A gatekeeper for learning

doi:10.1126/science.abf4523

GSTDTAP > 气候变化

DOI	10.1126/science.abf4523
	A gatekeeper for learning
	Flavio Donato
	2020-12-18
发表期刊	Science
出版年	2020
英文摘要	One of the most challenging goals of neuroscience is to unravel the biological underpinnings of cognitive functions. A necessary first step is to determine where in the brain cognition arises. In the past decades, the proposal that specific cognitive functions are localized within discrete brain regions has evolved into the theory that cognition is produced by the coordinated action of interconnected neurons across multiple areas. This is the case for the ability to learn from personal experiences, which might depend on the interplay between the parahippocampal-hippocampal network and the neocortex ([ 1 ][1]). On page 1435 of this issue, Doron et al. ([ 2 ][2]) reveal that inputs from the perirhinal cortex, one of the parahippocampal areas, modulate the activity of deep-layer pyramidal neurons in the primary somatosensory cortex, a region of the cerebral cortex that processes sensation, to control learning. Learning is a fundamental function of the brain because it supports the creation of internal representations of past experience that are used to guide future behavior. From a biological perspective, learning strengthens functional connectivity within a subset of coactive neurons, such that this subset of neurons—also known as a neuronal ensemble or engram—becomes the biological substrate on which a memory is stored ([ 3 ][3]). The interplay between external cues and the engram reinstates the spatiotemporal pattern of activity present during learning and supports memory retrieval ([ 4 ][4]). It was previously shown that upon learning an association between two stimuli, hippocampal activity was necessary for the reactivation of learning-related neuronal ensembles distributed throughout the cortex ([ 4 ][4]). However, the mechanisms by which the hippocampal network influenced the emergence of a brain-wide engram during learning remained obscure. ![Figure][5] Perirhinal inputs support learning Upon learning a hippocampus-dependent associative task, perirhinal inputs might act as a gate to modulate the excitability of apical dendrites and the impact of the feedback stream on layer 5 pyramidal neurons of the primary somatosensory cortex. GRAPHIC: KELLIE HOLOSKI/ SCIENCE To address this, Doron et al. trained rodents to report the artificial microstimulation of their primary somatosensory cortex (S1) by licking for a reward and showed that learning to lick in response to the stimulus relied on the activation of both the hippocampus and perirhinal cortex. Learning correlated with changes in the activity dynamics of a subpopulation of neurons anatomically located in layer 5 (L5) of S1. Indeed, during successful learning trials, a subset of L5 neurons exhibited increased rates of activity, which was temporally organized in stereotypical patterns of closely spaced action potentials called bursts. Silencing perirhinal projections to S1 reduced both the fraction of bursting neurons and the animal's success rate in the task. This observation is in line with a report that engram neurons are more “bursty” than non-engram neurons ([ 5 ][6]), which suggests that bursting might be a physiological signature of memory-bearing cells. How could bursting contribute to learning and memory processes? One possibility is that bursting might increase the reliability of communication between neurons ([ 6 ][7]). Indeed, because many central synapses are unreliable and a single action potential often fails to produce a response, a rapid sequence of many action potentials (i.e., a burst) could facilitate synaptic transmission. In this model, bursting might endow memory-bearing ensembles with an efficient, precise, and reliable neural code that allows these neurons to transmit information to downstream readers with high efficacy. Doron et al. found that applying a high-frequency burst stimulation to individual L5 neurons in S1 after learning was sufficient to induce memory retrieval, whereas a similar number of more dispersed action potentials was not. However, the stimulated neurons did