DRDC-RDDC-2014-P113

Direct comparison of the primary blast response of a physical head model with post-mortem human subjects

Simon Ouellet1, Cynthia Bir2 and Amal Bouamoul1

1Defence Research and Development Canada-Valcartier Research Center, 2459 Bravoure road, Québec, QC, Canada, simon.ouellet@drdc-rddc.gc.ca

2Keck School of Medicine, University of Southern California, Los Angeles, CA, USA ABSTRACT As the gathering of information on the prevalence of blast-induced traumatic brain injuries (bTBI) continues, there is a need for the development and validation of a physical model (headform) reproducing the mechanical response of the human head to the direct loading from a blast wave. The chain of events leading to an injuries following direct exposure to a blast wave is very complex and its full determination is still the topic of several research efforts. The first step in the injury cascade is necessarily the mechanical insult of the blast wave to the human head. With a combination of representative anatomical features, adequate material selection and careful instrumentation, a validated physical model could measure real external pressure field history and predict resulting intra-cranial pressures (ICP) for any blast loading scenario. In addition, a physical model has the unique ability to measure quantitatively the effect of protective headwear. The following article discusses the validation of the BI2PED (Blast-Induced Brain Injury Protection Evaluation Device) response against post-mortem human subjects (PMHS). Previously reported PMHS blast wave generator tests were methodically replicated in the same facility using the BI2PED. Loading conditions, instrumentation type and position as well as the head mounting technique were reproduced to ensure that the only difference between the two series of experiments was the model itself. A direct comparison of measured ICP histories is presented for two loading orientations and three loading magnitudes. It is demonstrated that the physical model response is in good agreement with that of the PMHS response. From signal analysis, additional evidences supporting skull deformation as the main contributor to ICP variations are discussed. Finally, external pressure fields from the blast wave generator experiments are compared to full scale free-field tests. 1. INTRODUCTION There is a preponderance of clinical and experimental evidences that suggest that traumatic brain injury (TBI) can occur as a result of a direct exposure to blast wave [1]. The chain of events leading to such injuries is likely very complex and its full determination is still the topic of several multi-disciplinary research efforts [1]. Highly controlled laboratory experiments on post-mortem human subject as well as the development of physical and numerical head models can certainly provide useful information on the first step in the cascade of events leading to injury, which is considered to be the mechanical insult of the blast wave to the human head. As a thorough understanding of the injury mechanism is developed, the need for physical models capable of reproducing the mechanical response of the human head under blast loading increases. A validated physical model combining representative anatomical features, adequate material selection and careful instrumentation provides three clear benefits. First, it can help to characterize the head external pressure field history (i.e. the loading) for any blast scenario. Operational blast scenarios are infinite and it is only by characterizing the real mechanical input to the head that it will be possible to distinguish between them. Second, a physical model can help estimate the magnitude of the stresses developed in the brain. This is essential to provide a link between external mechanical insult and the potential for injury. Finally, it can help evaluating the performance of protective headwear systems.

The direct exposure of the head to a blast wave creates a very short duration high amplitude loading that is very different in nature from loadings resulting from impacts seen in automotive or sport accidents. Willinger et al. [2,3] have comprehensively discussed how the duration of a loading on the head can determine the nature of the strain and stress fields in the brain. They distinguished between three lesion mechanisms, each of which is particular to a range of loading duration. For long

duration loading above 10-12 ms, distributed lesions throughout the brain are attributed to the generation of intra-cranial stresses from inertial forces. In such regime, the whole head is subjected to the same translational and rotational acceleration field. For duration between 4-10 ms, the skull motion and brain motion are decoupled. This regime has been the subject of numerous studies since the development of the rapid skull motion theory by Viano et al. [4]. Willinger stated that for such loading durations, the first resonance frequency of the head, which he cites as being between 100-150 Hz, may be excited. Skull to brain relative translational and rotational motion can cause bridge vein shearing and contusion if direct contact occurs. Finally, for impact duration below 4 ms, the loading can excite a second resonant frequency observed between 700-800 Hz [3-6]. This is a wave-dominated regime where local skull deformation occurs. In the context of blast TBI research, such mechanism has also been referred to as “skull flexure” or simply “direct stress transmission” [7-10]. Flexure happens when the loading energy is delivered rapidly enough so that the skull does not have time to reach force equilibrium throughout its structure. The skull doesn’t move as a rigid body but rather sustain local deformation, which then propagates in the structure as waves. Local deformations in the skull may generate compressive and tensile intra-cranial stresses and lesions in nearby brain regions.

