Return to BCS Main Page.
These two modeling papers illustrate a theoretical model of the mechanisms underlying high-dimensional decorrelation of odor stimuli in the olfactory bulb. The first (2006) is the more important for journal club purposes.
During the discussion, Patrick Gill brought up an additional related paper:
Today several people asked me to send out a 2005 paper (Fusi, Drew, Abbott 2005, Cascade models of synaptically stored memories_)_ showing why a variety of memory maintenance mechanisms with different timescales are better than having just one or two simple memory maintenance mechanisms. Here it is._ There's also a followup paper showing why it didn't matter that they used synapses with binary weighting in the 2005 paper:_ Stefano Fusi & L F Abbott "Limits on the memory storage capacity of bounded synapses" Nat Neuro 10 (4) April 2007 p 485.
For Spring semester 2009-2010, the Behavioral, Computational, and Systems Neuroscience (BCS) Journal Club will meet on Tuesdays from 12:00 to 1:15 pm in Uris Hall 205.
The overarching theme this semester is Oscillations. Adhering to this theme is not required, but recommended. Please interpret it broadly. Theories of gamma, beta, and theta oscillations in the nervous system are the centroid of intent, but alpha, sleep, circadian, etc. rhythms are also spot-on as are reasonably accessible dynamical systems topics from math and engineering.
To add yourself to the BCS-L mailing list, send a plain-text email to bcs-L-request@cornell.edu with the body of the message saying simply join. The subject line doesn't matter. Sending the message leave instead will unsubscribe you from the list. See Cornell's Lyris HowTo page for further details.
Please contact Thomas Cleland with any questions (David Smith is on walkabout this semester). BCS meeting archive.
In discussion, Patrick brought up Li et al (2004), a paper from Mu-ming Poo's lab showing that those presynaptic neurons that spike early, and that consequently (via STDP) have their synaptic weights upon a postsynaptic neuron strengthened, are also themselves rendered more excitable. That is, by this mechanism STDP can "work on the presynaptic neuron" as well as on its output synapse. The net effect of this can be to durably group a set of early-firing neurons together into a fully synchronous ensemble evoking activity in that postsynaptic cell.
In discussion, Guoshi brought up this modeling paper analyzing the property that visual input (lip reading) facilitates auditory input most effectively under moderate noise conditions.
NOTE: It's coming time to start thinking of next semester's overarching theme. One possibility is "Synaptic plasticity", a broad topic that could include molecular mechanisms as well as population-level patterns of perceptual learning, LTP as well as neuromodulator-regulated changes. As a BCS topic, of course, the intent would be to choose approaches relevant to behavioral, computational, and systems-level questions. The relevance/necessity of bidirectional regulation of plasticity would be a prime topic. Following up on synchrony-dependent timing properties of synaptic plasticity mechanisms would establish a common thread with this semester's theme. Another, somewhat related theme is a neuroscience version of statistical learning: anything from perceptual learning to Bayesian representations of coding to temporal difference learning to dopamine (Schultz model) to the underlying synaptic rules that give rise to relevant population level learning properties to optimality of the Bayesian brain. Thoughts? other ideas? Send them to Thom or bring them up at BCS.
Another idea: Mechanisms of memory consolidation and reconsolidation -- perhaps a more focused version of "synaptic plasticity" as above. These topics are much more well understood and diverse than they were even a few years ago, and they are leading to a number of exciting hypotheses about systems and behavioral integrative mechanisms. For example, see Nader & Einarsson (2010) Ann NY Acad Sci 1191:27-41, as well as Jonathan L.C. Lee's recent Nature Neuroscience paper (2008) and Trends in Neurosciences opinion (2010).
For Fall Semester 2010-2011, the Behavioral, Computational, and Systems Neuroscience (BCS) Journal Club will meet on Tuesdays from 11:45 to 1:00 pm in Uris Hall 205.
