proximitynewworld

BrainshapeconvergenceintheadaptiveradiationofNewWorldmonkeysLeandroAristidea,SergioFurtadodosReisb,AlessandraC.
Machadoc,InayaLimac,RicardoT.
Lopesc,andS.
IvanPereza,1aDivisiónAntropología,FacultaddeCienciasNaturalesyMuseo,UniversidadNacionaldeLaPlata,ConsejoNacionaldeInvestigacionesCientíficasyTécnicas,1900LaPlata,BuenosAires,Argentina;bDepartamentodeBiologiaAnimal,UniversidadeEstadualdeCampinas,CEP13.
083-862Campinas,SoPaulo,Brazil;andcLaboratóriodeInstrumentaoNuclear,CentrodeTecnologia,UniversidadeFederaldoRiodeJaneiro,IlhadoFundo,CEP21949-900RiodeJaneiro,RiodeJaneiro,BrazilEditedbyChetC.
Sherwood,TheGeorgeWashingtonUniversity,Washington,DC,andacceptedbytheEditorialBoardJanuary6,2016(receivedforreviewJuly22,2015)Primatesconstituteoneofthemostdiversemammalianclades,andanotablefeatureoftheirdiversificationistheevolutionofbrainmorphology.
However,theevolutionaryprocessesandecologicalfactorsbehindthesechangesarelargelyunknown.
Inthiswork,weinvestigatebrainshapediversificationofNewWorldmonkeysduringtheiradaptiveradiationinrelationtodifferentecologicaldimensions.
Ourresultsrevealthatbraindiversificationinthiscladecanbeexplainedbyinvokingamodelofadaptivepeakshiftstouniqueandsharedoptima,definedbyamultidi-mensionalecologicalnichehypothesis.
Particularly,weshowthattheevolutionofconvergentbrainphenotypesmayberelatedtoecologicalfactorsassociatedwithgroupsize(e.
g.
,socialcomplex-ity).
Together,ourresultshighlightthecomplexityofbrainevolutionandtheecologicalsignificanceofbrainshapechangesduringtheevolutionarydiversificationofaprimateclade.
comparativemethod|primates|adaptiveevolution|geometricmorphometrics|PlatyrrhiniAdaptiveradiation,definedastherapidandexceptionaladaptivediversificationofasinglephylogeneticlineageintoavarietyofdifferentecologicalniches(1,2),isthoughttobeoneofthemainevolutionaryprocessesgeneratingbiodiversityonEarth(3).
Althoughseveraladaptiveradiationshavenowbeenthoroughlystudied(e.
g.
,Africancichlids,Caribbeananoles,Gal-apagosfinches,etc.
),westillneedmorestudiesfromwhichgen-eralizationsonthisprocesscanbedrawn(2).
Particularly,underwhatconditionsalineagewillundergoadaptiveradiationhaslongbeendebated,althoughbothempiricalandtheoreticalmodelspointtotheexistenceofecologicalopportunityasamajorfactor(2,3).
Thisopportunitymayappear,amongothers,throughthecolonizationofanewarea,theextinctionofastrongecologicalcompetitor,ortheevolutionofanewtrait—a"keyinnovation"—thatallowstheutilizationofresourcesinwaysthatwerenotpreviouslypossible(3,4).
Therefore,asignificantdimensionofadaptiveradiationsisthediversificationofecologicallyrelevantphenotypictraits(2,5).
Amongthestudiedcasesofadaptivera-diations,severalecologicallyrelevanttraitshavebeenidentified.
Forexample,relativelimbsizeinanoleslizards(6),beakshapeinDarwin'sfinches(7),oroverallbodyshapeincichlidfishes(8).
Remarkably,atraitthathasreceivedlessattentioninthestudyofadaptiveradiationsamongvertebratecladesisbrainmorphol-ogy.
Brainshavesubstantialecologicalandadaptiveimportancebecausetheyunderliethebehaviorthatallowsananimaltosuc-cessfullyinteractwithitsenvironment.
Inthissense,primatesconstituteanotableexample,astheevolutionofbrainmorphologyisoneofthemostprominentfeaturesoftheirdiversification(9).
Primatesgenerallyengageincomplexforagingandsocialbehaviors(10),andtherefore,theevolutionofenhancedcognitivecapacitiesassociatedwithenlargedand/ormorecomplexbrainsmayconsti-tuteamajoraxisoftheiradaptiveecologicaldiversification.
Forexample,possessingalargebrainisperhapsthesinglemostrel-evantphenotypictraitofourownspecies,Homosapiens.
Specifically,previousworkshavepointedoutthattheevolutionofenlargedbrainscouldbeimportantecologicallybecausethesechangesarerelatedtotheacquisitionofthecognitiveabilitiesrequiredtosustaincomplexsocialinteractions—abehavioraltraitprobablyinvolvedintheoriginandmaintenanceoftheevolu-tionarysuccessofprimates(thesocialbrainhypothesis)(11).
However,althoughbrainsizehasbeentraditionallythepreferredmeasuredtraitinevolutionarystudies,brainsarenotuniformstructuresbutareconstitutedbyseveralanatomicallydistinctfunc-tionalsystemsormodules(e.
g.
,neocortex,cerebellum,etc.
).
Pre-viousworkhassuggestedthatthesemodulescanvaryintheirrelativesizeamongspeciesofsomemammalianclades(includingprimates)(12–14)and,moreover,thatthesemosaicchangescanbetterexplainneuraldiversitythantotal—absoluteorrelative—brainsizechangesalone(e.
g.
,amongprimates)(14).
Additionally,inseveralvertebrateclades,partofthismosaicvariationhasbeenrelatedtoparticularbehavioralcapacities(e.
g.
,refs.
15and16).
Thus,braindiversificationisprobablyacomplexprocessinvolvingseveraldimensionsofchangeoccurringduringthedivergenceofthespeciesandalongmultipleecologicalaxes.
Moreover,asinpreviousadaptiveradia-tionstudieswhereshapeisthemostecologicallyrelevanttrait(6,7),brainshape(i.
e.
,therelativepositionandsizeofindividualbraincomponents)isprobablyamoreimportantaspectofbrainevolu-tionthanitstotalrelativesize.
However,despitethepotentialecologicalrelevanceofbrainshapevariationforprimates,itsevo-lutionarydynamicshavebeenlessstudiedinmacroevolution(14).
