2:14 am � ce

Q5

print&

web4C=FPO

Fig. 8. Offline 3-D renderings in a single subject (subject 5). (a) Aberrated and (b) corrected renderings show similarvascular anatomy in transverse views, enabling visualization of the circle of Willis. Rotating the same data sets producesthe oblique views (c and d). A registration error is visible in the anterior cerebral artery (ACA) in (c). Additionally, thecorrected rendering in (d) indicates improved visualization of the right posterior cerebral artery (PCA), right middle cere-

bral artery (MCA) and left MCA relative to (c).

3-D transcranial US with aberration correction d B. D. LINDSEY et al. 9

993994995996997998999

1000100110021003100410051006100710081009101010111012101310141015101610171018101910201021102210231024102510261027102810291030103110321033103410351036103710381039104010411042104310441045104610471048104910501051105210531054

10551056105710581059106010611062106310641065106610671068106910701071107210731074107510761077107810791080108110821083108410851086108710881089109010911092109310941095109610971098109911001101110211031104110511061107110811091110111111121113111411151116

both amplitude and phase as a function of frequency(Aubry et al. 2001; Clement and Hynynen 2002; Whiteet al. 2005; Wu and Fink 1991), we do not havea means of implementing amplitude correction becausethe scanning system used cannot transmit arbitrarywaveforms on a per channel basis.

Aberration correction increased the ability to detectblood flow in the major cerebral arteries in 60% of casesafter correction, compared with 46% of cases withoutcorrection. Given that color flow imaging is an inherentlylow-SNR technique which is further limited by the skulland 3-D ultrasound as discussed in the Introduction,this improvement is helpful toward achieving the endgoal of developing a 3-D ultrasound imaging systemsensitive enough to distinguish normal and abnormalflow in all individuals. The results in this study suggestthat the most clinically helpful effect of this aberrationcorrection technique lies in its ability to fill in vesselswith missing segments (Figs. 5–7), though some

FLA 5.2.0 DTD � UMB9643_proof �

dropout remains. Although the presence of this dropoutof the Doppler signal in some of the presented imagesindicates that this goal has not yet been achieved,the aberration correction improvement presented here(�2–3 dB) might be combined with our earlier effortsat improving sensitivity by reducing cable lengths(�7 dB) (Lindsey et al. 2011) and reducing transducerfrequency (�3–8 dB, patient dependent) (Lindsey et al.2013).

It should also be noted that the primary advantageof this aberration correction technique is that multipleisoplanatic patches are corrected, signifying structuresare not lost at the edges of the volume as in other adaptiveimaging techniques. This effect was observed in thisstudy. For example, in Figure 8, the circle of Willis,despite its large spatial extent, remains entirely visiblewith aberration correction.

Although all data presented in this article have beenDoppler magnitude (i.e., power Doppler) as we examined

4 October 2013 � 2:14 am � ce

print&web4C=FPO

Fig. 9. Measured root-mean-square (RMS) bilateral aberration strengths for each subject (a and b) featuring the mean andstandard deviation of five isoplanatic patch measurements. A larger standard deviation indicates a larger variation in aber-rator strength among different isoplanatic patches within a single subject. Measured full width half-maximum (FWHM)correlation lengths are also provided for each subject (c and d). A larger standard deviation indicates greater variation in

the maximum spatial frequency content among different isoplanatic patches within a single subject.

10 Ultrasound in Medicine and Biology Volume -, Number -, 2013

11171118111911201121112211231124112511261127112811291130113111321133113411351136113711381139114011411142114311441145114611471148114911501151115211531154115511561157115811591160116111621163116411651166116711681169117011711172117311741175117611771178

11791180118111821183118411851186118711881189119011911192119311941195119611971198119912001201120212031204120512061207120812091210121112121213121412151216121712181219122012211222122312241225122612271228122912301231123212331234123512361237123812391240

the question of sensitivity, quantitative measurement ofcerebral blood flow velocities is most clinically useful.This measurement is most commonly performed usingpulsed spectral Doppler, to which this aberration correc-tion technique is directly applicable and should producesimilar improvements. In general, the primary challengewith performing a 3-D spectral Doppler exam lies in posi-tioning a small Doppler gate within a large 3-D volumeusing either only B-mode anatomy or a frame rate-reducing triplex mode.