The overarching theme this semester is Systems of neuronal representation and learning. Adhering to this theme is not required, but is strongly recommended. Please interpret it broadly. It is intended to include such diverse topics as: the systematic regulation of synaptic plasticity, Bayesian representations (including sensory representations as probability estimates), Bayesian and/or energetic optimality in neural encoding or transmission, perceptual learning, decision-making (including reward harvesting), temporal difference learning/dopamine (Schultz model), synaptic rules that give rise to systems-level learning properties.
\[OPTIONAL\] [Issa EB, Wang X (2008)|BCS Journal Club^J Neurosci 2008 Issa.pdf] Sensory responses during sleep in primate primary and secondary auditory cortex. _Journal of Neuroscience_ 28(53):14467-14480. |
[Shouval HZ, Wang SS-H, Wittenberg GM (2010)|BCS Journal Club^STDP-A consequence of more fundamental rule.pdf] Spike timing dependent plasticity: a consequence of more fundamental learning rules _Frontiers in Computational Neuroscience_ 4:19. \[This is a review article\] |
\[OPTIONAL - BACKGROUND\] [Hikosaka O, Sesack SR, Lecourtier L, Shepard PD (2008)|BCS Journal Club^Habenula crossroad between the basal ganglia and the limbic system.pdf]. Habenula: crossroad between the basal ganglia and the limbic system. _Journal of Neuroscience_ 28(46): 11825-11829. |
For Spring Semester 2010-2011, the Behavioral, Computational, and Systems Neuroscience (BCS) Journal Club will meet on Tuesdays from 11:45 to 1:00 pm in Uris Hall 205.
Papers and notes from previous semesters can be found in the BCS meeting archive.
The overarching theme this semester is Cell assemblies, functional connectivity, and feedback. Please interpret it broadly. It is intended to include such diverse topics as: cortical microcircuits, the dynamic reconfiguration of neural circuitry, decision making by such circuitry, the interactions between complex circuits and neuromodulatory nuclei (e.g., the basis for specificity in neuromodulatory effects), and the relationship between specific neuromodulators and nominal functions such as "attention" or "arousal".
To add yourself to the BCS-L mailing list, send a plain-text email to bcs-L-request@cornell.edu with the body of the message saying simply join. The subject line doesn't matter. Sending the message leave instead will unsubscribe you from the list. See Cornell's Lyris HowTo page for further details.
You can enroll in the BCS Journal Club for graduate or undergraduate credit (2 CR, S/U) as a Topics in Biopsychology seminar: PSYCH 6271-102. The course requires that you present at least once during the semester and participate actively overall.
Please contact Thomas Cleland or David Smith with any questions.
This is a tough one folks, so be extra sure to read it ahead of time.
Followup papers and URL of potential interest:
An interesting exploration of the functional connectivity between two memory systems.
For Fall Semester 2011-2012, the Behavioral, Computational, and Systems Neuroscience (BCS) Journal Club will meet on Tuesdays from 11:45 to 1:00 pm in Uris Hall 205.
The overarching theme this semester is yet to be determined, but our likely first speaker will discuss his own recent work relating to neural representations. Please interpret BCS themes broadly -- they are meant to focus rather than to exclude.
Starting in Fall 2011, BCS will try out a "minimal Powerpoint" policy. In order to make discussions more engaging and less formal, we encourage presentations to be primarily "chalk talks", in which concepts are sketched rather than figures shown. Mixed media are OK too, in which a complex figure can be put onto a slide or simply zoomed up on from the PDF file of the original paper, but drawing the figure tends to convey stronger understanding than does flashing a figure up on the wall. We also emphasize that you do not have to present papers in their entirety, much less multiple papers. Having everybody read up thoroughly on something small and focused usually makes for a better experience than everybody skimming one or more full papers. You may want to present only one exciting concept, exemplified by one or more figures drawn from one or more papers. That's great. Focus on the concepts, and don't feel compelled to master every detail of every paper that you want to include in your presentation. Do what you feel is best, but please do not just put the figures of a paper into a slide show and describe the paper.
That said, presenting your own work is always welcome, and in this case often it will be in Powerpoint format and formally organized. Not a problem.