NewWorldmonkeysorplatyrrhines,oneofthethreemajorprimateclades,constituteanexampleofamajormammalianadaptiveradiationthatunfoldedinisolationinCentralandSignificanceTheevolutionarydiversificationofbrainmorphologyisoneofthemostprominentfeaturesoftheprimateadaptiveradiationandalikelydeterminantofprimateevolutionarysuccess.
However,theecologicalfactorsresponsibleforthediversificationoftheprimatebrainarelargelyunknown.
Inthiswork,weuseacomparativeapproachtostudybraindiversificationduringtheadaptiveradiationofamajorprimateclade—theNewWorldmonkeys.
Weshowthatbrainmorphologyevolvedinassociationwiththeoccupationofseveraleco-logicalniches,andthatconvergenceinbrainmorphologyamongcladesmaybeassociatedwithanevolutionaryin-creaseinthecomplexityofsocialbehaviors.
Authorcontributions:L.
A.
andS.
I.
P.
designedresearch;L.
A.
andS.
I.
P.
performedresearch;A.
C.
M.
,I.
L.
,andR.
T.
L.
performedimageacquisition;S.
F.
d.
R.
,A.
C.
M.
,I.
L.
,andR.
T.
L.
con-tributednewreagents/analytictools;L.
A.
andS.
I.
P.
analyzeddata;andL.
A.
,S.
F.
d.
R.
,andS.
I.
P.
wrotethepaper.
Theauthorsdeclarenoconflictofinterest.
ThisarticleisaPNASDirectSubmission.
C.
C.
S.
isaguesteditorinvitedbytheEditorialBoard.
1Towhomcorrespondenceshouldbeaddressed.
Email:ivanperezmorea@gmail.
com.
Thisarticlecontainssupportinginformationonlineatwww.
pnas.
org/lookup/suppl/doi:10.
1073/pnas.
1514473113/-/DCSupplemental.
2158–2163|PNAS|February23,2016|vol.
113|no.
8www.
pnas.
org/cgi/doi/10.
1073/pnas.
1514473113DownloadedbyguestonDecember31,2020SouthAmericaduringthelast25–35Ma,resultinginbroadecologicalandmorphologicaldiversity(17,18).
Noticeably,previousworkshavepointedtotheputativeoccurrenceofsev-eralevolutionaryindependentprocessesofincreaseandde-creaseinrelativebrainsize(19–21)inthisclade,indicatinganextensivediversificationofbrainmorphology.
However,itisunknownhowbrainshapehasevolvedinthecontextofthisprimateadaptiveradiation.
Inthisstudy,weinvestigatetheprocessofbrainshapedi-versificationofNewWorldmonkeysduringtheiradaptiveradiationinrelationtodifferentecologicaldimensions.
Wefirstquantifybrainshapevariationusingvirtualreconstructionsofcranialendocastsandgeometricmorphometricsmethods,andthenexploretheevolu-tionaryprocessesunderlyingbraindiversificationusingphylogeneticcomparativemethods.
Wehypothesizedthat,ifbrainshapeisanecologicallyrelevanttraitforplatyrrhines,apatternofvariationthatdepartsfromneutralmodelsofevolutionwouldbeexpected.
Toaddressthishypothesis,weanalyzedtherelationshipbetweenphe-notypicvariationandthebranchingprocessofthespeciesbasedonaBrownian(random)modelofevolution.
Next,weinvestigatedwhetheramodelinvolvingchangesinadaptivepeaksinamacro-evolutionarylandscapeforbrainshapecanbetteraccountfortheobserveddiversity.
Moreover,consideringthatalikelyoutcomeofanadaptiveradiation,giventheimportanceofecologicalfactorsinshapingdiversity,istherepeatedevolutionofsimilarphenotypes(22),weexplicitlytestfortheexistenceofbrainmorphologicalconvergenceintheplatyrrhineradiation.
Finally,weexplorepre-viouslyhypothesizedecologicalfactors(e.
g.
,groupsize)(11)drivingbrainshapediversificationandconvergenceinthisprimateclade.
ResultsBrainShapeVariation.
Tostudythepatternofvariationinbrainshapeintheplatyrrhinecladeweperformedaprincipal-componentanalysis(PCA)ontheProcrustesshapecoordinatesof399land-marksandsemilandmarksmeasuredon49species(TableS1)thatdescribeeachspeciesexternalbrainmorphology(Fig.
1AandTableS2).
Fig.
1Ashowsthemaintrendsinshapevariation,asrepresentedbythefirsttwoprincipalcomponents(PCs),whichtogetheraccountfor61%ofthetotalvariation,alongaprojectionofthephylogenyonthemorphospace.
Ageneralseparationamongfamilies,andalargelyphylogeneticallystructuredoccupationofthemorphospacecanbeobserved,althoughremarkableshapeprox-imityoccursamongmembersfromthethreefamilies(theatelidsAteles,Brachyteles,andLagothrix;thepithecidsChiropotesandCacajao;andthecebidCebus)showinghighandsimilarenceph-alization(i.
e.
,totalrelativebrainsize)values[relativeendocranialvolume(rECV)](Figs.
S1andS2),suggestingtheexistenceofbrainshapeconvergenceamongtheseclades.
Furthermore,AlouattaandSaimirispecies,extremesontheencephalizationscale,alsoappearasextremesinshapespace.
Althoughtheseresultsmaysuggestthatshapeandtotalrelativesizedescribethesamemorphologicalpropertiesofthebrain,onlyafractionofthetotalshapevariationcanbeaccountedbyencephalization[phylogeneticgeneralizedleast-squaresregression(PGLS):λ=0.
74;R2=0.
11;P=0.
010].
BrainshapedifferencesalongPC1(Fig.
1BandMovieS1),whichexplains43%ofvariation,aremainlyrelatedtogeneralen-largementoftheneocortexareawithrespecttothestemandcerebellumareas;andtherelativepositionofthestem,whichcontraststheposteriorlocationinAlouattawiththemoreventralpositioninSaimiri.
Specifically,shapechangesintheneocortexareconcentratedmainlyinthefrontal,parietal,andoccipitallobes,withthelargestvariationobservedintheoccipitallobe,whichex-pandsposteriorlyandinadorsoventraldirection.