If a non-invasive, real-time 3-D ultrasound-basedtechnology could, through imaging and blood flowvelocity measurements, both confirm the presence of anocclusion and rule out hemorrhage, physicians wouldpossess all the information required to confidently forma treatment plan, including making an informed tPA deci-sion. Given the time-sensitive nature of stroke and thecorresponding urgency of this decision, the low costand high portability of ultrasound systems allow themto potentially be used in the field or ambulance,decreasing time to tPA decision (Bowman 2010;

FLA 5.2.0 DTD � UMB9643_proof �

Holscher et al. 2008; Schlachetzki et al. 2012).Although either a color Doppler approach, such asdescribed here, or a perfusion imaging approach, asdiscussed in the Introduction, may detect the presenceof a clot, recent progress at detecting hemorrhage withultrasound is encouraging for solving the second andequally important part of this problem (Kern et al.2008; Kukulska-Pawluczuk et al. 2012; Meyer-Wietheet al. 2009).

In summary, we have assessed the feasibility ofa phase aberration correction technique for bilateral 3-Dtranscranial ultrasound in a study of healthy volunteersusing microbubble contrast agent. Aberration correctionproduced an increase in Doppler voxels of 33.9%and enabled detection of blood flow in 60% of arteriesafter correction, compared with 46% of arteries beforecorrection. We hope that when combined with recentdevelopments in portable ultrasound, microbubble-basedimaging techniques and transducers, 3-D ultrasoundmight improve in quality and gain clinical acceptancefor use in stroke diagnosis.

4 October 2013 � 2:14 am � ce

Q6

Q7

3-D transcranial US with aberration correction d B. D. LINDSEY et al. 11

12411242124312441245124612471248124912501251125212531254125512561257125812591260126112621263126412651266126712681269127012711272127312741275127612771278127912801281128212831284128512861287128812891290129112921293129412951296129712981299130013011302

1303130413051306

Acknowledgments—The authors thank Vivek Patel, Albert Chen andMarko Jakovljevic for assistance moving the scanner, and SankalpGokhale for administering IVs This research was supported by GrantsR01 HL089507 and T32 EB001040 from the National Institutes ofHealth.

Q8

1307130813091310131113121313131413151316131713181319132013211322132313241325132613271328132913301331133213331334133513361337133813391340134113421343134413451346134713481349135013511352135313541355135613571358135913601361136213631364

REFERENCES

Aaslid R, Markwalder TM, Nornes H. Noninvasive transcranial Dopplerultrasound recording of flow velocity in basal cerebral arteries.J Neurosurg 1982;57:769–774.

Alexandrov AV, Wojner AW, Grotta JC. CLOTBUST: Design ofa randomized trial of ultrasound-enhanced thrombolysis for acuteischemic stroke. J Neuroimaging 2004;14:108–112.

American Institute of Ultrasound in Medicine (AIUM)/National Electri-cal Manufacturers Association (NEMA). Standard for real-timedisplay of thermal and mechanical acoustic indices on diagnosticultrasound equipment. Laurel, MD: AIUM; 1998.

Andersen G, Munch J, Veitch P. Compact, holographic correction ofaberrated telescopes. Appl Opt 1997;36:1427–1432.

Astheimer JP, Pilkington WC, Waag RC. Reduction of variance in spec-tral estimates for correction of ultrasonic aberration. IEEE Trans Ul-trason Ferroelectr Freq Control 2006;53:79–89.

Aubry JF, Tanter M, Gerber J, Thomas JL, Fink M. Optimal focusing byspatio-temporal inverse filter: II. Experiments. Application tofocusing through absorbing and reverberating media. J Acoust SocAm 2001;110:48–58.