To add yourself to the BCS-L mailing list, send a plain-text email to bcs-L-request@cornell.edu with the body of the message saying simply join. The subject line doesn't matter. Sending the message leave instead will unsubscribe you from the list. See Cornell's Lyris HowTo page for further details.
You can enroll in the BCS Journal Club for graduate or undergraduate credit (2 CR, S/U) as a Topics in Biopsychology seminar: PSYCH 6271-102. The course requires that you present at least once during the semester and participate actively overall.
Please contact Thomas Cleland or David Smith with any questions.
For additional background, if desired:
For Spring Semester 2011-2012, the Behavioral, Computational, and Systems Neuroscience (BCS) Journal Club will meet on Tuesdays from 11:45 to 1:00 pm in Uris Hall 205.
Papers and notes from previous semesters can be found in the BCS meeting archive|../../../../../../../../../../display/BCS/Past+BCS+Meetings||\.
The overarching theme this semester is likely to be "Active Sensation", perhaps to be modified at our first meeting. Please interpret BCS themes broadly -- they are meant to focus rather than to exclude.
BCS will continue its "minimal Powerpoint" policy of Fall 2011. In order to make discussions more engaging and less formal, we encourage presentations to be primarily "chalk talks", in which concepts are sketched rather than figures shown. Mixed media are OK too, in which a complex figure can be put onto a slide or simply zoomed up on from the PDF file of the original paper, but drawing the figure tends to convey stronger understanding than does flashing a figure up on the wall. We also emphasize that you do not have to present papers in their entirety, much less multiple papers. Having everybody read up thoroughly on something small and focused usually makes for a better experience than everybody skimming one or more full papers. You may want to present only one exciting concept, exemplified by one or more figures drawn from one or more papers. That's great. Focus on the concepts, and don't feel compelled to master every detail of every paper that you want to include in your presentation. Do what you feel is best, but please do not just put the figures of a paper into a slide show and describe the paper.
That said, presenting your own work is always welcome, and in this case often it will be in Powerpoint format and formally organized. Not a problem.
To add yourself to the BCS-L mailing list, send a plain-text email to bcs-L-request@cornell.edu with the body of the message saying simply join. The subject line doesn't matter. Sending the message leave instead will unsubscribe you from the list. See Cornell's Lyris HowTo page for further details.
You can enroll in the BCS Journal Club for graduate or undergraduate credit (1-2 CR, S/U) as a Topics in Biopsychology seminar: PSYCH 6271-101 (class number 8699). The course requires that you present at least once during the semester and participate actively overall. You are welcome to attend without enrolling, of course, but we do appreciate you enrolling if you plan to attend the whole semester and to present.
Please contact Thomas Cleland or David Smith with any questions.
For Fall Semester 2012-2013, the Behavioral, Computational, and Systems Neuroscience (BCS) Journal Club will meet on Tuesdays from 11:45 to 1:00 pm in Uris Hall 205.
Papers and notes from previous semesters can be found in the BCS meeting archive.
The semester's theme is assessing and comparing methods of inference regarding neural activity*. * How do local field potential (LFP) data relate to data from studies based on spike trains, BOLD, fMRI, calcium or voltage-dependent optical signals, immediate-early gene histology, or other measures of neural activity? What artifacts may arise from the use of particular techniques and do they threaten the supposed findings of a given paper? This has particular importance for relatively complex or new techniques -- e.g., studies of dynamical interregional communication in the brain, frequency-domain analyses, and other techniques that may be easily misinterpreted or overtrusted. As always, please interpret BCS themes broadly -- they are meant to focus rather than to exclude.
We are considering decision making and 'neuroeconomics' as a possible organizing theme for Spring 2013.