Thesechangescontributetoanoverallmoreglobularbrainshapeforpositivescores.
Additionally,thebrainbaseexhibitsamoreventrallyflexedmorphologyamongspecieswithmorepositivescores.
Furthermore,althoughPC1ishighlycorrelatedwithrECV(PGLS:λ=0.
87;R2=0.
77,P=0.
001;Fig.
S3),ourresultsclearlyshowthatPC1describescomplexmorphologicalchangesthatcannotbechar-acterizedmerelybyrelativetotalbrainvolumemeasurements.
ShapevariationalongPC2(18.
5%oftotalvariation)ismainlyrelatedtochangesintheprefrontalarea,whichbecomeslessex-panded,particularlyalongthebrainbasemidline,inthepositivescoresdirection;andwithageneralincreaseinglobularityfornega-tivescores(Fig.
1BandMovieS2).
Importantly,althoughNewWorldmonkeysexhibitaremarkablebodysizediversity(0.
01–10kg.
),brainshapevariationcannotbemerelyattributedtoevolutionaryallometriceffects(PGLS:λ=0.
69;R2=0.
15;P=0.
003).
BrainShapeConvergence.
Next,weexplicitlytestwhetherthepatternofshapeproximityexhibitedbymembersofthethreefamiliesdepartsfromwhatcanbeexpectedbychance.
Wecal-culatedtwopattern-basedmeasuresofmorphologicalconver-gence(C1andC5)usingthefirsttwoPCsandcomparedthemagainstanulldistributionobtainedusingsimulationsunderarandommodelofevolution(Brownianmotion).
Althoughonlythefirstmeasureshowedastrongconvergencesignal(C1=0.
63;P=0.
01;C5=5;P=0.
22),theseanalysespointtotheexistenceofnonrandomevolutionaryprocessesgeneratingbrainshapesimilarityamongtheseclades.
TheEvolutionofBrainShape.
Phylogeneticsignalanalysesallowedtofurtherunraveltheevolutionarydynamicsofbrainshapediversification.
Complementingtheobservedcladewiseconver-gence,KstatisticvaluesbothforPC1(K=1.
30,P=0.
001)andPC2(K=2.
32,P=0.
001)showedthatvariancewasconcentratedamongcladesratherthanwithinclades,indicating,besidesaconsiderablephylogeneticsignal,apatternofearlybrainmor-phologicaldiversificationthatdepartsfromneutralexpectations.
Disparity-through-time(DTT)plotsfurthersupportthisview:anearlyfastdiversificationofthemorphologicalaspectsrepresentedbyPC2isobserved(Fig.
2);whereas,forPC1,adropindisparitybeyondtheneutralexpectationcanbeseenbetween17and12Ma,pointingtoaburstofbrainshapeevolutionapproximatelyassociatedwiththeoriginoftheextantsubfamilies(Fig.
2andFig.
S2),followedbyaslowdownindisparitychangesuntilpresenttimes.
Moreover,theseresultsalsopointtotheexistenceofsep-arateburstsofbrainevolutionduringtheplatyrrhineradiation,eachaffectingdifferentaspectsofbrainmorphology.
Finally,weexplicitlyaskedwhetheramodelofadaptiveevo-lutioncouldbeinvokedtoexplainbrainshapediversificationinplatyrrhinemonkeys,andparticularly,toexplaintheobservedpatternofshapeconvergence.
Todothis,weimplementamodelselectionapproach,inwhichthelikelihoodsofseveralalterna-tiveevolutionaryscenariosarecomparedusinganinformationcontentcriterion[correctedAkaikeinformationcriterion(AICc)]tofindthebest-fittingmodel.
Theexploredmodelswerearandomwalk[modeledasBrownianmotion(BM)];earlyburst(wheretherateofBrownianevolutiondecaysexponentiallythroughtime);andseveraladaptiveevolutionscenariosmodeledasOrnstein–Uhlenbeck(OU)processes(Fig.
S4).
InOUmod-els,nodesandbranchesofthephylogenyareassignedtodif-ferentselectiveregimesrepresentingphenotypicadaptivepeaksonaSimpsonianmacroevolutionarylandscape.
Thisway,di-rectionalandstabilizingselectioncanbemodeled.
Toconstructthesescenarios,wecombinedaprioribiologicalhypotheseswiththeuseoftheSURFACEmethod(SupportingInformationandFigs.
S4andS5),which,usingadata-drivenalgorithm,findsthemodelthathasthebeststatisticalfit.
Bycombiningtheseap-proaches,wegeneratedalikelybiologicalhypothesiswithahighabsolutestatisticalsupport,excludingthepossibilityofchoosingonlythebestamongseveralbadbiologicalhypotheses(seeSupportingInformationfordetails).
OurresultssupportasthebestfitanOUmodelbasedonamultidimensionalecologicalnichehypothesis(AICcweight=0.
94;Fig.
3andTableS3)inwhichdietcomposition,locomotionstrategy,andgroupsizeareAristideetal.
PNAS|February23,2016|vol.
113|no.
8|2159EVOLUTIONDownloadedbyguestonDecember31,2020themainadaptivelandscape-definingvariables.
Moreover,thissuggeststhatecologicalfactorsrepresentedbygroupsizecouldberesponsiblefortheobservedcladewiseconvergenceinplatyr-rhines,asfurtherindicatedbysignificantshapedifferencesbe-tweenspecieswithlarge(i.
e.
,morethan15individuals)andsmallgroups(phylogeneticmultivariateANOVAforPC1andPC2;P=0.
001,R2=0.
74).
Overall,ourresultsindicatethatadaptivebrainshapediversificationinplatyrrhinescanbeinvokedasaplausibleexplanationfortheobservedpatterns,andparticularly,toaccountforbrainshapeconvergenceintheclade.
DiscussionTheevolutionofrelativelylargeandcomplexbrainsisperhapsoneofthekeydeterminantsofprimateevolutionarysuccess.
Additionally,severalextantandextinctprimatecladesconstituteputativeexamplesofadaptiveradiations(10),whereoneofthePC1(42.
3%)PC1PC2PC2(18.