Operational blast waves may excite the head structure through more than one frequency band and potentially create lesions from a combination of the aforementioned mechanisms. For example, skull deformation may occur along with relative brain-skull motion, at different times but still within a single event. An important question to answer is which of these mechanisms dominates in the generation of brain stresses following blast exposure. Clearly, simple physical models, such as automotive test devices (ATD), which only records head global accelerations, cannot provide the answer to such question. Even though they can be representative of certain shock severity level and therefore certain risk of cerebral lesions, they cannot distinguish between lesion mechanisms [3]. The study of the effect of blast on the brain requires a more direct measurement.

When a high-pressure blast wave travels across a body (head), it reflects off and diffracts around to form a transient pressure field that is unique to each individual blast scenario. This pressure field is influenced by the blast propagation direction, magnitude and duration at the location of interaction, but also by surrounding reflecting surfaces, including the ground. The stresses generated within the brain following blast exposure therefore also depend on all of these characteristics.

The BI2PED headform shown in Figure 1 has been developed by Defence Research and Development Canada (DRDC) Valcartier research center to characterize head external pressure field history following any given blast exposure, to predict resulting global head acceleration and brain intra-cranial pressure (ICP) variations. It is understood that biological material failure in a real scenario may also occur under shear loads, but it is assumed that ICP is representative of the magnitude of injurious stresses generated within the brain. Shear stresses or strain within brain or brain-like materials would be very difficult to measure experimentally. It has been shown that the BI2PED presents the necessary physiological feature and representative selection of material to estimate head response to blast [11]. Nevertheless, a critical aspect of this physical model to fully meet its purpose is the validation of its response against real human head response (post-mortem). Until this is achieved, there is no guarantee the evolution of ICP follows that from a human head.

Bir et al. [10] has recently reported blast wave generator tests where instrumented post-mortem human subjects (PMHS) heads were subjected to blast waves of 3 different intensities and in 4 different orientations. Fluctuation in intra-cranial pressures was monitored at 4 locations within the brain and skull strains were measured at 5 locations. This parametric approach rendered a dataset that is well fitted to begin the validation process of the BI2PED. It is understood that the full injury cascade cannot be assessed using PMHS, but the BI2PED model focuses on reproducing the mechanical response of the head and the stress transmission mechanism, not the injury itself. Through a collaboration agreement, DRDC Valcartier Research Center and Wayne State University (WSU) Biomedical Engineering Department teamed up to replicate the reported PMHS tests using the BI2PED headform. Loading conditions, instrumentation type and position as well as head mounting technique were reproduced to ensure that the only difference between the two series of experiments was the use of BI2PED or PMHS. Direct comparison of measured ICP histories was therefore possible.

2. EX The summskulmemmatebottowateare fside,at vausinmodin fuwithand compossthe h 2.1 I For origiversmonwhento re(0.00linea

FISOmon(Figudiamextrefounreasobend30-4doesinfor

Figure

XPERIMEN

BI2PED consmarized here l, cerebrospin

mbranes are oerial propertieom of the craner, which actsflush mounted, left side, bacarying locatiog very thin w

dified to ensurull scale blast h 6 accelerome

rotational accmparison of intsible; The natuhead orientatio

Instrumentat

the internal ininal PMHS sion of the BI2

nitoring ICP. Tn adequately esolve rise tim001% of full-ar output for p

The originaO fiber-opticnitor ICP. Ture 2) sensor

meter. It has themely small. nd to be veon, FISO dding the fibe40 mm radius not providrmation abou

1: The Blast-I

NTAL METH

struction, matefor the purpos

nal fluid (CSFf uniform thices of their bionial cavity and as the CSF. Td with the skick and top of ons depending

wires which arre that the sens

testing, the heters at the bacelerations. Totra-cranial preure, orientatioon and mount

tion

nstrumentationstudy, namely2PED uses moThey are 1.9 casted in the bme on the ord-scale output ppressures up to

al WSU stuc pressure seThe FISO Fr is 550 mihe advantage However, it

ery fragile. does not recer-optic wire us. The mande detailed ut the transduc

Induced Brain

HOD

erials and instse of this artic

F), membranesckness and ar

ological countd is held in plThe headformn in a customthe head. Pres

g on the purpoe pulled out osing diaphragm

headform is tyase of the necko replicate thessures, threeon and positionting technique

n, focus was y the frontalodified Kulitemm in diamebrain surrogatder of ten miper g) to be ao 14 bars and

udy used ensor to

FOP-MIV icrons in of being was also For that

commend above a

nufacturer technical cer aside

p

n Injury Prote

trumentation acle. The currens and brain stre made from terpart. The brlace by the me

m is instrumenm design sleevssure transducose of the triaof the gel durim is in intima

ypically mounk which allowe original PMsets of param

n of the instrue.

