Reinforcement Learning of Targeted Movement in a Spiking Neuronal Model of Motor Cortex 英文参考文献.docVIP

下载本文档

7
0
约3.7万字
约 8页
2017-05-13 发布于上海
举报
版权申诉

Reinforcement Learning of Targeted Movement in a Spiking Neuronal Model of Motor Cortex 英文参考文献.doc

1、原创力文档（book118）网站文档一经付费（服务费），不意味着购买了该文档的版权，仅供个人/单位学习、研究之用，不得用于商业用途，未经授权，严禁复制、发行、汇编、翻译或者网络传播等，侵权必究。。
2、本站所有内容均由合作方或网友上传，本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺！文档内容仅供研究参考，付费前请自行鉴别。如您付费，意味着您自己接受本站规则且自行承担风险，本站不退款、不进行额外附加服务；查看《如何避免下载的几个坑》。如果您已付费下载过本站文档，您可以点击这里二次下载。
3、如文档侵犯商业秘密、侵犯著作权、侵犯人身权等，请点击“版权申诉”（推荐），也可以打举报电话：400-050-0827(电话支持时间：9:00-18:30)。
4、该文档为VIP文档，如果想要下载，成为VIP会员后，下载免费。
5、成为VIP后，下载本文档将扣除1次下载权益。下载后，不支持退款、换文档。如有疑问请联系我们。
6、成为VIP后，您将拥有八大权益，权益包括：VIP文档下载权益、阅读免打扰、文档格式转换、高级专利检索、专属身份标志、高级客服、多端互通、版权登记。
7、VIP文档为合作方或网友上传，每下载1次，网站将根据用户上传文档的质量评分、类型等，对文档贡献者给予高额补贴、流量扶持。如果你也想贡献VIP文档。上传文档

Reinforcement Learning of Targeted Movement in a Spiking Neuronal Model of Motor Cortex 英文参考文献

ReinforcementLearningofTargetedMovementin aSpikingNeuronalModelofMotorCortex GeorgeL.Chadderdon1*,SamuelA.Neymotin1,2,CliffC.Kerr1,3,WilliamW.Lytton1,4,5 1DepartmentofPhysiologyandPharmacology,StateUniversityNewYorkDownstate,Brooklyn,NewYork,UnitedStatesofAmerica,2DepartmentofNeurobiology,Yale UniversitySchoolofMedicine,NewHaven,Connecticut,UnitedStatesofAmerica,3SchoolofPhysics,UniversityofSydney,Sydney,Australia,4DepartmentofNeurology, StateUniversityNewYorkDownstate,Brooklyn,NewYork,UnitedStatesofAmerica,5KingsCountyHospital,Brooklyn,NewYork,UnitedStatesofAmerica Abstract Sensorimotorcontrolhastraditionallybeenconsideredfromacontroltheoryperspective,withoutrelationtoneurobiology. Incontrast,hereweutilizedaspiking-neuronmodelofmotorcortexandtrainedittoperformasimplemovementtask, which consisted of rotating a single-joint ‘‘forearm’’ to a target. Learning was based on a reinforcement mechanism analogoustothatofthedopaminesystem.Thisprovidedaglobalrewardorpunishmentsignalinresponsetodecreasingor increasing distance from hand to target, respectively. Output was partially driven by Poisson motor babbling, creating stochasticmovementsthatcouldthenbeshapedbylearning.Thevirtualforearmconsistedofasinglesegmentrotated aroundanelbowjoint,controlledbyflexorandextensormuscles.Themodelconsistedof144excitatoryand64inhibitory event-basedneurons,eachwithAMPA,NMDA,andGABAsynapses.Proprioceptivecellinputtothismodelencodedthe2 musclelengths.Plasticitywasonlyenabledinfeedforwardconnectionsbetweeninputandoutputexcitatoryunits,using spike-timing-dependenteligibilitytracesforsynapticcreditorblameassignment.Learningresultedfromaglobal3-valued signal: reward (+1), no learning (0), or punishment (21), corresponding to phasic increases, lack of change, or phasic decreasesofdopaminergiccellfiring,respectively.Successfullearningonlyoccurredwhenbothrewardandpunishment wereenabled.Inthiscase,5targetangleswerelearnedsuccessfullywithin180sofsimulationtime,withamedianerrorof 8degrees.Motorbabbli