One of the key findings supporting the assembly hypothesis was found in recordings from the primary motor cortex of behaving monkeys involved in a delayed pointing task [A. Riehle, S. Gru¨n, M. Diesmann, A. Aertsen, Spike synchronization and rate modulati



Neurocomputing70(2007)2064–2068
doc.guandang.net
Validationoftask-relatedexcessofspikecoincidences
basedonNeuroXidence
GordonPipaa,Ã,AlexaRiehleb,SonjaGrunc,d
a
FrankfurtInstituteforAdvancedStudies&MPIforBrainResearch,Max-von-Laue-Str.1,60438Frankfurt/Main,Germany
b
MediterraneanInstCognitiveNeuroscience,CNRS&Aix-MarseilleUniversities,Marseille,France
c
ComputationalNeuroscienceGroup,RIKENBrainScienceInstitute,WakoCity,Japan
d
BernsteinCenterforComputationalNeuroscience,Berlin,Germany
Availableonline16November2006
Abstract
Oneofthekey ndingssupportingtheassemblyhypothesiswasfoundinrecordingsfromtheprimarymotorcortexofbehavingmonkeysinvolvedinadelayedpointingtask[A.Riehle,S.Grun,M.Diesmann,A.Aertsen,Spikesynchronizationandratemodulationdifferentiallyinvolvedinmotorcorticalfunction,Science278(1997)1950–1953].Basedontheunitaryevent(‘UE’)method,theauthorshaveshownthatexcesscoincidencesbetweensimultaneouslyrecordedneuronsoccurdynamicallyatbehaviorallyrelevantpointsintime.However,sensitivityoftheUEmethodfornon-stationaritiesandregularityofspiketrainscausedfearthattheresultspresentedmightbe,atleastinpart,falsepositives.Wereanalyzedthesamedatawiththenewnon-parametricmethodNeuroXidence,whichisrobustagainst ring-ratemodulations,ratechangesacrosstrials,regularityorburstiness,aswellaslowrates.OurresultsbasedonNeuroXidencecon rmtheresultspresentedinRiehleetal.[Spikesynchronizationandratemodulationdifferentiallyinvolvedinmotorcorticalfunction,Science278(1997)1950–1953]anddemonstratebehaviorallymodulatedexcessesofcoincidencesonatimescaleof3–5ms.
r2006ElsevierB.V.Allrightsreserved.
Keywords:Spikesynchronization;Statisticalsigni cance;Non-parametric;Surrogatedata;NeuroXidence;Monkeymotorcortex
1.Introduction
Inrecentyears,therehasbeenagreatdealofcontroversialdiscussionabouttheassemblyhypothesis.Ontheonehand,evidencefor ne-temporalstructureinspikingactivityisaccumulating.Ontheotherhand,therearestillconcernsabouttheappropriateanalysistechniquestodemonstratetheexistenceofprecisetemporalcoordina-tionofspikingactivitiesofneuronsandtodiminishthefearoffalsepositivesduetoviolationsofinherentassumptionsoftheanalysismethod[2,5,13].Oneofthekey ndingsthatsupportstheassemblyhypothesiswasfoundinrecordingsfromtheprimarymotorcortexofbehavingmonkeysinvolvedinadelayedpointingtask[12].Itwasshownthatexcessjoint-spikeactivityofsimulta-neouslyrecordedneuronsoccursdynamicallyatbehavio-Correspondingauthor.
rallyrelevantpointsintime.Theanalysisofthedatawasbasedontheunitaryeventanalysismethod(‘UE’;[3,4]).2.UEmethod
ThecoreideaoftheUEmethodisbasedonastatisticalhypothesistestthatcomparestheamountofempiricallymeasuredcoincidenceswithinsimultaneouslyrecordedneuronswiththeexpectednumber.Thelatterisestimatedbasedonthenull-hypothesis(H0)thatassumesthattherecordedspiketrainswithintheanalysiswindowcanbedescribedbyindependentandstationaryBernoullipro-cesses[3].Tocomparetheexpectedandtheempiricalnumbersbyastatisticalsigni cancetest,theUEmethodassumestheexpectednumberstobePoissondistributed[3].Thus,aviolationofH0canhavedifferentcauses:spikesofthesimultaneouslyrecordedneuronsarecorrelated,orindividualspiketrainsviolateinherentassumptionsofthemethod.Ofcourse,thelatterhastobeavoidedbecausethis
E-mailaddress:pipa@mpih-frankfurt.mpg.de(G.Pipa).
0925-2312/$-seefrontmatterr2006ElsevierB.V.Allrightsreserved.doi:10.1016/j.neucom.2006.10.142
Nikoli? , submitted c [2] Validation of task-related excess of spike coincidences based on NeuroXidence. G. Pipa, A. Riehle and S. Gr¨ n. ...
Nikoli? , submitted c [2] Validation of task-related excess of spike coincidences based on NeuroXidence. G. Pipa, A. Riehle and S. Gr¨ n. ...
热门文档