not necessarily need to be part of the learning-related ensemble, which suggests that bursting of any L5 neuron has the potential to elicit memory retrieval in expert animals but not in naïve ones. If, upon learning, bursting of any neuron can induce memory retrieval, what is special about engram neurons? Doron et al. suggest that even though memory retrieval naturally relies on the activation of engram neurons, a larger fraction of the network could be involved in the same processes when properly stimulated. This might make retrieval robust to the progressive change in neuronal representations taking place over time in the cortex ([ 7 ][8]). An alternative hypothesis could be that bursting of a non-engram neuron might reactivate the original engram, possibly through lateral connections among excitatory and inhibitory neurons. Disentangling such hypotheses awaits future investigation. Because perirhinal inputs target the superficial layers of S1, they might modulate bursting and learning by increasing activity at the apical tuft (farthest from the cell body along the ascending trunk) of L5 neuron dendrites. In support of this idea, Doron et al. found a fraction of apical dendrites exhibiting large calcium transients upon stimulus presentation after learning. Notably, disrupting apical dendritic activity during learning in rodents reduced performance. What is the role of apical dendritic activity in learning? Inputs to the neocortex are organized so that distinct information streams terminate within distinct layers ([ 8 ][9]). Feedforward streams carry information about sensory stimuli and terminate in the middle layers. Feedback streams carry information about context or predictions about the task (generated according to an internal model) and have impacts on the outer layers. Such segregation is reproduced at the level of single L5 neurons, where feedforward and feedback streams segregate within the basal and apical dendrites, respectively ([ 8 ][9]). The rules by which each information stream shapes neuronal responses might change with learning. At the network level, responses of layer 2 and layer 3 neurons align to feedback instead of feedforward streams as a result of learning ([ 9 ][10]). Doron et al. reveal that perirhinal inputs might act as a gate that modulates apical dendritic excitability during learning, and might enhance the influence of the feedback streams on the activity of L5 pyramidal neurons (see the figure). What type of information is carried by the feedback streams, and how this information is used during learning, are issues that remain unresolved. Some indications might come from work on machine learning, which postulates that during learning, higher-level information (i.e., the feedback streams) could modulate lower-level activity (i.e., the feedforward streams) toward outputs that are more consistent with the higher-level predictions ([ 10 ][11], [ 11 ][12]). The purpose of this process would be to minimize the difference (i.e., the error) between predicted and obtained outputs—a process that might be central to the inner workings of the brain, too ([ 12 ][13]). Such an algorithm, known as “backpropagation of error,” is used to train artificial neural networks ([ 10 ][11]). In the future, it will be interesting to determine whether learning-related changes like those observed by Doron et al. indeed serve the purpose of calculating errors at the level of individual L5 neuron dendrites and thereby drive learning processes in biological neural networks akin to artificial ones. 1. [↵][14]1. J. L. McClelland, 2. B. L. McNaughton, 3. R. C. O'Reilly , Psychol. Rev. 102, 419 (1995). [OpenUrl][15][CrossRef][16][PubMed][17][Web of Science][18] 2. [↵][19]1. G. Doron et al ., Science 370, eaaz3136 (2020). [OpenUrl][20][Abstract/FREE Full Text][21] 3. [↵][22]1. S. A. Josselyn, 2. S. Tonegawa , Science 367, eaaw4325 (2020). [OpenUrl][23][Abstract/FREE Full Text][24] 4. [↵][25]1. K. Z. Tanaka et al ., Neuron 84, 347 (2014). [OpenUrl][26][CrossRef][27][PubMed][28] 5. [↵][29]1. K. Z. Tanaka et al ., Science 361, 392 (2018). [OpenUrl][30][Abstract/FREE Full Text][31] 6. [↵][32]1. J. E. Lisman , Trends Neurosci. 