5%)S.
bicolorS.
midasS.
mystaxS.
fuscicollisC.
jacchusC.
penicillataC.
kuhliiC.
auritaM.
humeraliferM.
chrysoleucusM.
argentatusC.
pygmaeaC.
goeldiiL.
chrysomelasL.
rosaliaA.
azaraeA.
nigricepsA.
vociferansC.
apellaC.
robustusC.
albifronsC.
capucinusS.
sciureusS.
boliviensisA.
geoffroyiA.
marginatusA.
paniscusA.
chamekL.
lagotrichaB.
hypoxanthusB.
arachnoidesA.
macconnelliA.
seniculusA.
carayaA.
belzebulA.
guaribaA.
palliataA.
pigraC.
molochC.
brunneusC.
donacophilusC.
personatusC.
calvusC.
melanocephalusC.
satanasC.
chiropotesP.
monachusP.
irrorataP.
pitheciaCebidaeAtelidaePithecidaeAotus+RelativeECV++++ABFig.
1.
MorphometricanalysisofNewWorldmonkey'sbrainshape.
(A)Ordinationof49platyrrhinespeciesinthemorphospacedefinedbythefirsttwoprincipalcomponents(PCs)ofbrainshapevariation,whichtogetheraccountfor61%oftotalvariance.
EndocastatLeftshowsthemeasuredlandmarks(red),andcurves(green)andsurfaces(blue)semilandmarksoneachindividual(seealsoFig.
S6andTableS2).
Encephalizationvalues(relativeECV)foreachspeciesaredepictedbythecolorofeachdatapoint.
Additionally,thethreeplatyrrhinefamiliesareindicated,alongwiththegenusAotus,whosepositionintheplatyrrhinetreeiscontentious.
(B)Brainshapechangesassociatedwiththemainaxesofvariation.
ModelswereobtainedbywarpingasurfacemodelofthemeanplatyrrhineshapealongPC1andPC2scores.
SeealsoMoviesS1andS2.
2160|www.
pnas.
org/cgi/doi/10.
1073/pnas.
1514473113Aristideetal.
DownloadedbyguestonDecember31,2020expectedoutcomesisthemarkedadaptivediversificationofeco-logicallyrelevantphenotypictraits.
However,brainmorphologyhasreceivedlittleattentioninthestudyofprimateadaptivera-diationsdespiteitsmanifestecologicalimportance.
Moreover,mostpreviouscomparativestudiesofprimatebrainevolutionhavemainlyexploredthediversificationofrelativebrainsize,havinggenerallyoverlookedtheevolutionaryandecologicalimportanceofvariationinotherphenotypicdimensionsoftheprocessofbrainevolution,e.
g.
,variationintherelativesizeandposition—shape—ofthedifferentbrainmodules.
OurresultsindicatethatNewWorldmonkeyspresentanex-tensivevariationinbrainshapeand,importantly,thatonlyafractionofthisvariationisrelatedtototalrelativebrainsize,indicatingthatasignificantamountofinformationisnotcap-turedbythisvariable,asotherworkershavepointedout(23).
Relativebrainsizechangesaremostlyassociatedwitharelativeenlargementoftheneocortex,withmostlocalizedshapechangesconcentratedonthefrontalandoccipitalareas,andcorrelatedchangesinotherbrainmodules(Fig.
S3).
Additionally,inspeciesexhibitingthisneocortexenlargement,thebrainstemshiftsitsrelativepositiontoamoredownwardlyorientedlocation(Fig.
1BandMovieS1).
Therefore,althoughourapproach(i.
e.
,vir-tualendocastsandgeometricmorphometricsmethods)onlyallowedquantificationofexternalbrainmorphology,itcancontributevaluableinformationthatmaybeotherwiseover-looked,astheconjointchangesinrelativesizeandpositionofbraincomponentsorstructures—asdescribedabove—cannotbemeasuredonlywithvolumetricassessmentsofthewholebrain(i.
e.
,itsabsoluteorrelativevolume)orbrainparts.
Ourresultsalsoshowthatplatyrrhinesdisplayasignificantlyhighphylogeneticstructureinbrainshapevariation,asinotherpreviouslystudiedphenotypictraits(e.
g.
,refs.
21and24).
Al-thoughthisisnotunexpectedatmacroevolutionaryscales(25),ourresultsadditionallyshowthatseveralcladesexhibitsignif-icantlyconvergentbrainmorphologies,suggestingtheexistenceofecologicalornonneutralevolutionaryfactorsstructuringbrainshapevariation.
Moreover,ourmodelselectionresultsindicatethat,beyondthisclearphylogeneticstructure,braindiversificationamongplatyrrhinescanbeexplainedinvokingaSimpsonianmodelofadaptivepeakshiftstouniqueandsharedoptima,mainlydefinedbyamultidimensionalecologicalhy-pothesis.
Particularly,ourresultssuggestthatbrainshapedi-versificationintheplatyrrhineradiationmayhaveunfoldedinatleasttwostages(Fig.
3):anearlydifferentiation(i.
e.
,amongfamilies)withinanecologicaladaptivelandscapemainlyde-finedbydietandlocomotionstrategyaxes,followedbyseveralshiftsatthesubfamilyleveltoasharedadaptivepeakdefinedbysocialgroupsize,generatingshapeconvergence.
Thefirststagemayberelated—althoughnotexclusively—tothechangesrepresentedbyPC2,whichshowsastrongearlydiversificationpattern,strikinglysimilartothatofbodymass(Fig.
2),whichpreviousworksproposedtohavediversifiedwithinanadaptivelandscapeassociatedwithdietandlocomotionstrategies(17,18).
0510152025MaSaguinusbicolorSaguinusmidasSaguinusmystaxSaguinusfuscicollisCallithrixjacchusCallithrixpenicillataCallithrixkuhliiCallithrixauritaMicohumeraliferMicochrysoleucusMicoargentatusCebuellapygmaeaCallimicogoeldiiLeontopithecuschrysomelasLeontopithecusrosaliaAotusazaraeAotusnigricepsAotusvociferansCebusapellaCebusrobustusCebusalbifronsCebuscapucinusSaimirisciureusSaimiriboliviensisAtelesgeoffroyiAtelesmarginatusAtelespaniscusAteleschamekLagothrixlagotrichaBrachyteleshypoxanthusBrachytelesarachnoidesAlouattamacconnelliAlouattaseniculusAlouattacarayaAlouattabelzebulAlouattaguaribaAlouattapalliataAlouattapigraCallicebusmolochCallicebusbrunneusCallicebusdonacophilusCallicebuspersonatusCacajaocalvusCacajaomelanocephalusChiropotessatanasChiropoteschiropotesPitheciamonachusPitheciairrorataPitheciapitheciaFig.