---
BCS will continue its "minimal Powerpoint" policy of Fall 2011. In order to make discussions more engaging and less formal, we encourage presentations to be primarily "chalk talks", in which concepts are sketched rather than figures shown. Mixed media are OK too, in which a complex figure can be put onto a slide or simply zoomed up on from the PDF file of the original paper, but drawing the figure tends to convey stronger understanding than does flashing a figure up on the wall. We also emphasize that you do not have to present papers in their entirety, much less multiple papers. Having everybody read up thoroughly on something small and focused usually makes for a better experience than everybody skimming one or more full papers. You may want to present only one exciting concept, exemplified by one or more figures drawn from one or more papers. That's great. Focus on the concepts, and don't feel compelled to master every detail of every paper that you want to include in your presentation. Do what you feel is best, but please do not just put the figures of a paper into a slide show and describe the paper.
That said, presenting your own work is always welcome, and in this case often it will be in Powerpoint format and formally organized. Not a problem.
To add yourself to the BCS-L mailing list, send a plain-text email to bcs-L-request@cornell.edu with the body of the message saying simply join. The subject line doesn't matter. Sending the message leave instead will unsubscribe you from the list. See Cornell's Lyris HowTo page for further details.
You can enroll in the BCS Journal Club for graduate or undergraduate credit (1-2 CR, S/U) as a Topics in Biopsychology seminar: PSYCH 6271-101 (class number 8699). The course requires that you present at least once during the semester and participate actively overall. You are welcome to attend without enrolling, of course, but we do appreciate you enrolling if you plan to attend the whole semester and to present.
Please contact Thomas Cleland or David Smith with any questions.
Bonus content based on discussion:
More bonus content based on discussion -- the 2011 point/counterpoint articles in Neuron on the putative role of neurogenesis in pattern separation in olfactory bulb:
For Spring Semester 2012-2013, the Behavioral, Computational, and Systems Neuroscience (BCS) Journal Club will meet on Tuesdays from 11:45 to 1:00 pm in Uris Hall 205.
Papers and notes from previous semesters can be found in the BCS meeting archive.
The semester's theme is Neural representations: are they? what are they? and how are they formed? This can be approached from a mechanistic level (what neuronal mechanisms underlie the metric(s) by which neural representations are formed), a psychological level (generalization, discrimination, psychophysical evidence), a systems neuroscience level (coordinated cortical learning systems), or other levels TBD. Is the concept accurate? is it useful? misleading? Sensory inputs and learning change neural activity and mediate our experience - is this process well described by the concept of representations? As always, please interpret BCS themes broadly -- they are meant to focus rather than to exclude.
---
BCS will continue its "minimal Powerpoint" policy, in place since Fall 2011. In order to make discussions more engaging and less formal, we encourage presentations to be primarily "chalk talks", in which concepts are sketched rather than figures shown. Mixed media are OK too, in which a complex figure can be put onto a slide or simply zoomed up on from the PDF file of the original paper, but drawing the figure tends to convey stronger understanding than does flashing a figure up on the wall. We also emphasize that you do not have to present papers in their entirety, much less multiple papers. Having everybody read up thoroughly on something small and focused usually makes for a better experience than everybody skimming one or more full papers. You may want to present only one exciting concept, exemplified by one or more figures drawn from one or more papers. That's great. Focus on the concepts, and don't feel compelled to master every detail of every paper that you want to include in your presentation. Do what you feel is best, but please do not just put the figures of a paper into a slide show and describe the paper.
That said, presenting your own work is always welcome, and in this case often it will be in Powerpoint format and formally organized. Not a problem.
To add yourself to the BCS-L mailing list, send a plain-text email to bcs-L-request@cornell.edu with the body of the message saying simply join. The subject line doesn't matter. Sending the message leave instead will unsubscribe you from the list. See Cornell's Lyris HowTo page for further details.
You can enroll in the BCS Journal Club for graduate or undergraduate credit (1-2 CR, S/U) as a Topics in Biopsychology seminar: PSYCH 6271. The course requires that you present at least once during the semester and participate actively overall. You are welcome to attend without enrolling, of course, but we do appreciate you enrolling if you plan to attend the whole semester and to present.
Please contact Thomas Cleland or David Smith with any questions.