Three-LayerVolumeConductorModelforSurfaceEMG567
FIGURE1.Three-layeredcylindricalvolumeconductorcon- gurationwitheccentricsource.Thecrosssectionontheleft perpendiculartothecylinderaxis…illustratestherela-tionbetweenthelocalcoordinatesRandr,thecylindricalcoordinates and ,andthepositionsofaxis,source,andobservationpoint.Thecrosssectionontheright paralleltoandincludingtheaxis…showstheparametersa,b,andc,describingthecon guration,aswellasthecylindricalcoor-dinatez.Thesource ber…isshownwithitsmotorendplate,i.e.,thepositionoftheneuromusculartransition,halfwayalongthe ber.ThedepthofthesourceequalscÀR.Elec-trodesaredimensionless pointelectrodes….
countthreedifferentlayers brain,skull,andskin .Inapreviouspaper,23wehypothesizedthatthepotentialspreadinginsEMGwasduetoasimilarcauseasinEEGandextendedthetwo-layermodeltothreelayers.Thesethreelayersthencanbetakentorepresentalargeinnercylinderofmusculartissue,anintermediatelayerofsub-cutaneousfat,andanouterlayerofrelativelywell-conductingskin seeFig.1 .ThisconceptwasrecentlyfollowedbyFarinaandMerlettiintheirimplementationofathree-layermodelinCartesiancoordinates.9
Thepresentpaperdescribesthemathematicalbasisofthecylindricalvolumeconductormodel,anddiscussesanumberofaspectsregardingitsimplementation.Throughout,thefocuswillbeonvolumeconduction.TheotheressentialingredientofanysEMGmodel,anaccuratedescriptionofthebioelectricsource,isdis-cussedinAppendixA.Tofacilitatethescienti candeducationaluseofawell-testedversionofthemodelinandbeyondourowndepartment,wedevelopedasoft-warepackagewithauser-friendlyinterface.Thisanalyti-calvolumeconductor ANVOLCON softwarepackageispresentedinAppendixB.
METHODS
MathematicsoftheThree-LayerBoundedVolume
ConductorWithsEMGmodels,weseektounderstandcertainphenomenaobservedinexperiments.Forthatpurpose,themodelshavetoincorporaterelevant‘‘real-life’’char-acteristics.Itdependsonthepurposeofthemodelwhichtheseshouldbe.29Toreproducevolumeconductedpo-tentialdistributionsproperly,itisessentialthatthege-ometryofthemodeldesignre ectstherecordingcon-
gurationofexperiments.Often,theseexperimentsareperformedonhumanlimbsthat,asa rstapproximation,maybeconsideredcylindrical.Theyare nite,bothinlengthand,moreimportantly,inradialextension.SincemeasuredsEMGsignalsaredominatedbythecontribu-tionsofsuper cialmotorunits,themodelshouldsupporttheuseofeccentricsources.Itshouldalsoallowforthein uenceofthetwotissuetypes—skintissueandsubcu-taneousfat—thatliebetweenthesourceandtherecord-ingelectrode.Infact,especiallyforsuper cialsources,thesetwotypesmayconstitutethemajorpartoftheintermediatetissueandtheircombinedeffectsonthevolumeconductedpotentialarecertainlynotnegligible.Finally,tissueanisotropyneedstobeincorporated,espe-ciallytheanisotropyofmusculartissuewithpoorcon-ductivityinthedirectionperpendiculartothe berscom-paredtothatinthe berdirection.
Inahumanlimb,thethreetissuetypesarearrangedinamoreorlessconcentriccon guration.Fortheconstruc-tionofmodels,thiscylindricalsymmetryimpliesthatitispossibletouseananalyticaldescriptionofthevolumeconductedpotential.Thus,Heringaetal.14tackledtheproblemofaneccentricsourceina nite,anisotropicbutsingle-layervolumeconductor.Thiscon gurationisre-alisticinthesensethatittakesintoaccountthe nitenessofthevolumeconductoraswellastherelativelysuper- cialsourcespertainingtoexperiments.Thethree-layermodelpresentedhereisanelaboration12ofthetwo-layermodeldescribedbyGootzenetal.ItincorporatesalloftheaforementionedelementsandisbasedonthePoissonequation,astaticapproximationofthequasistaticvol-umeconductionphenomena.Followingtheapproachoutlinedinpreviouspapers,12,14,15thegeneralsolutionofthisequationforeachofthethreelayersofthecon gu-rationofFig.1isgivenby
d
1z1 , ,k2 Kr10
G k
r 1 kr
n
e
in
An k I 1zn
k ,1r
1
d
2 , ,k
2K r20
r 2z2 k G k
r
n
e
in
Cn k I 2zn
k
2r
D2zn k Kn
2 k
, 2

r
热门文档