put on replicalobe, parietal

e piezo-resistiveter and uni-dte. The transducroseconds. T

adequate for ta resonant fre

Figure 2: piezo-resistive

ection Evaluat

are detailed innt version of ttructures. Thedifferent pol

rain is made embrane strucnted with five ve. They are pcers are also cal. These transing curing. Thate contact witnted on a Hybws for the calcMHS study frometers needed umentation, th

ating three of l lobe and ocve XCL-072 pdirectionnal. Tucers have a s

They are suffithe headform equency over 5

Relative size e XCL-072 se

tion Device (B

n Ouellet et althe headform re skin, skull, fymers selecteof a silicone

ctures. It is sligflat pressure

positioned on tcasted within tsducers are cahese intracranth the gel as it

brid III neck. Iulation of resu

om WSU and to be replica

he loading con

f the ICP meaccipital lobe pressure sensoThey proved tsufficient banficiently accel

operational ra550 kHz.

of pressure se

ensors and b. F

BI2PED).

l. [11] and arereplicates the flax and tentoed to match egel. It rests oghtly suspendtransducers wthe forehead, the brain surroarefully positi

nial transducert cures. WhenIt is then equiultant translatallow for a d

ted as precisenditions and fi

surement fromICPs. The cuors (Figure 2ato be very reldwidth (> 25 leration insensange. They ha

ensors a. KuliFISO FOP sen

e only skin,

orium lastic n the

ded in which

right ogate ioned rs are

n used ipped tional direct ely as inally

m the urrent a) for liable kHz) sitive ave a

ite nsor.

from1500whiccapaeffecbothwouif a wereand

the FcentThe reachfor tcoulopticInevassuincida solongof mcomrelatwith

fromoutenot imagbut fthe wand

m its operatin0 kPa. The sench has a maxacity of the Fcts) and that f

h the piezo-eluld provide va

sensor was toe aligned and processed sig

Replicating FISO pressureer of the headdownside to

hing into the the possibilityld perform sevc wires was n

vitably, the reumption was thdence with regolid gives a mgitudinal stressmagnitude hig

mponent of strtively orientath the stress dir

The positionm the nasion ber skull surfacperfectly matging. Figure 3follow the samwire was achifinal position

ng range. Intensor needs to ximum samplISO sensors t

fragility of theectric and fiblidation of pao break durinboth sensors

nals from both

the original oe sensors throd. This was ob

this method frontal lobe.

y to run the wiveral repetitioot exceeded, iesulting orienhat the effect gards to the blmeasure of sts. Nevertheles

gher than theirress likely dotion independrection is such

n of sensors inbone towards ce. These distatch the PHMS3 shows the x-me route as theved by exitin.

Figure 3* Fiber op

Table 1, Sen

ernal DRDC be used in co

ling rate of 2to monitor the sensors had ber optic sensst and currentg the test seriwere attache

h sensors wer

orientation of tough the skull,bviously the lis that it expIn a physical ires together,

ons in the test it was decidedntation is diffof a different last directiontress in the dss, brain tissuer shear moduominates the dent. The onlyh that the trans

n the original the back of thances were keS head dimen-ray photograp

he piezo-resisting all wires fr

3: BI2PED inteptic sensors a

nsor positions

calibration sonjunction wi250 kHz. Givehe ICP variatiobeen a problesors side by t FISO sensories. Great car

ed to each othre compared.

the sensors w, perpendiculaless intrusive poses the sensmodel, castinin the oppositseries and to

d to run the wfferent than in

orientation mwas maintain

direction of mes as well as slus, with a Postress state a

y effect that csducer itself im

PMHS study he head, distanept similar evnsions. Final ph of the BI2Pive sensor wirrom the back

ernal pressure

are not appare

for both PMH

showed a lineith FISO data en the lack oons adequatel

em in the origiside in the B

r measurementre was taken ther at the base

as challengingar to the skullway to positisor wires signng the sensorste direction frmake sure thires out togethn the origina

may be minimaed. A unidirec

measurement. silicone gel haoisson’s ratio

and the measucould be seen mpedes the me

was specifiednce from the

ven though thsensor positio

PED. The FISres. It can be nof the skull. T

e transducers pent under x-ray

HS and BI2PE

ear output foacquisition sy

of technical inly (e.g. bandwinal study, it w

BI2PED headfts and provideto ensure thate. Following t

g. The original surface tangeion the ICP senificantly, pars into the surrrom the blast. hat the bendingher toward theal PMHS stual if a certain ctional pressuIt is technic

ave a bulk mod approaching ure from the is when the a

easurement.

d in the three ahead mid-plane BI2PED lenons were vali

SO sensors arenoticed that mTable 1 lists a

positions, y imaging.