3.
Time-calibratedphylogenetictreeforthestudiedNewWorldmonkeyspeciesshowingadaptiveregimesforthebest-fittingmodelofbrainshapeevolution.
Regimeswereassignedbasedondietcomposition,locomotionstrategy,andgroupsizewithexceptionofSaimiri,whichwasassignedbasedontheSURFACEresults(SupportingInformation).
Drawingsbroadlydepicttheecologicalcategoriesthatdefinetheseregimesandarenotintendedtorepresentancestralstates.
Dietsarerepresentedbyniche-definingfooditems(17).
Locomotionisrepresentedbytypicalbehaviors.
Groupsize(smallandlarge)isrepresentedbyencircleddots.
Convergentregimesaredefinedbyhavinglargegroupsize.
AotusandCallicebusareconsideredbysomeworkerstobesisterclades.
0.
00.
51.
01.
5RelativedisparityBMPC1PC2rECV0510152025Time(Ma)Fig.
2.
Disparity-through-time(DTT)plotsforbrainshape,encephalization(rECV),andlogbodymass(BM).
Relativedisparityateachpointindicatestheaverageextantdisparityofthesubcladesthathadanancestoratthattimewithrespecttothewholecladedisparity.
DashedlineandgrayshadowrepresenttheexpectationunderaBMmodelofevolution,estimatedviasimulations,andits95%confidenceinterval,respectively.
Aristideetal.
PNAS|February23,2016|vol.
113|no.
8|2161EVOLUTIONDownloadedbyguestonDecember31,2020Thissimilarityinthepatternsofdiversificationofbodymassandacomponentofbrainshapemaypointtotheactionofcommonselectivefactorsonbothtraits,oralternatively,totheexistenceofdevelopmentalconstraints.
Interestingly,weonlyfoundevidenceforamoderateevolutionaryallometrylinkingvariationinbothtraits(PGLS:λ=0.
92;R2=0.
25;P=0.
001);thus,modularoverintegratedexplanationsmaybefavored.
However,wecannotruleoutcompletelytheexistenceofintegratedchangesaffectingbrainmorphology.
Forexample,correlatedstructuralchangeswiththeface—whichmayreflectdietaryadaptations—couldbeamajorsourceoftheecologicalsignalinbrainshape.
Followingthisinitialdivergenceprocess,ourresultssuggestthatensuingshiftstoasharedadaptivepeak(Fig.
3),generatingcla-dewiseconvergenceatthesubfamilylevelinparticularaspectsofbrainmorphology(e.
g.
,expansionoftheneocortexanditscor-relatedincreaseinrelativebrainsize;Fig.
S3),couldbeasignif-icantmodeofbrainshapechangeinplatyrrhineevolution.
Particularly,amodelexplicitlyincorporatingbrainshapeconver-gence(Fig.
3)inrelationtotheseveralecologicalfactorsthatarelikelyrepresentedbygroupsize(e.
g.
,socialcomplexity;seebelow)(16)wasstronglyfavoredoverseveralnonconvergentmodels(Fig.
S4).
Recentstudieshavesuggestedthatphenotypicconvergencecouldbeafrequentphenomenoninmacroevolutionaryradiations(26–28),althoughthisistypicallyobservedinradiationswheretheconvergentspeciesevolvedingeographicisolationfromeachother.
Althoughsometheoreticalmodels(29)andempiricalex-amples(8)describetheemergenceofmorphologicalconvergenceinthesamegeographicarea,thesecasesrefertocommunitieswherethenumberofspecieslargelyexceedsthenumberofavailableecologicalniches.
Alternatively,ourresultssuggestthatbrainshapeconvergencewaslikelyattainedatalatestageoftheradiation,asdescribedabove,onlyafterdivergenceinseveralotherdimensionsofthenichespacewasachieved.
Thisscenariomaybeinagreementwithproposedgeneralmodelsandexamplesofvertebrateadaptiveradiations(4,30),inwhichcladesdiversifyinstages,forexample,firstdivergingintraitsallowingdifferentialexploitationofthemacrohabitat[e.
g.
,bodysizeforplatyrrhines(17,18)],followedbymorenarrowse-quentialpartitioningofnichedimensions(e.
g.
,microhabitat,dietspecializations,behavior,communication,etc.
).
Inthecaseofplatyrrhines,oneofthislatenichedimensionscouldberepre-sentedbythecomplexityofsocialinteractions(i.
e.
,acognitivedimensionoftheniche).
Convergentphenotypesexhibit,althoughnotexclusively,encephalization-relatedshapechanges.
Particu-larly,theneocortexisthemainstructureexhibitingarelativeenlargement.
Becausebrainsizeinprimatesscalesalmostisometricallywithcellnumber(31),relativelylargerbrainstruc-tureswouldhaverelativelyhighercomputationalpower(14),in-dicatinganincreaseinthecognitiveabilitiesrelatedwiththeenlargingneocortex.
Inthissense,ourresultssuggeststhatthesocialbrainhypothesis,whichpositsthatthelargeneocortexofprimatesevolvedinresponsetothecognitivedemandsarisingfromlivingincomplexsocialgroups(16,32),canbeinvokedtobetterunderstandtheevolutionatmacroevolutionaryscalesofbrainshapeandencephalizedphenotypesinanecologicalcontext,andparticularly,thefactorsgeneratingbrainmorphologicalcon-vergenceamongprimatespecies.
InthecaseofSaimiri,whichisthemostencephalizedgenusandwhichourresultsshowedtooccupyitsownadaptivepeak,additionalfactorsbeyondthoseconsideredheremaybedrivingitsevolution.
Summarizing,weshowthatbrainshapeconstitutesaneco-logicallyrelevantphenotypictraitintheplatyrrhineadaptiveradiation,andparticularly,thattheevolutionofspecificaspectsofbrainshapeprobablyallowedtheexploitationofadditionaldimensionsoftheplatyrrhineecologicalnichespace.
Overall,thisfurtherindicatesthatbrainshapeevolutioncouldbeofrelevanceforunderstandingothercasesofprimateand,moregenerally,vertebrateadaptiveradiations.