Azucena O, Crest J, Cao J, Sullivan W, Kner P, Gavel D, Dillon D,Olivier S, Kubby J. Wavefront aberration measurements and correc-tions through thick tissue using fluorescent microsphere referencebeacons. Opt Express 2010;18:17521–17522.

Baumgartner RW, Arnold M, Gonner F, Staikow I, Herrmann C,Rivoir A, Muri RM. Contrast-enhanced transcranial color-codedduplex sonography in ischemic cerebrovascular disease. Stroke1997;28:2473–2478.

Becker A, Kuhnt D, Bakowsky U, Nimsky C. Contrast-enhanced ultra-sound ventriculography. Neurosurgery 2012;71:ons296–ons301;discussion ons301.

Beckers JM. Increasing the size of the isoplanatic patch with multicon-jugate adaptive optics. In: Proceedings, ESO Symposium on LargeTelescopes and Their Instrumentation. European Southern Observa-tory, Garching, Germany 1988;693–703.

Bedggood PA, Ashman R, Smith G,Metha AB.Multiconjugate adaptiveoptics applied to an anatomically accurate human eye model. OptExpress 2006;14:8019–8030.

Bogousslavsky J, Van Melle G, Regli F. The Lausanne Stroke Registry:Analysis of 1,000 consecutive patients with first stroke. Stroke 1988;19:1083–1092.

Bowman J. Ultrasound Applications in EMS. J Emerg Med Serv 2010;35:36–47.

Chenaitia H, Squarcioni C, Marie BP, Emgan Q, Tomislav P. Transcra-nial sonography in prehospital setting. Am J Emerg Med 2011;29:1231–1233.

Clement GT, Hynynen K. A non-invasive method for focusing ultra-sound through the human skull. PhysMed Biol 2002;47:1219–1236.

Clement GT, White PJ, Hynynen K. Enhanced ultrasound transmissionthrough the human skull using shear mode conversion. J Acoust SocAm 2004;115:1356–1364.

Dahl JJ, Feehan TJ. Direction of arrival filters for improved aberrationestimation. Ultrason Imaging 2008;30:1–20.

Dahl JJ, SooMS, Trahey GE. Spatial and temporal aberrator stability forreal-time adaptive imaging. IEEE Trans Ultrason Ferroelectr FreqControl 2005;52:1504–1517.

Davis SM, Donnan GA. 4.5 hours: The new timewindow for tissue plas-minogen activator in stroke. Stroke 2009;40:2266–2267.

Dubinin A, Cherezova T, Belyakov A, Kudryashov A. Human retinaimaging: Widening of high resolution area. J Mod Opt 2008;55:671–681.

Flax S, O’Donnell M. Phase-aberration correction using signals frompoint reflectors and diffuse scatterers: Basic principles. IEEE TransUltrason Ferroelect Freq Control 1988;35:758–767.

FLA 5.2.0 DTD � UMB9643_proof �

Fry FJ, Barger JE. Acoustic properties of the human skull. J Acoust SocAm 1978;63:1576–1590.

Gahn G, Gerber J, Hallmeyer S, Hahn G, Ackerman RH, Reichmann H,von Kummer R. Contrast-enhanced transcranial color-coded du-plexsonography in stroke patients with limited bone windows.AJNR Am J Neuroradiol 2000;21:509–514.

Gauss RC, Trahey GE, Soo MS. Wavefront estimation in the humanbreast. In: Proc SPIE 4325, Medical imaging 2001: Ultrasonicimaging and signal processing, 172 (May 30, 2001). doi: 10.1117/12.428194; http://dx.doi.org/10.1117/12.428194.

Hacke W, Kaste M, Bluhmki E, Brozman M, Davalos A, Guidetti D,Larrue V, Lees KR, Medeghri Z, Machnig T, Schneider D, vonKummer R, Wahlgren N, Toni D. Thrombolysis with alteplase 3 to4.5 hours after acute ischemic stroke. N Engl J Med 2008;359:1317–1329.