20, 38 (1997). [OpenUrl][33][CrossRef][34][PubMed][35][Web of Science][36] 7. [↵][37]1. M. E. Rule, 2. T. O'Leary, 3. C. D. Harvey , Curr. Opin. Neurobiol. 58, 141 (2019). [OpenUrl][38][CrossRef][39] 8. [↵][40]1. M. Larkum , Trends Neurosci. 36, 141 (2013). [OpenUrl][41][CrossRef][42][PubMed][43][Web of Science][44] 9. [↵][45]1. H. Makino, 2. T. Komiyama , Nat. Neurosci. 18, 1116 (2015). [OpenUrl][46][CrossRef][47][PubMed][48] 10. [↵][49]1. Y. LeCun, 2. Y. Bengio, 3. G. Hinton , Nature 521, 436 (2015). [OpenUrl][50][CrossRef][51][PubMed][52] 11. [↵][53]1. T. P. Lillicrap, 2. A. Santoro, 3. L. Marris, 4. C. J. Akerman, 5. G. Hinton , Nat. Rev. Neurosci. 21, 335 (2020). [OpenUrl][54] 12. [↵][55]1. R. P. Rao, 2. D. H. Ballard , Nat. Neurosci. 2, 79 (1999). [OpenUrl][56][CrossRef][57][PubMed][58][Web of Science][59] Acknowledgments: We thank S. Arber for helpful input during writing. [1]: #ref-1 [2]: #ref-2 [3]: #ref-3 [4]: #ref-4 [5]: pending:yes [6]: #ref-5 [7]: #ref-6 [8]: #ref-7 [9]: #ref-8 [10]: #ref-9 [11]: #ref-10 [12]: #ref-11 [13]: #ref-12 [14]: #xref-ref-1-1 "View reference 1 in text" [15]: {openurl}?query=rft.jtitle%253DPsychological%2Breview%26rft.stitle%253DPsychol%2BRev%26rft.aulast%253DMcClelland%26rft.auinit1%253DJ.%2BL.%26rft.volume%253D102%26rft.issue%253D3%26rft.spage%253D419%26rft.epage%253D457%26rft.atitle%253DWhy%2Bthere%2Bare%2Bcomplementary%2Blearning%2Bsystems%2Bin%2Bthe%2Bhippocampus%2Band%2Bneocortex%253A%2Binsights%2Bfrom%2Bthe%2Bsuccesses%2Band%2Bfailures%2Bof%2Bconnectionist%2Bmodels%2Bof%2Blearning%2Band%2Bmemory.%26rft_id%253Dinfo%253Adoi%252F10.1037%252F0033-295X.102.3.419%26rft_id%253Dinfo%253Apmid%252F7624455%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [16]: /lookup/external-ref?access_num=10.1037/0033-295X.102.3.419&link_type=DOI [17]: /lookup/external-ref?access_num=7624455&link_type=MED&atom=%2Fsci%2F370%2F6523%2F1410.atom [18]: /lookup/external-ref?access_num=A1995RL52000001&link_type=ISI [19]: #xref-ref-2-1 "View reference 2 in text" [20]: {openurl}?query=rft.jtitle%253DScience%26rft.stitle%253DScience%26rft.aulast%253DDoron%26rft.auinit1%253DG.%26rft.volume%253D370%26rft.issue%253D6523%26rft.spage%253Deaaz3136%26rft.epage%253Deaaz3136%26rft.atitle%253DPerirhinal%2Binput%2Bto%2Bneocortical%2Blayer%2B1%2Bcontrols%2Blearning%26rft_id%253Dinfo%253Adoi%252F10.1126%252Fscience.aaz3136%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [21]: /lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6Mzoic2NpIjtzOjU6InJlc2lkIjtzOjE3OiIzNzAvNjUyMy9lYWF6MzEzNiI7czo0OiJhdG9tIjtzOjIzOiIvc2NpLzM3MC82NTIzLzE0MTAuYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9 [22]: #xref-ref-3-1 "View reference 3 in text" [23]: {openurl}?query=rft.jtitle%253DScience%26rft.stitle%253DScience%26rft.aulast%253DJosselyn%26rft.auinit1%253DS.%2BA.%26rft.volume%253D367%26rft.issue%253D6473%26rft.spage%253Deaaw4325%26rft.epage%253Deaaw4325%26rft.atitle%253DMemory%2Bengrams%253A%2BRecalling%2Bthe%2Bpast%2Band%2Bimagining%2Bthe%2Bfuture%26rft_id%253Dinfo%253Adoi%252F10.1126%252Fscience.aaw4325%26rft_id%253Dinfo%253Apmid%252F31896692%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [24]: /lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6Mzoic2NpIjtzOjU6InJlc2lkIjtzOjE3OiIzNjcvNjQ3My9lYWF3NDMyNSI7czo0OiJhdG9tIjtzOjIzOiIvc2NpLzM3MC82NTIzLzE0MTAuYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9 [25]: #xref-ref-4-1 "View reference 4 in text" [26]: {openurl}?query=rft.jtitle%253DNeuron%26rft.volume%253D84%26rft.spage%253D347%26rft_id%253Dinfo%253Adoi%252F10.1016%252Fj.neuron.2014.09.037%26rft_id%253Dinfo%253Apmid%252F25308331%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [27]: /lookup/external-ref?access_num=10.1016/j.neuron.2014.09.037&link_type=DOI [28]: /lookup/external-ref?access_num=25308331&link_type=MED&atom=%2Fsci%2F370%2F6523%2F1410.atom [29]: #xref-ref-5-1 "View reference 5 in text" [30]: {openurl}?query=rft.jtitle%253DScience%26rft.stitle%253DScience%26rft.aulast%253DTanaka%26rft.auinit1%253DK.%2BZ.