MaterialsandMethodsSample.
Oursampleconsistedof179adultskullsofbothsexesfrom49platyrrhinespeciesbelongingtoall17broadlyrecognizedgenera(TableS1).
Thisrepresentsalargesampleofplatyrrhinephylogeneticdiversity(40–60%ofextantspecies)(33,34).
MorphologicalAnalyses.
ForeachspecimenX-raycomputedtomographyormicro-computedtomographyscanswereacquiredordownloadedfrompublicrepositories(21).
Three-dimensionalvirtualendocastsweregeneratedfromthesedatafollowingathreshold-based2Dsegmentationprocedure(21,35).
Fromthesegmentedimages,asurface3Dmodel—thevirtualendocast—wasgenerated,andtheendocranialvolume(ECV)wasmeasuredasthevolumeenclosedbythissurface.
Atotalof26anatomicallandmarks(TableS2)and373semilandmarksalongcurvesandsurfacesweredigitizedoneachendocast(Fig.
1AandFig.
S6).
Todigitizethesurfacesemilandmarks,ameshof268roughlyequidis-tantpointswasgeneratedautomaticallyononespecimenandthenpro-jectedontoeveryotherendocast,usingthelandmarksandcurvesemilandmarksasreferencepointsforalignmentandathin-platesplineinterpolation(36),asimplementedinthegeomorphpackageforR(37).
Toremovenonshapevariation,ageneralizedProcrustesanalysis(GPA)wasperformed.
GPAalsoreturnsthecentroidsize(CS)ofeachconfiguration,ameasureofthesizeofthestructure.
Additionally,semilandmarkswereallowedtoslidealongtangentsandtangentplanestothecurvesandsur-faces,respectively,minimizingthebendingenergydistancebetweensemi-landmarksineachspecimenandtheconsensusconfiguration(38).
TheresultingProcrustesshapecoordinateswereextracted.
VariabilityinbrainshapewascharacterizedbymeansofaPCAofthespecies-averagedProcrustescoordinates(DatasetS1).
Thisreducesthedata-setdimensionalityandgeneratesuncorrelatedaxes(PCs)describingthemaintrendsinshapevariationamongspecies.
Subsequently,PCscoreswereusedinthefollowinganalysesasbrainshapevariables.
Wemeasuredencephalization(orrelativebrainsize)astheresidualvalues(rECV)fromaphylogeneticregressionofECVonbasicranialCS(21).
En-cephalizationhasbeenmeasuredinmanydifferentwaysinprimatestudies,andthechoicemainlydependsonthequestionbeingaskedasallpresentstrengthsandweaknesses(39).
Ourmeasureisconceptuallydifferentfromthatofotherplatyrrhinestudies,whichmainlycorrectedforbodymass(e.
g.
,refs.
19and20)butissimilartothatusedbyotherworkers(e.
g.
,refs.
40and41).
Usingthismeasurement,wehaveamorestructuralencephalizationdefinition,avoidingtheuseofbodymassdata,whichcanbesubjectedtodifferentselectivepressurestothatexperiencedbybrainsize,affectingevolutionaryinterpretations(42).
ComparativeAnalyses.
Thephylogenyfortheplatyrrhinespecieswasobtainedfromthefossil-calibratedBayesianmoleculartreeinAristideetal.
(18)andwasprunedtomatchthespeciesinourmorphometricdataset(DatasetS2).
PGLSswereusedtoassesstheassociationbetweenthevariousanalyzedvariables.
PGLSaccountsfortheexpectedlackofindependenceamongsamplesarisingfromphylogeneticstructure(43)bymodelingresidualvari-ationassumingBrownianevolution.
Thisassumptioncanbeamelioratedbyincorporatinganadditionalparametertotheregression,λ[rangingfrom0to1andestimatedbymaximumlikelihood(44)],whichmeasuresthephylogeneticsignalintheregressionresiduals.
Convergencewasassessedusingtwopattern-basedteststatistics(C1andC5),whichquantifyindependentlyevolvedsimilaritywithoutmakingas-sumptionsabouttheprocessesgeneratingthissimilarity(45).
Thisway,in-vestigationoftheprocessesdrivingconvergentevolutioncanbeseparatedfromitsidentification.
C1quantifiesthedegreetowhichtheputativelyconvergenttaxahaveevolvedtobemoresimilarthantheirreconstructedancestorswere.
C5simplycountsthenumberoflineagesenteringtheregionofthemorphospaceoccupiedbythehypothesizedconvergenttaxa.
Bothmeasureswerecomparedagainstanullexpectationgeneratedusing999BMtraitsimulationsalongthephylogeny(45).
PhylogeneticsignalintraitdatawereestimatedbyBlomberg'sKstatistic(25).
Kmeasurestheassociationbetweenthepatternofvariationinthedataandthestructureofthephylogenetictree,consideringaBMmodelofevolutionasanullexpectation.
ValuesofKnear1indicateastrongphy-logeneticsignal,whereasvaluesnear0indicateadecouplingofphyloge-neticandphenotypicdivergence.
Kvaluesabove1indicatethatspeciesaremoresimilarthanexpectedunderBM,apatternthatmaybeassociatedwithearlyniche-fillingscenarios(46).
DTTplots(47)weregeneratedintheGEIGERpackageforR(48)toex-aminethetemporalpatternofchangeinrelativephenotypicdisparityalong2162|www.
pnas.
org/cgi/doi/10.
1073/pnas.
1514473113Aristideetal.
DownloadedbyguestonDecember31,2020theplatyrrhinephylogeny.
DTTanalysesallowcomparingtheobservedpatternofintracladeversusamong-cladesdisparitythroughtimewithanexpectationunderBM.
Highrelativedisparityvaluesareindicativeofex-tensivewithin-cladediversificationandamong-cladesoverlap,whereasvaluesnear0suggestthatvariationismainlypartitionedamongclades(47).
ModelselectionanalyseswereperformedwiththemvMORPHpackageforR(49),whichallowsfittingseveralevolutionarymodelstotraitdataandaphylogenyinamultivariateframework.
Foreachmodel,therelativefitwasassessedusingthesamplesizeAICc(50).
Severalmodelswereevaluated,withBMevolutionasthesimplest.