Hacke W, Kaste M, Fieschi C, Toni D, Lesaffre E, von Kummer R,Boysen G, Bluhmki E, Hoxter G, MahagneMH, Hennerici M. Intra-venous thrombolysis with recombinant tissue plasminogen activatorfor acute hemispheric stroke: The European Cooperative AcuteStroke Study (ECASS). JAMA 1995;274:1017–1025.

Hashimoto H, Etani H, NakaM, Kinoshita N, Nukada T. [Assessment ofthe rate of successful transcranial Doppler recording through thetemporal windows in Japanese with special reference to aging andsex]. Nippon Ronen Igakkai Zasshi 1992;29:119–122.

Haworth KJ, Fowlkes JB, Carson PL, Kripfgans OD. Towards aberrationcorrection of transcranial ultrasound using acoustic droplet vapor-ization. Ultrasound Med Biol 2008;34:435–445.

Heidenreich PA, Trogdon JG, Khavjou OA, Butler J, Dracup K,Ezekowitz MD, Finkelstein EA, Hong Y, Johnston SC, Khera A,Lloyd-Jones DM, Nelson SA, Nichol G, Orenstein D, Wilson PW,Woo YJ. Forecasting the future of cardiovascular disease in theUnited States: A policy statement from the American Heart Associ-ation. Circulation 2011;123:933–944.

Holscher T. Prehospital use of portable ultrasound for stroke diagnosisand treatment initiation. Air Rescue 2012;2:48–50.

Holscher T, Schlachetzki F, Zimmermann M, Jakob W, Ittner KP,Haslberger J, Bogdahn U, Boy S. Transcranial ultrasound from diag-nosis to early stroke treatment: 1. Feasibility of prehospital cerebro-vascular assessment. Cerebrovasc Dis 2008;26:659–663.

Holscher T, Wilkening W, Draganski B, Meves SH, Eyding J, Voit H,Bogdahn U, Przuntek H, Postert T. Transcranial ultrasound brainperfusion assessment with a contrast agent-specific imaging mode:Results of a two-center trial. Stroke 2005;36:2283–2285.

Hoyer HX, Vogl S, Schiemann U, Haug A, Stolpe E, Michalski T. Pre-hospital ultrasound in emergency medicine: Incidence, feasibility,indications and diagnoses. Eur J Emerg Med 2010;17:254–259.

Ivancevich NM, Dahl JJ, Smith SW. Comparison of 3-Dmulti-lag cross-correlation and speckle brightness aberration correction algorithmson static and moving targets. IEEE Trans Ultrason FerroelectrFreq Control 2009;56:2157–2166.

Ivancevich NM, Dahl JJ, Trahey GE, Smith SW. Phase-aberrationcorrection with a 3-D ultrasound scanner: Feasibility study. IEEETrans Ultrason Ferroelectr Freq Control 2006;53:1432–1439.

Ivancevich NM, Pinton GF, Nicoletto HA, Bennett E, Laskowitz DT,Smith SW. Real-time 3-D contrast-enhanced transcranial ultrasoundand aberration correction. Ultrasound Med Biol 2008;34:1387–1395.

Jensen JA. Estimation of blood velocities using ultrasound. Cambridge:Cambridge University Press; 1996.

Johnston DC, Welsh BM. Analysis of multiconjugate adaptive optics.J Opt Soc Am 1994;11:394–408.

Kern R, Diels A, Pettenpohl J, Kablau M, Brade J, Hennerici MG,Meairs S. Real-time ultrasound brain perfusion imaging with anal-ysis of microbubble replenishment in acute MCA stroke. J CerebBlood Flow Metab 2011;31:1716–1724.

Kern R, Kablau M, Sallustio F, Fatar M, Stroick M, Hennerici MG,Meairs S. Improved detection of intracerebral hemorrhage withtranscranial ultrasound perfusion imaging. Cerebrovasc Dis 2008;26:277–283.

