《Scalablearchitectureinmammalianbrains在哺乳动物的大脑可扩展的体系结构学习教案》由会员分享，可在线阅读，更多相关《Scalablearchitectureinmammalianbrains在哺乳动物的大脑可扩展的体系结构学习教案（42页珍藏版）》请在金锄头文库上搜索。
1、会计学1Scalablearchitectureinmammalianbrains在哺乳动物的大脑在哺乳动物的大脑(dno)可扩展的可扩展的体系结构体系结构第一页，共42页。Optical physiology andOptical physiology and synaptic plasticity: synaptic plasticity:Shari GelberShari GelberDaniel OConnorDaniel OConnorDmitry SarkisovDmitry SarkisovShy ShohamShy ShohamMegan SullivanMegan Sulli
2、vanGayle WittenbergGayle WittenbergBrain scaling andBrain scaling and evolution: evolution:Mark BurishMark BurishDamon ClarkDamon ClarkKim Hatch HarrisonKim Hatch HarrisonHao Yuan KuehHao Yuan KuehJennifer ShultzJennifer ShultzMatt WagersMatt WagersKrysta WyattKrysta WyattLaboratory of Sam WangLabor
3、atory of Sam Wang第1页/共41页第二页，共42页。Too many learning rules!Too many learning rules!Abbott and Nelson,Nat. Neurosci.第2页/共41页第三页，共42页。第3页/共41页第四页，共42页。CA3-CA1 synapse of hippocampusRich history of extracellular and single-cell recordingThe cell biology and plasticity literature is vastOne synapse per c
4、onnectionHas AMPA, NMDA, mGluR,第4页/共41页第五页，共42页。Plasticity at CA3-CA1: a separable process?Plasticity at CA3-CA1: a separable process?第5页/共41页第六页，共42页。n nIs timing between single spikes sufficient to Is timing between single spikes sufficient to describe the actual learning rule?describe the actual
5、learning rule?n nHow different are the requirements for How different are the requirements for potentiation and depression? Are these full potentiation and depression? Are these full inverses of each other?inverses of each other?n nHow might the learning rule map to behavior?How might the learning r
6、ule map to behavior?n nCan cell biology help clarify these issues Can cell biology help clarify these issues conceptually?conceptually?Questions at the CA3-CA1 synapseQuestions at the CA3-CA1 synapse第6页/共41页第七页，共42页。Nishiyama et al. 2000 Nature 408:58480 pairings at 5 Hz Cesium pipet solutionPike et
7、 al. 1999Journal of Physiology 518:57140 pairings at 5 Hz Four epochs (10 pairings each) at 10-s intervalsChristie et al. 1996Learning and Memory 3:160900 pairings at 3 HzEpochs (30-60 pairings each) at 5-s intervalsAt the CA3-CA1 synapse, pairing single APs givesLTP.no change.LTD.第7页/共41页第八页，共42页。A
8、t CA3-CA1 synapses, pairing of At CA3-CA1 synapses, pairing of single spikes fails to induce single spikes fails to induce potentiationpotentiationSingle pre+post APs100 pairings with Dt=0-7 ms (pre before post)Pairing frequency 0.1-5 Hz第8页/共41页第九页，共42页。HoweverIn Nishiyama et al. Cesium replaced Pot
9、assiumin the intracellular pipette solutionThis will-depolarize neuron -block A-type K channels in dendrite-broaden action potentials第9页/共41页第十页，共42页。PotassiumCesiumCs+K+pairings given at 5 Hz第10页/共41页第十一页，共42页。Cs+K+can bursts of APs emulate Cs+?第11页/共41页第十二页，共42页。LTP induction requires postsynaptic
10、 bursting第12页/共41页第十三页，共42页。CA1 neurons fire bursts during theta oscillation第13页/共41页第十四页，共42页。Pairings must be given at high frequencyDt第14页/共41页第十五页，共42页。Observations on the LTP learning ruleObservations on the LTP learning rulen nPotentiation requires Potentiation requires postsynaptic burstspost
11、synaptic burstsn nPairings are effective Pairings are effective at 5 Hz (near theta)at 5 Hz (near theta)does depression have these requirements?does depression have these requirements?第15页/共41页第十六页，共42页。LTD occurs under many timing conditions All pairings given at 5 Hz第16页/共41页第十七页，共42页。RequirementT
12、hetaBurstingLTP yes yesLTD no no第18页/共41页第十九页，共42页。Conclusions so farConclusions so farn nPotentiation and depression have different Potentiation and depression have different general induction requirements.general induction requirements.n nPotentiation: active exploration (causal, bursts, 5 Hz)Pote
13、ntiation: active exploration (causal, bursts, 5 Hz)n nDepression: many conditionsDepression: many conditionsn nDuring theta, the rule acts like a phase detector.During theta, the rule acts like a phase detector.n nAt other times, the rule is depression-only.At other times, the rule is depression-onl
14、y.第19页/共41页第二十页，共42页。Why not explore parameter space exhaustively?because its exhausting!第20页/共41页第二十一页，共42页。Shy ShohamUncaging in spatial patterns: High-throughput screening of neural responsesResolution: 1 Hz10 Hz10 Hz第36页/共41页第三十七页，共42页。The CA3-CA1 learning ruleThe CA3-CA1 learning rulebutdissectable?Complex第37页/共41页第三十八页，共42页。第38页/共41页第三十九页，共42页。内容(nirng)总结会计学。Abbott and Nelson,。Nat. Neurosci.。at 5-s intervals。-broaden action potentials。RequirementThetaBursting。Depression: many conditions。Resolution: 1 micron。Complex第四十二页，共42页。