Morecomplexmodelsincludedearlyburst(51),weretheratesofBrownianevolutiondecaysexponentiallywithtime,mimickingniche-fillingscenarios;andseveraladaptiveOUmodels(Fig.
S4).
TheSURFACEmethod(52)wasusedtoexploretheOUmodelspacetocorroboratethatourbest-fittingbiologicalhypothesisisnotmarkedlydif-ferentfromthebestpossiblestatisticalhypothesis(SupportingInformationandFig.
S5).
OUmodelsdescribeprocesseswheretraitsvaluesareconstrainedaroundoneorseveraloptimathatcanbeconsideredadaptivepeaksinamacroevolutionarylandscape.
ThesemodelsincorporatetothestochasticBMmodeladeterministictermrepresentingtheactionofselectiontowardanop-timumvalue,andconstitutethesimplestmathematicalexpressionofanevo-lutionarymodelwithselection(53).
EachOUmodelisconstructedbyassigninghypothesizedadaptiveregimestoeachbranchandnodeofthephylogenyfollowingalternativeevolutionaryscenariosforthetraitsunderstudy.
ACKNOWLEDGMENTS.
WethankJ.
deOliveira(MuseuNacionaldoRiodeJaneiro),D.
Flores(MuseoArgentinodeCienciasNaturales),M.
deVivo(MuseudeZoologiadaUniversidadedeSoPaulo),andtheDigitalMorphologyMu-seum(KyotoUniversity)forgrantingusaccesstothecollectionsundertheircare.
WearegratefultoA.
Rosenbergerandtwoanonymousreviewersforusefulcomments.
ThisresearchwassupportedbygrantsfromtheFondoparalaInves-tigaciónCientíficayTecnológica,ConselhoNacionaldeDesenvolvimentoCientí-ficoeTecnológico,andFundaodeAmparoàPesquisadoEstadodeSoPaulo.
1.
FutuymaDJ(1998)EvolutionaryBiology(SinauerAssociates,Sunderland,MA).
2.
GavriletsS,LososJB(2009)Adaptiveradiation:Contrastingtheorywithdata.
Science323(5915):732–737.
3.
SimpsonGG(1953)TheMajorFeaturesofEvolution(ColumbiaUnivPress,NewYork).
4.
LososJB(2010)Adaptiveradiation,ecologicalopportunity,andevolutionaryde-terminism.
AmericanSocietyofNaturalistsE.
O.
WilsonAwardAddress.
AmNat175(6):623–639.
5.
JnssonKA,etal.
(2012)EcologicalandevolutionarydeterminantsfortheadaptiveradiationoftheMadagascanvangas.
ProcNatlAcadSciUSA109(17):6620–6625.
6.
LososJB(2009)LizardsinanEvolutionaryTree:EcologyandAdaptiveRadiationofAnoles(UnivofCaliforniaPress,Berkeley,CA).
7.
SoonsJ,etal.
(2015)IsbeakmorphologyinDarwin'sfinchestunedtoloadingde-mandsPLoSOne10(6):e0129479.
8.
MuschickM,IndermaurA,SalzburgerW(2012)Convergentevolutionwithinanadaptiveradiationofcichlidfishes.
CurrBiol22(24):2362–2368.
9.
BartonRA(2006)Primatebrainevolution:Integratingcomparative,neurophysio-logicalandethologicaldata.
EvolAnthropol15(6):224–236.
10.
FleagleJG(2013)PrimateAdaptationandEvolution(Academic,SanDiego).
11.
DunbarRIM(1998)Thesocialbrainhypothesis.
EvolAnthropol6(5):178–190.
12.
BartonRA,HarveyPH(2000)Mosaicevolutionofbrainstructureinmammals.
Nature405(6790):1055–1058.
13.
deWinterW,OxnardCE(2001)Evolutionaryradiationsandconvergencesinthestructuralorganizationofmammalianbrains.
Nature409(6821):710–714.
14.
SmaersJB,SoligoC(2013)Brainreorganization,notrelativebrainsize,primarilycharacterizesanthropoidbrainevolution.
ProcBiolSci280(1759):20130269.
15.
HuberR,vanStaadenMJ,KaufmanLS,LiemKF(1997)Microhabitatuse,trophicpatterns,andtheevolutionofbrainstructureinAfricancichlids.
BrainBehavEvol50(3):167–182.
16.
DunbarRIM,ShultzS(2007)Evolutioninthesocialbrain.
Science317(5843):1344–1347.
17.
RosenbergerAL(1992)EvolutionoffeedingnichesinNewWorldmonkeys.
AmJPhysAnthropol88(4):525–562.
18.
AristideL,RosenbergerAL,TejedorMF,PerezSI(2015)Modelinglineageandphe-notypicdiversificationintheNewWorldmonkey(Platyrrhini,Primates)radiation.
MolPhylogenetEvol82(PtB):375–385.
19.
HartwigW,RosenbergerAL,NorconkMA,OwlMY(2011)Relativebrainsize,gutsize,andevolutioninNewWorldmonkeys.
AnatRec(Hoboken)294(12):2207–2221.
20.
AllenKL,KayRF(2012)Dietaryqualityandencephalizationinplatyrrhineprimates.
ProcBiolSci279(1729):715–721.
21.
AristideL,etal.
(2015)Encephalizationanddiversificationofthecranialbaseinplatyrrhineprimates.
JHumEvol81:29–40.
22.
SchluterD(2000)TheEcologyofAdaptiveRadiation(OxfordUnivPress,Oxford).
23.
HealySD,RoweC(2007)Acritiqueofcomparativestudiesofbrainsize.
ProcBiolSci274(1609):453–464.
24.
PerezSI,KlaczkoJ,RocattiG,DosReisSF(2011)Patternsofcranialshapedi-versificationduringthephylogeneticbranchingprocessofNewWorldmonkeys(Primates:Platyrrhini).
JEvolBiol24(8):1826–1835.
25.
BlombergSP,GarlandT,Jr,IvesAR(2003)Testingforphylogeneticsignalincom-parativedata:Behavioraltraitsaremorelabile.
Evolution57(4):717–745.
26.
MahlerDL,IngramT,RevellLJ,LososJB(2013)Exceptionalconvergenceonthemacroevolutionarylandscapeinislandlizardradiations.
Science341(6143):292–295.
27.