Krejza J, Swiat M, Pawlak MA, Oszkinis G, Weigele J, Hurst RW,Kasner S. Suitability of temporal bone acoustic window:

4 October 2013 � 2:14 am � ce

Q9

12 Ultrasound in Medicine and Biology Volume -, Number -, 2013

13651366136713681369137013711372137313741375137613771378137913801381138213831384138513861387138813891390139113921393139413951396139713981399140014011402140314041405140614071408140914101411141214131414141514161417141814191420142114221423142414251426

1427142814291430143114321433143414351436143714381439144014411442144314441445144614471448144914501451145214531454145514561457145814591460146114621463146414651466146714681469147014711472147314741475147614771478147914801481148214831484

Conventional TCD versus transcranial color-coded duplex sonog-raphy. J Neuroimaging 2007;17:311–314.

Kukulska-Pawluczuk B, Ksiazkiewicz B, Nowaczewska M. Imaging ofspontaneous intracerebral hemorrhages by means of transcranialcolor-coded sonography. Eur J Radiol 2012;81:1253–1258.

Lindsey BD, Light ED, Nicoletto HA, Bennett ER, Laskowitz DT,Smith SW. The ultrasound brain helmet: New transducers andvolume registration for in vivo simultaneous multi-transducer 3-Dtranscranial imaging. IEEE Trans Ultrason Ferroelectr Freq Control2011;58:1189–1202.

Lindsey BD, Nicoletto HA, Bennett E, Laskowitz DT, Smith SW. Simul-taneous bilateral real-time 3-D transcranial ultrasound imaging at 1MHz through poor acoustic windows. Ultrasound Med Biol 2013;39:721–734.

Lindsey BD, Smith SW. Pitch-catch phase aberration correction ofmultiple isoplanatic patches for 3-D transcranial ultrasoundimaging. IEEE Trans Ultrason Ferroelectr Freq Control 2013;60:463–480.

Liu DL, Waag RC. Time-shift compensation of ultrasonic pulse focusdegradation using least-mean-square error estimates of arrivaltime. J Acoust Soc Am 1994;95:542–555.

Maciak A, Kier C, Seidel G, Meyer-Wiethe K, Hofmann UG. Detectingstripe artifacts in ultrasound images. J Digit Imaging 2009;22:548–557.

Marchetti E, Brast R, Delabre B, Donaldson R, Fedrigo E, Frank C,Hubin N. On-sky testing of the multi-conjugate adaptive opticsdemonstrator. Messenger 2007;129:8–13.

Marinoni M, Ginanneschi A, Forleo P, Amaducci L. Technical limits intranscranial Doppler recording: Inadequate acoustic windows.Ultrasound Med Biol 1997;23:1275–1277.

Meyer-Wiethe K, Sallustio F, Kern R. Diagnosis of intracerebralhemorrhage with transcranial ultrasound. Cerebrovasc Dis 2009;27(Suppl. 2):40–47.

Miller-Jones SM. Automated arrival time correction for ultrasoniccephalic imaging. Ph.D. dissertation, Durham, NC: Department ofBiomedical Engineering, Duke University, 1980.

Muller RA, Buffington A. Real-time correction of atmosphericallydegraded telescope images through image sharpening. J Opt SocAm 1974;64:1200–1209.

Nedelmann M, Stolz E, Gerriets T, Baumgartner RW, Malferrari G,Seidel G, Kaps M. Consensus recommendations for transcranialcolor-coded duplex sonography for the assessment of intracranialarteries in clinical trials on acute stroke. Stroke 2009;40:3238–3244.

Pichardo S, Sin VW, Hynynen K. Multi-frequency characterization ofthe speed of sound and attenuation coefficient for longitudinal trans-mission of freshly excised human skulls. Phys Med Biol 2011;56:219–250.