ED test series.

or pressures uystem (Velocnformation onwidth, accelerwas decided tform. This mee back-up readt sensing elemthe test series

al PHMS student and towarensors in a PMrticularly the

rogate brain alTo ensure tha

g limit of the e back of the h

udy. Howeverrange of angl

ure sensor castally a measudulus that is o0.5. The presensor shoul

angle of incid

axis using distne and depth

ngth and widthidated using xe invisible to

minimal bendinll sensors inte

up to e 50) n the ration o use ethod dings ments , raw

dy ran rd the MHS.

ones llows at we fiber head. r, the les of ted in

ure of orders essure ld be dence

tance from h did x-ray x-ray ng of

ended

frontrepresign

studyfrequprevlow- 2.2 T The for tsectiusedgas increabou

studyusedincid140 was threeoverduringauggenetypicduramilliexpagenebelothat repreblastcondhistochar

propdirecpres

Unfortunatetal sensor waesentative of als.

Post-process

y. All ICP suency of 7 kH

vious free-field-pass Butterw

Test conditio

blast wave gethe BI2PED vaion and a drivd to fill the drexpands and deased to increut the blast wa

Similarly toy, three mem

d to generatdent overpreskPa respection the order

e shock magrpressure histng calibrationge at the taerated from cally have ation in theiseconds. Wansion sectionerator is ablow the 10 ms m

the geneesentative of t profile fducted at DRories for the 8rge detonated

The BI2PEDpagation direcction (i.e. righsure levels.

ely, due to acas significantlthe PMHS se

sing of the ICsignals were Hz. External pd experimentsorth filter with

ns

enerator facilialidation test ven expansionriver section udrives a shockease the ruptuave generator

Figure 4

o the originambrane thickne

te shocks wssures of 85,ively. Shock of 7.5 to 8.5 m

gnitudes. Fulltories were n shots usingarget locatio

classic sholong positiv

e order of With its cun, the current be to generatmark. In ordererated shockan operationa

from free-fieRDC-Valcartie85 kPa shock 0.2 m off the

D was evaluatction (i.e. fronht side exposu

ccidental pullily displaced censor position

CP signals wasfiltered usin

pressure signas with the heah a 10 kHz cu

ity of WSU useries. The sh

n section sepaup to a pressuk wave downure pressure adesign can be

4: Wayne Sta

al PHMS esses were with peak 120 and

duration ms for the l incident measured

g a pencil on. Shock ock tubes ve phase

tens of stom-built blast wave te shocks r to ensure ks were al blast, a eld tests

er research cefrom the blasground and at

ted in two orntal exposure)

ure). Three rep

ing of the wcompared to tns and was pr

s done accordig an 8-pole als were postadform [11], wut-off frequenc

sed in the orighock wave genarated by a myure causing thn the expansioand conseque

e found in [12]

te University

enter was uset wave generat a 5 m distanc

rientations; w) and with th

peats were obt

Figure 5: Cofrom shock

ires during pthe PHMS. Treferred for th

ing to the metlow-pass Bu

-processed fowhich consistecy.

ginal PMHS snerator is a twylar membran

he membraneon section. Thently the shoc].

shock wave g

ed for compaator and for ace.

with the head he head long-atained for each

omparison of k wave generat

5m (0

pouring of theThe occipital she comparativ

thod used in tutterworth filtollowing the sed of a filterin

study (Figure wo part flared ne. Compresseto rupture. U

he thickness ock wave press

generator.

arison. Figure a free-field bla

long-axis paraxis perpendih orientation a

f overpressure tor and from 5

0.2m HOB).

e silicone gelsensor is the ve analysis of

the original PMter with a cusame method ng using an 8

4) was used atube with a d

ed helium gasUpon rupturing

f the mylar casure. More d

5 shows preast from a 5 k

rallel to the scular to the sat all three inc

history gener5kg C4 charge

l, the most

f ICP

MHS ut-off as in -pole

again driver s was g, the an be etails

essure kg C4

shock shock cident

rated e at

2.3 H In thvertetop awereBI2Pmouwas the c

3. R 3.1 E The analysurroanalyaspeintradiffrat thsignof talignshocacqusampmisameaof tu2.4 aheliuoverfronttheoratioobtaforehordefieldparathe htherehead

Head mounti

he original PMebrae. They wand bottom ae centered in PED along wiunted in a sim

protected by connectors. Fi

Figure

RESULTS AN

External pres

analysis of ysis of the exounding the ysis is import

ects from the a-cranial presraction symmhe overlappingals. Synchronthe left and nment of the ck. Given tuisition, whicple every malignment cousurements. Seube boundarieand 2.6 times um, the theorerpressure is ext surface and

ory. Previous o between 2.6ained in very dhead pressure

er of 1.2 to 1.3d. The side prallel to the blaheadform. In efore, it is codform. Finally

ing

MHS study, hwere pressurizand stretched the tube as mith the acceler

milar fashion. Tbubble wrap

igure 6 shows

6: Reproduc

ND DISCUSS

ssure field

the results xternal transien

headform (Ftant to confirmtest, which casure variationetry is confirmg right and le