MoenDS,IrschickDJ,WiensJJ(2013)Evolutionaryconservatismandconvergencebothleadtostrikingsimilarityinecology,morphologyandperformanceacrosscontinentsinfrogs.
ProcBiolSci280(1773):20132156.
28.
GrundlerMC,RaboskyDL(2014)Trophicdivergencedespitemorphologicalconver-genceinacontinentalradiationofsnakes.
ProcBiolSci281(1787):20140413.
29.
SchefferM,vanNesEH(2006)Self-organizedsimilarity,theevolutionaryemergenceofgroupsofsimilarspecies.
ProcNatlAcadSciUSA103(16):6230–6235.
30.
StreelmanJT,DanleyPD(2003)Thestagesofvertebrateevolutionaryradiation.
TrendsEcolEvol18:126–131.
31.
Herculano-HouzelS,CollinsCE,WongP,KaasJH(2007)Cellularscalingrulesforprimatebrains.
ProcNatlAcadSciUSA104(9):3562–3567.
32.
DunbarRIM(1992)Neocortexsizeasaconstraintongroupsizeinprimates.
JHumEvol22(6):469–493.
33.
GrovesCP(2001)PrimateTaxonomy(SmithsonianInstitutionPress,Washington,DC).
34.
InternationalUnionforConservationofNature(2014)IUCNRedListofThreatenedSpecies,Version2015.
2.
Availableatwww.
iucnredlist.
org.
AccessedMarch17,2015.
35.
NeubauerS,GunzP,HublinJ-J(2009)Thepatternofendocranialontogeneticshapechangesinhumans.
JAnat215(3):240–255.
36.
MitteroeckerP,GunzP(2009)Advancesingeometricmorphometrics.
EvolBiol36:235–247.
37.
AdamsDC,Otarola-CastilloE(2013)geomorph:AnRpackageforthecollectionandanalysisofgeometricmorphometricshapedata.
MethodsEcolEvol4(4):393–399.
38.
GunzP,MitteroeckerP,BooksteinFL(2005)Semilandmarksinthreedimensions.
ModernMorphometricsinPhysicalAnthropology,edSliceDE(KluwerAcademic/PlenumPublishers,NewYork),pp73–98.
39.
LefebvreL(2012)Primateencephalization.
ProgressinBrainResearch,edsHofmanMA,FalkD(Elsevier,Amsterdam),Vol195,pp393–412.
40.
GouldSJ(1977)OntogenyandPhylogeny(HarvardUnivPress,Cambridge,MA).
41.
RossCF,RavosaMJ(1993)Basicranialflexion,relativebrainsize,andfacialkyphosisinnonhumanprimates.
AmJPhysAnthropol91(3):305–324.
42.
SmaersJB,DechmannDK,GoswamiA,SoligoC,SafiK(2012)Comparativeanalysesofevolutionaryratesrevealdifferentpathwaystoencephalizationinbats,carnivorans,andprimates.
ProcNatlAcadSciUSA109(44):18006–18011.
43.
MartinsEP,HansenTF(1997)Phylogeniesandthecomparativemethod:Ageneralapproachtoincorporatingphylogeneticinformationintotheanalysisofinterspecificdata.
AmNat149(4):646–667.
44.
FreckletonRP,HarveyPH,PagelM(2002)Phylogeneticanalysisandcomparativedata:Atestandreviewofevidence.
AmNat160(6):712–726.
45.
StaytonCT(2015)Thedefinition,recognition,andinterpretationofconvergentevolution,andtwonewmeasuresforquantifyingandassessingthesignificanceofconvergence.
Evolution69(8):2140–2153.
46.
RevellLJ,HarmonLJ,CollarDC(2008)Phylogeneticsignal,evolutionaryprocess,andrate.
SystBiol57(4):591–601.
47.
HarmonLJ,SchulteJA,2nd,LarsonA,LososJB(2003)Tempoandmodeofevolu-tionaryradiationiniguanianlizards.
Science301(5635):961–964.
48.
HarmonLJ,WeirJT,BrockCD,GlorRE,ChallengerW(2008)GEIGER:Investigatingevolutionaryradiations.
Bioinformatics24(1):129–131.
49.
ClavelJ,EscarguelG,MerceronG(2015)mvMORPH:AnRpackageforfittingmulti-variateevolutionarymodelstomorphometricdata.
MethodsEcolEvol6(11):1311–1319.
50.
BurnhamKP,AndersonDR(2002)ModelSelectionandMultimodelInference:APracticalInformation-TheoreticApproach(Springer,NewYork).
51.
HarmonLJ,etal.
(2010)Earlyburstsofbodysizeandshapeevolutionarerareincomparativedata.
Evolution64(8):2385–2396.
52.
IngramT,MahlerDL(2013)SURFACE:Detectingconvergentevolutionfromcom-parativedatabyfittingOrnstein–UhlenbeckmodelswithstepwiseAkaikein-formationcriterion.
MethodsEcolEvol4(5):416–425.
53.
ButlerMA,KingAA(2004)Phylogeneticcomparativeanalysis:Amodelingapproachforadaptiveevolution.
AmNat164:683–695.
54.
CresslerCE,ButlerMA,KingAA(2015)DetectingadaptiveevolutioninphylogeneticcomparativeanalysisusingtheOrnstein–Uhlenbeckmodel.
SystBiol64(6):953–968.
55.
BondM,etal.
(2015)EoceneprimatesofSouthAmericaandtheAfricanoriginsofNewWorldmonkeys.
Nature520(7548):538–541.
56.
KayRF,etal.
(2012)PaleobiologyofSantacrucianprimates.
EarlyMiocenePaleobiologyinPatagonia:High-LatitudePaleocommunitiesoftheSantaCruzFormation,edsVizcaínoS,KayRF,BargoM(CambridgeUnivPress,Cambridge,UK),pp306–330.
57.
RevellLJ(2013)Twonewgraphicalmethodsformappingtraitevolutiononphy-logenies.
MethodsEcolEvol4:754–759.
58.
SantanaSE,AlfaroJL,AlfaroME(2012)AdaptiveevolutionoffacialcolourpatternsinNeotropicalprimates.
ProcBiolSci279(1736):2204–2211.
Aristideetal.
PNAS|February23,2016|vol.
113|no.
8|2163EVOLUTIONDownloadedbyguestonDecember31,2020