Pinton G, Aubry JF, Fink M, Tanter M. Effects of nonlinear ultrasoundpropagation on high intensity brain therapy. Med Phys 2011;38:1207–1216.

Postert T, Federlein J, Przuntek H, Buttner T. Insufficient and absentacoustic temporal bone window: Potential and limitations of trans-cranial contrast-enhanced color-coded sonography and contrast-enhanced power-based sonography. Ultrasound Med Biol 1997;23:857–862.

Roger VL, Go AS, Lloyd-Jones DM, Benjamin EJ, Berry JD,Borden WB, Bravata DM, Dai S, Ford ES, Fox CS, Fullerton HJ,Gillespie C, Hailpern SM, Heit JA, Howard VJ, Kissela BM,Kittner SJ, Lackland DT, Lichtman JH, Lisabeth LD, Makuc DM,Marcus GM, Marelli A, Matchar DB, Moy CS, Mozaffarian D,Mussolino ME, Nichol G, Paynter NP, Soliman EZ, Sorlie PD,Sotoodehnia N, Turan TN, Virani SS, Wong ND, Woo D,Turner MB. Heart disease and stroke statistics—2012 Update: Areport from the American Heart Association. Circulation 2011;125:e2–e220.

Schlachetzki F, Herzberg M, Holscher T, Ertl M, Zimmermann M,Ittner KP, Pels H, Bogdahn U, Boy S. Transcranial ultrasound

FLA 5.2.0 DTD � UMB9643_proof �

1485

from diagnosis to early stroke treatment: Part 2. Prehospital neuro-sonography in patients with acute stroke: The Regensburg StrokeMobile Project. Cerebrovasc Dis 2012;33:262–271.

Schlachetzki F, Holscher T, Koch HJ, Draganski B, May A, Schuierer G,Bogdahn U. Observation on the integrity of the blood-brain barrierafter microbubble destruction by diagnostic transcranial color-coded sonography. J Ultrasound Med 2002;21:419–429.

Seidel G, Algermissen C, Christoph A, Claassen L,Vidal-Langwasser M, Katzer T. Harmonic imaging of the humanbrain: Visualization of brain perfusion with ultrasound. Stroke2000;31:151–154.

Seidel G, Kaps M, Gerriets T. Potential and limitations of transcranialcolor-coded sonography in stroke patients. Stroke 1995;26:2061–2066.

Sims JR, Rordorf G, Smith EE, Koroshetz WJ, Lev MH, Buonanno F,Schwamm LH. Arterial occlusion revealed by CT angiographypredicts NIH stroke score and acute outcomes after IV tPA treat-ment. AJNR Am J Neuroradiol 2005;26:246–251.

Smith SW, Ivancevich NM, Lindsey BD, Whitman JJ, Light ED,Fronheiser MP, Nicoletto HA, Laskowitz DT. The ultrasound brainhelmet: Feasibility study of multiple simultaneous 3-D scans ofcerebral vasculature. Ultrasound Med Biol 2009;35:329–338.

Smith SW, Pavy HG, von Ramm OT. High speed ultrasound volumetricimaging system: Part I. Transducer design and beam steering. IEEETrans Ultrason Ferroelectr Freq Control 1991;38:100–108.

Smith SW, Trahey GE, von RammOT. Phased array ultrasound imagingthrough planar tissue layers. Ultrasound Med Biol 1986;12:229–243.

Steinberg BD. Radar imaging from a distorted array: The radio cameraalgorithm and experimenls. IEEE Trans Antennas Propagat 1981;29:740–748.

TanerMT, Koehler F. Alhilali. Estimation and correction of near-surfacetime anomalies. Geophys 1974;39:442–463.

Tanter M, Aubry JF, Gerber J, Thomas JL, Fink M. Optimal focusing byspatio-temporal inverse filter: I. Basic principles. J Acoust Soc Am2001;110:37–47.

Tillett JC, Astheimer JP, Waag RC. A model of distributed phase aber-ration for deblurring phase estimated from scattering. IEEE TransUltrason Ferroelectr Freq Control 2010;57:214–228.