nicity betweenright signal

headform witthe samplingch was set at

microseconds), uld be detececond, the poes was studied

the benchmaetical reflectedxpected to be clearing effecexplosive tes

6 and 3.2 fordifferent ambie is in line w3 times the peressures are slast flow direct

general, the oncluded thaty, shock speed

heads were dzed and mounttightly by two

much as possibrometer brackThe net was cand a Kevlarboth models

cing the origin

ION

starts with tnt pressure fieFigure 7). Th

m a few differean influence tns. First, shomed by lookieft side pressun time of arrivs confirms tth the incomig rate of tt 1000 kHz

even a sligcted from theossible influend. The peak oark peak incidd to incident pbetween 2.6

cts are expectests with the hr a peak inciient conditionith what coulak incident ovlightly over thtion and the flpressure field

t the tube boud can be mea

disarticulated fted upside-doo chains that ble. The Hybrket. The headcut out aroundr™ sleeve, anmounted in th

 nal PMHS stud

the eld his ent the ock ing ure val the ing the (1

ght ese nce verpressure reent overpressupressure ampland 3. Natura

ed to make theheadform in fident overpre

n (dry air at arld be expecteverpressure, whe incident ov

flow is affected is very simiundaries havesured from th

Fig

from the torsown in a soft n

were fixed torid III neck w

dform was putd the exterior d taped to the

he shock wave

dy test conditi

ecorded on thure of 85, 120lification ratioally, the headfe reflected prefree-field conessure around round 5 deg Ced. The side pwhich is similaverpressure bd by the diffrilar to what we limited effehe delay betw

gure 7: Pressupressure sh

o between thnet. The net wo the tube wa

was similarily t upside downpressure sens

e tube wall toe generator.

ions with the B

he forehead w0 and 140 kPao, for this rangform is not pe

essure history nditions gener

85 kPa. HowC). In general,peak overpresar to what wa

because they araction from thwould be seenect on the loaeen the foreh

ure field for 8hock (filtered

he third and fwas attached aall. The specimremoved from

n in a soft nesors areas. Cao avoid slappin

BI2PED.

was on averagea respectivelyge of peak incerfectly flat odeviate away

rated amplificwever, these , the measure ssures were ons observed in are not complhe front portion in free-fieldading seen by

head signal an

85 kPa incidensignals).

fourth at the mens m the t and

abling ng of

e 2.2, y. For cident on the

from cation were peak n the free-

letely on of d and y the

nd the

nt

rightwas 3.2 E The spacof raprocresomorerespsignHz bcontverypeaknois

Fig

3.3 R Repethat throuexterThisadeqorigirepre85 kconsgenesecoloweattribwavfollohisto 3.4 P Fronparaanalywas simi

t/left pressuremeasured at 0

Effect of sens

responses of cing of approxaw occipital ICcessed signal. lve pressure e high frequenonse that wasificant contenband, which dtent above 400y similar, withk. In general, e from the sig

gure 8: Examp

Repeatability

eatability of mthe instrume

ughout the ternal damage

s is critical toquately and drinal PMHSesentative exa

kPa peak insistency was erator and the ond peaks of er than 5%,butable to vae. Slightly hig

owing peaks (ories. Overall

PMHS-BI2PE

ntal, parietal allel and perpysis in the cuthe closest to

ilar in shape b

es, which are m0.22 ms, whic

sor type and p

the piezo-resiximately 4 mmCP measuremThe results influctuations incy noise, as s on par with

nt up to approxdominates the 00 Hz comparh the exceptiothe low-pass

gnals.

mple of compar

y of ICP meas

measurements entation is rest series and

to the heado be able to raw relevant c

S data. Fiample of threncident pres

achieved wheadform. Vathe ICP sign

, part of wariations in tgher variabilit(up to 10%) the repeatabil

ED ICP comp

and occipital pendicular oriurrent article wo the same senbut significan

measured on tch yields a sho

post-processi

istive and the m, were comp

ment from bothndicate that thin the brain mseen in figure

h the validatedximately 4000early time do

red to the noison of an occa filter used in

rison of intra-sensor an

surements

is verified to esponding co

d that no intedform has oc

interpret the comparison wigure 9 showee shots donessure level.

with the blastariability on finals were ge

which can althe generatedty is observedand rest of thlity was consid

parison

ICP signals ientation and was to put on nsor location ntly lower tha

the same horiock speed of 4

ing

fiber-optic separed for all teh sensors alonhe fiber-opticmaterial. The e 8a and 8b. Nd piezo-resist0 Hz, with an omain responssier fiber-optisional variation the original

-cranial pressud piezo-resist

ensure orrectly ernal or ccurred.