%26rft.volume%253D361%26rft.issue%253D6400%26rft.spage%253D392%26rft.epage%253D397%26rft.atitle%253DThe%2Bhippocampal%2Bengram%2Bmaps%2Bexperience%2Bbut%2Bnot%2Bplace%26rft_id%253Dinfo%253Adoi%252F10.1126%252Fscience.aat5397%26rft_id%253Dinfo%253Apmid%252F30049878%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [31]: /lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6Mzoic2NpIjtzOjU6InJlc2lkIjtzOjEyOiIzNjEvNjQwMC8zOTIiO3M6NDoiYXRvbSI7czoyMzoiL3NjaS8zNzAvNjUyMy8xNDEwLmF0b20iO31zOjg6ImZyYWdtZW50IjtzOjA6IiI7fQ== [32]: #xref-ref-6-1 "View reference 6 in text" [33]: {openurl}?query=rft.jtitle%253DTrends%2Bin%2Bneurosciences%26rft.stitle%253DTrends%2BNeurosci%26rft.aulast%253DLisman%26rft.auinit1%253DJ.%2BE.%26rft.volume%253D20%26rft.issue%253D1%26rft.spage%253D38%26rft.epage%253D43%26rft.atitle%253DBursts%2Bas%2Ba%2Bunit%2Bof%2Bneural%2Binformation%253A%2Bmaking%2Bunreliable%2Bsynapses%2Breliable.%26rft_id%253Dinfo%253Adoi%252F10.1016%252FS0166-2236%252896%252910070-9%26rft_id%253Dinfo%253Apmid%252F9004418%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [34]: /lookup/external-ref?access_num=10.1016/S0166-2236(96)10070-9&link_type=DOI [35]: /lookup/external-ref?access_num=9004418&link_type=MED&atom=%2Fsci%2F370%2F6523%2F1410.atom [36]: /lookup/external-ref?access_num=A1997VZ21000010&link_type=ISI [37]: #xref-ref-7-1 "View reference 7 in text" [38]: {openurl}?query=rft.jtitle%253DCurr.%2BOpin.%2BNeurobiol.%26rft.volume%253D58%26rft.spage%253D141%26rft_id%253Dinfo%253Adoi%252F10.1016%252Fj.conb.2019.08.005%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [39]: /lookup/external-ref?access_num=10.1016/j.conb.2019.08.005&link_type=DOI [40]: #xref-ref-8-1 "View reference 8 in text" [41]: {openurl}?query=rft.jtitle%253DTrends%2BNeurosci.%26rft.volume%253D36%26rft.spage%253D141%26rft_id%253Dinfo%253Adoi%252F10.1016%252Fj.tins.2012.11.006%26rft_id%253Dinfo%253Apmid%252F23273272%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [42]: /lookup/external-ref?access_num=10.1016/j.tins.2012.11.006&link_type=DOI [43]: /lookup/external-ref?access_num=23273272&link_type=MED&atom=%2Fsci%2F370%2F6523%2F1410.atom [44]: /lookup/external-ref?access_num=000316710400001&link_type=ISI [45]: #xref-ref-9-1 "View reference 9 in text" [46]: {openurl}?query=rft.jtitle%253DNat.%2BNeurosci.%26rft.volume%253D18%26rft.spage%253D1116%26rft_id%253Dinfo%253Adoi%252F10.1038%252Fnn.4061%26rft_id%253Dinfo%253Apmid%252F26167904%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [47]: /lookup/external-ref?access_num=10.1038/nn.4061&link_type=DOI [48]: /lookup/external-ref?access_num=26167904&link_type=MED&atom=%2Fsci%2F370%2F6523%2F1410.atom [49]: #xref-ref-10-1 "View reference 10 in text" [50]: {openurl}?query=rft.jtitle%253DNature%26rft.volume%253D521%26rft.spage%253D436%26rft_id%253Dinfo%253Adoi%252F10.1038%252Fnature14539%26rft_id%253Dinfo%253Apmid%252F26017442%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [51]: /lookup/external-ref?access_num=10.1038/nature14539&link_type=DOI [52]: /lookup/external-ref?access_num=26017442&link_type=MED&atom=%2Fsci%2F370%2F6523%2F1410.atom [53]: #xref-ref-11-1 "View reference 11 in text" [54]: {openurl}?query=rft.jtitle%253DNat.%2BRev.%2BNeurosci.%26rft.volume%253D21%26rft.spage%253D335%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [55]: #xref-ref-12-1 "View reference 12 in text" [56]: {openurl}?query=rft.jtitle%253DNature%2Bneuroscience%26rft.stitle%253DNat%2BNeurosci%26rft.aulast%253DRao%26rft.auinit1%253DR.%2BP.%26rft.volume%253D2%26rft.issue%253D1%26rft.spage%253D79%26rft.epage%253D87%26rft.atitle%253DPredictive%2Bcoding%2Bin%2Bthe%2Bvisual%2Bcortex%253A%2Ba%2Bfunctional%2Binterpretation%2Bof%2Bsome%2Bextra-classical%2Breceptive-field%2Beffects.%26rft_id%253Dinfo%253Adoi%252F10.1038%252F4580%26rft_id%253Dinfo%253Apmid%252F10195184%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [57]: /lookup/external-ref?access_num=10.1038/4580&link_type=DOI [58]: /lookup/external-ref?access_num=10195184&link_type=MED&atom=%2Fsci%2F370%2F6523%2F1410.atom [59]: /lookup/external-ref?access_num=000078409300017&link_type=ISI
领域	气候变化 ; 资源环境
URL	查看原文
引用统计
文献类型	期刊论文
条目标识符	http://119.78.100.173/C666/handle/2XK7JSWQ/308348
专题	气候变化资源环境科学
推荐引用方式 GB/T 7714	Flavio Donato. A gatekeeper for learning[J]. Science,2020.
APA	Flavio Donato.(2020).A gatekeeper for learning.Science.
MLA	Flavio Donato."A gatekeeper for learning".Science (2020).