Vignon F, Aubry JF, Tanter M, Margoum A, Fink M. Adaptive focusingfor transcranial ultrasound imaging using dual arrays. J Acoust SocAm 2006;120:2737–2745.

Vignon F, Shi WT, Yin X, Hoelscher T, Powers JE. The stripe artifact intranscranial ultrasound imaging. J Ultrasound Med 2010;29:1779–1786.

Von Ramm OT, Smith SW, Pavy HG. High speed ultrasound volumetricimaging system part II: Parallel processing and display. IEEE TransUltrason Ferroelectr Freq Control 1991;38:109–115.

Waag RC, Astheimer JP. Statistical estimation of ultrasonic propagationpath parameters for aberration correction. IEEE Trans Ultrason Fer-roelectr Freq Control 2005;52:851–869.

Wessels T, Bozzato A, Mull M, Klotzsch C. Intracranial collateral path-ways assessed by contrast-enhanced three-dimensional transcranialcolor-coded sonography. Ultrasound Med Biol 2004;30:1435–1440.

White DN, Stevenson RJ. The acoustic characteristics of the skull.Ultrasound Med Biol 1978;4:225–252.

White J, Clement GT, Hynynen K. Transcranial ultrasound focus recon-struction with phase and amplitude correction. IEEE Trans UltrasonFerroelectr Freq Control 2005;52:1518–1522.

White PJ, Clement GT, Hynynen K. Longitudinal and shear mode ultra-sound propagation in human skull bone. Ultrasound Med Biol 2006;32:1085–1096.

Wilson MH, Levett DZ, Dhillon S, Mitchell K, Morgan J, M.P.W. G,Imray C. Stroke at high altitude diagnosed in the field using portableultrasound. Wilderness Environ Med 2011;22:54–57.

World Health Organization. The top 10 causes of death. Geneva:Author; 2011. Fact Sheet No. 310.

4 October 2013 � 2:14 am � ce

148614871488

Our reference: UMB 9643 P-authorquery-v9

AUTHOR QUERY FORM

Journal: UMB

Article Number: 9643

Please e-mail or fax your responses and any corrections to:

E-mail: L.Bernazzani@elsevier.com

Fax: 570-409-4520

Dear Author,

Please check your proof carefully and mark all corrections at the appropriate place in the proof (e.g., by using on-screen

annotation in the PDF file) or compile them in a separate list. Note: if you opt to annotate the file with software other than

Adobe Reader then please also highlight the appropriate place in the PDF file. To ensure fast publication of your paper please

return your corrections within 48 hours.

For correction or revision of any artwork, please consult http://www.elsevier.com/artworkinstructions.

Any queries or remarks that have arisen during the processing of your manuscript are listed below and highlighted by flags in

the proof.

Location

in articleQuery / Remark: Click on the Q link to find the query’s location in text

Please insert your reply or correction at the corresponding line in the proof

If there are any drug dosages in your article, please verify them and indicate that you have done so by

initialing this query

Q1 Is short title OK?

Q2 Did you mean Lindsey et al. (2011) here? there is no Lindsey et al. (2009).

Q3 Please add Cochrane et al. 2010 and Savord and Solomon 2003 to reference list.

Q4 Do you mean Vignon et al. 20010 or 2006 here? there is no Vignon et al. 2008

Q5 Please add Wu and Fink 1991 to Reference list or delete citation.

Q6 Please confirm that the Beckers reference added to the list is the one you cited in text.

Q7 Please confirm that Dubinin reference added to list is the one you cited in text.

Q8 Please confirm that the Gauss reference added is the one you cited in text.

Q9 Please confirm that the Miller-Jones reference added to the list is the one you cited in text.

Q10 Please confirm that given names and surnames have been identified correctly.

Please check this box or indicate

your approval if you have no

corrections to make to the PDF file ,

Thank you for your assistance.