results with the ws a

e at the Great

t wave irst and

enerally lso be

d shock d on the he ICP dered excellen

were comparfor all threethe occipital in the PMHS

an in the PM

Fsig

izontal plane. 450 m/s.

ensors, which ests. Figure 8 ng with their fr

sensor probafiber-optic s

Nevertheless, tive sensor. Bimportant con

se. The piezo-ic sensors. Onon on the ordl PMHS study

ure measuremtive sensor.

nt.

red with the incident presince the sens

S. In general, HS. This was

Figure 9, Exagnals from thr

lo

The delay for

were casted sshows a repre

frequency respably had sufficsensor signalsthe sensors sh

Both FFTs forntribution wit-resistive sens

nce filtered, boder of 10% ony only remov

ments obtained

original PMHessure levels.sor final positfrontal ICP ms expected co

ample of filterree different teoading conditi

r the 85 kPa s

side by side wesentative exaponse and the cient bandwids had considehowed a frequr this test revthin the 1000-sors show minoth responses n the first preed high frequ

d from fiber op

HS study for The focus o

tion in the BI2

measurements onsidering tha

red occipital ICests using the ions

shock

with a ample post-

dth to erably uency ealed -1200 nimal were

essure uency

ptic

both of the 2PED were

at the

CP same

BI2PICP 3.4.1 The incidwhicnega

phasstrescontattribwavtensifromPMHin ththe eobsein thvariafrequinitiareach0.4 m

Figu

was resuincrean inthe modthe dhavehighFigusignorigiThe 700 resoncont

PED frontal seshowed a lev

1 Parallel orie

direct compadent overpressch was the timative peaks we

The signals ses of compresses in the reatre-coup injurbuted to the ce at the back sile wave can b

m the front reaHS study alreahe brain [10]. external blast erved is likelyhe exterior trations due to uency phenomal tensile phahing the backms).

ure 10 , Direc

In terms of in very good

lts over the ease in blast increase in BIPMHS. The

dels was on thdamping of the a stiffer respher frequencyure 11 shows tals for the 85in of the diffePMHS show

Hz whereas nance peak tent around b

ensors were cel of correlati

entation – Occ

arison of occisure is shown meframe wherere measured.indicate that

ession and tear portion of thry) has been compressive wskull free surfbe transmittedaches the backady suggestedIt is importanwave even re

y linked to a nransient pressrelative brain

menon [2-4]. ase is of a higk of the head a

ct comparison for a pa

ICP magnitudd agreement w

three testedincident peakI2PED ICPs tmain discrep

he main oscillhat oscillation.ponse than they and for a sthe Fast Fouri5 kPa shock. Terence in the rws a diffuse

the BI2PEDaround 110

etween 50-20

considerably mon to the PMH

cipital region

ipital ICPs foin Figure 10.

re most of the the brain mat

ension at a nohe brain followreported in t

wave initiated face. Because d to the back ok. Based on sd that there want to understaneach the back

natural frequensure history, tn-skull motionIt is interestin

gher magnitudat the same tim

of intra-cranarallel exposu

de, the BI2PEDwith the origid blast severk overpressurethat was verypancy betweelation frequen. The BI2PEDe PMHS, oscishorter periodier Transform This plot con

response of boresonance pe has a simil

00-1200 Hz. 00 Hz, respon

more distant fHS that was cl

ICP

or parallel shoThe ICP histo

e ICP variation

terial is first loticeably specwing shock lothe literature at the front ostress waves

of the brain bekull strain meas a direct cornd that stressek of the head.ncy of the skuthey are the n would occung to note thde in every sime as the sku

ial pressure hures to differen

D response inal PMHS rities. The e generated y similar to en the two ncy and on appears to illating at a d of time. of the ICP

nfirmed the oth models. eak around lar diffuse Frequency

nsible for the

from the skulllose to that of

ocks of 85 kPories are plottns were obser

loaded in tenscific frequencoading of the

already [7-9of the skull wtravel faster in

efore the initiaeasurements arrelation betws in the skull The specific

ull itself. Sincresult of inte

ur on a longerhat the first coignal. This co

ull back face i

history betweent overpressu

ICP fluctuati

Figure 1occipital I

BI2PED he

l surface (seef the occipital

Pa, 120 kPa aed over the firrved and wher

sion and then cy. The develfront surface

9]. This phenwhich is reflect

n the skull thaal compressiveat various loca

ween skull stracan travel forfrequency of

ce these oscillaernal wave ar time frame ompressive phould be due ts unloaded (a

en PMHS and re level.

ions over lon

11: Direct comICP FFT betw

eadform for a pto a 85 kPa s

e Table 2). PaICPs.

and 140 kPa rst 8 ms of sigre the positive

oscillates betwlopment of teof the skull (c

nomenon has ted off as a tean in the braine wave transmations, the oriain and ICP hirth and back bf oscillation thations are notctivities. Thesince it is a lhase followinto the shock iafter approxim

BI2PED headf

nger intervals,

mparison of thween PMHS aparallel exposshock.

arietal

peak gnals, e and

ween ensile coup-been

ensile n, this mitted iginal istory

before hat is t seen e ICP lower g the in air

mately

dform

, also

he and sure

showlongapprare TherPMHThe tightis noand

Willdefoto gassemis a influsuturthis physresp 3.4.2 Occioverparaoscilphascastedefothe bby in

perpwhicresonfrequobsewiththe sthe sBI2P

Figu

wed some disger time scaleroximately 50likely attriburefore, they arHS head fromfact that this

ten in the net. ot attached to the flax and te

The diffuse linger [2,3] wormations (i.e global head rmbly is slightgood match

uences its stiffres surely affesecond reson

sical model. Tonse may fall

2 Perpendicul

ipital ICP corpressure are sallel shock, thellating betweese is expecteded in the right

ormations whibrain stress stncreasing blas

A level of copendicular oriech ICPs fluctunance arounduency (500 Herved that damh the bone on skull may respstress state in PED, which ha

ure 12: Direc

crepancies. Ine in the time

0 Hz and was utable to a mre inevitably i

m the body. Ths was not obsMore so, the any spine-likeentorium mem

700 Hz resonwhich associa

skull flexure,esponse. Thetly too stiff. Tfor skull mat

fness. Indeed, ect the local a

nance frequencThe present co within the sc

lar orientation

omparisons fshown in Figue results showen phases of ted since the right hemisphere. ich loads the btate following st severity is s

orrelation betwentation. Peakuate is again

d 1100-1200 HHz) than in thmped harmoniwhich the str

pond quasi-inthe brain und

as a uniform h

ct comparison for perpe

n particular, ae domain in tnot seen with

more global minfluenced by hese motions aserved on the BI2PED brain

e structure in tmbrane.

nant frequencyated the respo, direct stress e higher resonhe magnitudeterial. Howevthe human sk

and global strucy should be omparison is batter of a larg

n – Occipital r

for perpendicure 12. The IC

w that the rightension and coht side of the Very rapid cobrain region d exposure to a

similar to wha

ween PHMS ak ICPs are ver

higher in theHz, the PMHhe parallel oric oscillationsrain gauge wandependently oderneath [10]. homogenous s

of intra-craniendicular expo

a lower oscillthe original P

h the BI2PEDmotion of the

the head mouare most proba

BI2PED couln is not expectthe cranial cav

y observed on onse of the htransmission)nance frequen

e of the ICP bever, the homokull is not a houctural responconsidered b

based on a veer population

region ICP

cular shocks CP histories wt occipital reg

ompression at head is expos

ompressive lodirectly undera blast is senst was seen in

and BI2PED sry similar for e headform. W

HS ICPs oscilrientation (70s seen in the sas mounted. Itof each other This would e

skull compare

ial pressure hosures to diffe

atory fluctuatPMHS resultsheadform. Thhead or to

unting techniqably not repreld simply indted to move invity. Its move

the PMHS ishead to high ) rather than tncy in the Being similar, igeneous geomomogenous strnse. Neverthelefore trying t

ery limited amresponse.

of 85, 120 were again plotgion is now fir

a specific freqsed first to the

oading on one rneath in comsitive to orienthe parallel or

similar to the pthe three blasWhile the BI2

lation appear00 Hz). In theskull strain sigt was hypotheand that the l

explain the unid to the PHM

history betweenrent overpress

tion was obses. This fluctuhese low freqrelative brain

que and the disesentative of adicate that it wn a biofidelic fement is only l

in accordancvelocity load

to relative skuBI2PED suggeit is likely thatmetry of the Bructure. The dess, a study oto match it to

mount of PMH

and 140 ktted over 8 msrst loaded in cquency. The ie shock and thside of the sk

mpression. Thintation. The ICrientation.

parallel shockst severities bu2PED ICPs aed to occur ae original PMgnals appeare

esized that theocalized defoiform frequenS.

n PMHS and Bsure level.

erved over a muation occurrequency fluctuan-skull movemsarticulation o

a realistic respwas more strofashion yet sinlimited by the

ce with the woding to local ull-brain motioests that the t the material BI2PED skulldifferent bonen the variabili

oo precisely wHS and the BI2

kPa peak incs. Contrarily t

compression binitial comprehe ICP sensor

kull generates is emphasizesCP increase ca

k is observed iut the frequen

again show a at a slightly l

MHS study, ited to be assoce bones compormation may

ncy response o

BI2PED headf

much ed at ations ment. of the onse. ongly nce it

e CSF

ork of skull on or skull itself

l also s and ity of

with a 2PED

cident to the

before ession rs are local

s how aused

in the ncy at

clear lower t was ciated osing drive

of the

dform

4. CONCLUSION Previously reported PMHS blast tests were methodically replicated using the DRDC Valcartier research center physical head model named BI2PED. The loading conditions, instrumentation type and position as well as head mounting technique were reproduced to ensure that the only difference between the two series of experiments was the use of BI2PED or PMHS. The loading obtained from the blast wave generator was compared to free-field blast loading and it was confirmed that the generator produced relevant loading magnitude and duration for the study of blast-induced TBI. Excellent reproducibility was obtained on the BI2PED intra-cranial pressure measurements. A direct comparison of measured occipital ICP revealed good agreement between the headform and PMHS for three blast intensities (80, 100 and 120 kPa) and two blast orientations (parallel and perpendicular). The ICP magnitudes were particularly close to the PMHS ones, while the frequency of the oscillations was slightly higher in the headform. The BI2PED exhibited specific ICP oscillations around 1000-1200 Hz in both orientations while the PMHS ICP oscillations were respectively around 700 Hz and 500 Hz in the parallel and perpendicular orientations. Based on previous work correlating skull local deformation to early-time ICP variations, these frequencies are linked to the skull natural resonance frequencies. The results suggest that the BI2PED skull assembly may be slightly too stiff compared with the chosen PMHS. The design of the BI2PED skull could be refined but matching its response of a very limited amount of PMHS specimen should be avoided and a survey on the variability of human skull modal response should be done first. There were a few discrepancies on the longer duration ICP variations. However, those variations are not believed to be representative of real human response since the head is no longer attached to the neck and body. The neck provides a compliance that is very different from the soft net in which the heads were mounted for the tests. Lower frequency ICP variations are associated with skull-brain relative motion and head global motion. The BI2PED brain is currently not attached to any spine-like structure in the cranial cavity and its movement is only constrained by the CSF and membranes so it is not currently designed to replicate relative brain-skull relative motion. However, the representative mass and center of gravity of the headform should ensure that a representative global motion can be obtained in an ideal test configuration. Overall, the BI2PED was in very good agreement with the PMHS on the early-time ICP variations, which is where the highest pressure peaks are observed. Assuming that blast-induced TBI may be correlated with these early peaks in ICP, the BI2PED would represent a very useful tool to assess the severity of different blast scenarios and help with the performance evaluation of protective headwear systems. 5. REFERENCES [1] Bass C.R., Panzer M.B., Rafaels K.A., Wood G., Shridharani J., Capehart B., Brain injury from

blast, Annals of biomed Eng., 40(1) (2012), 185-202. [2] Willinger R., Taleb L., Kopp C., J of Neurotrauma,12(4) (2012), 743-754 [3] Willinger R., Biomechanical aspects of head trauma, Schmidt-Romhild, Lubeck, 1997, p29-57 [4] Viano D.C, Biomechanics of head injury: Towards a theory linking head dynamic motion. Brain

tissue deformation and neural trauma. Proceedings of the Stapp Car Crash Conf. 32: 1-20, 1988 [5] Hodgson V.R., Gurdjian E.S., The response characteristics of the head with emphasis on

mechanical impedance., Proceedings of the Stapp Car Crash Conference 11: 79-83, 1967 [6] Stalnaker R.L., Fagel J.L., Driving point impedance characteristics of the head, J. of biomech., 4

(1971), 127-139. [7] MossWC. King MJ, Blackman EG, Skull flexure from blast waves: a mechanism for brain injury

with implication on for helmet design, Phys Rev Let, 103(2009), 108702. [8] Ganuple S., Alai A.,Plougonven E., Chandra N., Mechanics of blast loading on the head models in

the study of traumatic brain injury using experimental and computational approaches, Biomechanics and Modelling in Mechanobiology, Springer, 2012.

[9] Gupta R.K., Przekwas A., Mathematical models of blast-induced TBI, Frontiers in neurology, 4 (2013).

[10] Bir C., Bolander R., Leonardi A., Ritzel D., Vandevord P., Dingell J.D., A biomechanical prospective of blast injury neurotrauma, NATO RTO HFM-207 Symposium on A survey of Blast Injury across the full landscape of military operations, Halifax, 2010.

[11] Ouellet S., Bouamoul A., Gauvin R., Binette J.S., Williams K.V., Martineau L. "Development of a Biofidelic Head Surrogate for Blast-Induced Traumatic Brain Injury Assessment", Proceedings of the Personal Armor System Symposium, Nuremberg, Germany, 2012

[12] Leaonardi A., Bir C., Ritzel D., VandeVord P., Intracranial pressure increases during exposure to shock wave. J of Neurotrauma, 28 (2011), p.85-94.