m8ta
use https for features.
text: sort by
tags: modified
type: chronology
[0] Pawlak V, Kerr JN, Dopamine receptor activation is required for corticostriatal spike-timing-dependent plasticity.J Neurosci 28:10, 2435-46 (2008 Mar 5)

{1420}
hide / / print
ref: -2012 tags: octopamine STDP locust LTP LTD olfactory bulb date: 03-11-2019 18:59 gmt revision:5 [4] [3] [2] [1] [0] [head]

PMID-22278062 Conditional modulation of spike-timing-dependent plasticity for olfactory learning.

  • Looked at the synapes from the Muschroom body (Kenyon cells, sparse code) to the beta-lobe (bLN) in locusts.
  • Used in-vivo dendrite patch, sharp micropipette.
  • Found that, with a controlled mushroom body extracellular stim for plasticity induction protocol at the KC-> bLN synapese, were able to get potentiation and depression in accord with STDP.
  • This STDP became pure depression in the presence of octopamine
  • See also / supercedes: Synaptic Learning Rules and Sparse Coding in a Model Sensory System Luca A. Finelli ,Seth Haney, Maxim Bazhenov, Mark Stopfer, Terrence J. Sejnowski 2008

{1421}
hide / / print
ref: -0 tags: STDP dopamine hippocampus date: 01-16-2019 21:56 gmt revision:1 [0] [head]

PMID-26516682 Retroactive modulation of spike timing-dependent plasticity by dopamine.

  • Here we show that dopamine, a positive reinforcement signal, can retroactively convert hippocampal timing-dependent synaptic depression into potentiation.
  • This effect requires functional NMDA receptors and is mediated in part through the activation of the cAMP/PKA cascade.
  • Mouse horizontal slices.
  • Plasticity induced by 100 pairings of a single EPSP followed by a postsynaptic spike (heavy-handed?)
  • Pre-before-post @ 10ms -> LTP
  • Post-before-pre @ -20ms -> LTD
  • Post-before-pre @ -10ms -> LTP (?!)
    • Addition of Dopamine antagonist (D2: sulpiride, D1/D5: SCH23390) prevented LTP and resulted in LTD.
  • Post-before-pre @ -20ms -> LTP in the presence of 20 uM DA.
    • The presence of DA during coordinated spiking activity widense the timing interval for induction of LTP.
  • What about if it's applied afterward?
  • 20 uM DA applied 1 minute (for 10-12 minutes) after LTD induction @ -20 mS converted LTD into LTP.
    • This was corrected by addition of the DA agonists.
    • Did not work if DA was applied 10 or 30 minutes after the LTD induction.
  • Others have shown that this requires functional NMDA receptors.
    • Application of NMDA agonist D-AP5 after post-before-pre -20ms did not affect LTD.
    • Application of D-AP5 before DA partially blocked conversion of LTD to LTP.
    • Application of D-AP5 alone before induction did not affect LTD.
  • This is dependent on the cAMP/PKA signaling cascade:
    • Application of forskolin (andenylyl cyclase AC activator) converts LTD -> LTP.
    • Dependent on NMDA.
  • PKA inhibitor H-89 alsoblocked LTD -> P.

{715}
hide / / print
ref: Legenstein-2008.1 tags: Maass STDP reinforcement learning biofeedback Fetz synapse date: 04-09-2009 17:13 gmt revision:5 [4] [3] [2] [1] [0] [head]

PMID-18846203[0] A Learning Theory for Reward-Modulated Spike-Timing-Dependent Plasticity with Application to Biofeedback

  • (from abstract) The resulting learning theory predicts that even difficult credit-assignment problems, where it is very hard to tell which synaptic weights should be modified in order to increase the global reward for the system, can be solved in a self-organizing manner through reward-modulated STDP.
    • This yields an explanation for a fundamental experimental result on biofeedback in monkeys by Fetz and Baker.
  • STDP is prevalent in the cortex ; however, it requires a second signal:
    • Dopamine seems to gate STDP in corticostriatal synapses
    • ACh does the same or similar in the cortex. -- see references 8-12
  • simple learning rule they use: d/dtW ij(t)=C ij(t)D(t) d/dt W_{ij}(t) = C_{ij}(t) D(t)
  • Their notes on the Fetz/Baker experiments: "Adjacent neurons tended to change their firing rate in the same direction, but also differential changes of directions of firing rates of pairs of neurons are reported in [17] (when these differential changes were rewarded). For example, it was shown in Figure 9 of [17] (see also Figure 1 in [19]) that pairs of neurons that were separated by no more than a few hundred microns could be independently trained to increase or decrease their firing rates."
  • Their result is actually really simple - there is no 'control' or biofeedback - there is no visual or sensory input, no real computation by the network (at least for this simulation). One neuron is simply reinforced, hence it's firing rate increases.
    • Fetz & later Schimdt's work involved feedback and precise control of firing rate; this does not.
    • This also does not address the problem that their rule may allow other synapses to forget during reinforcement.
  • They do show that exact spike times can be rewarded, which is kinda interesting ... kinda.
  • Tried a pattern classification task where all of the information was in the relative spike timings.
    • Had to run the pattern through the network 1000 times. That's a bit unrealistic (?).
      • The problem with all these algorithms is that they require so many presentations for gradient descent (or similar) to work, whereas biological systems can and do learn after one or a few presentations.
  • Next tried to train neurons to classify spoken input
    • Audio stimului was processed through a cochlear model
    • Maass previously has been able to train a network to perform speaker-independent classification.
    • Neuron model does, roughly, seem to discriminate between "one" and "two"... after 2000 trials (each with a presentation of 10 of the same digit utterance). I'm still not all that impressed. Feels like gradient descent / linear regression as per the original LSM.
  • A great many derivations in the Methods section... too much to follow.
  • Should read refs:
    • PMID-16907616[1] Gradient learning in spiking neural networks by dynamic perturbation of conductances.
    • PMID-17220510[2] Solving the distal reward problem through linkage of STDP and dopamine signaling.

____References____

[0] Legenstein R, Pecevski D, Maass W, A learning theory for reward-modulated spike-timing-dependent plasticity with application to biofeedback.PLoS Comput Biol 4:10, e1000180 (2008 Oct)
[1] Fiete IR, Seung HS, Gradient learning in spiking neural networks by dynamic perturbation of conductances.Phys Rev Lett 97:4, 048104 (2006 Jul 28)
[2] Izhikevich EM, Solving the distal reward problem through linkage of STDP and dopamine signaling.Cereb Cortex 17:10, 2443-52 (2007 Oct)

{633}
hide / / print
ref: Pawlak-2008.03 tags: dopamine striatum cortex STDP plasticity NMDAR date: 10-08-2008 17:24 gmt revision:1 [0] [head]

PMID-18322089[0] Dopamine Receptor Activation Is Required for Corticostriatal Spike-Timing-Dependent Plasticity

  • Single action potentials (APs) backpropagate into the higher-order dendrites of striatal spiny projection neurons during cortically driven "up" states (Kerr and Plenz, 2004)
    • note: many 'up' states in the striatum do not contain an AP.
  • Blocking dopamine D1/D5 receptors prevented both LTD and LTP induction.
  • first paragraph has a ton of references! They note that burst spiking in cortical and striatal projection neurons is infrequent - mostly, there are single spikes - and so STDP investigations are more applicable than high frequency stimulation LTP induction.
  • tested in vitro -- para-horizontal sections into the dorsolateral striatum of young rat brain, whole-cell current clamp, GABA_A currents blocked.
  • striatal projection neurons (SPNs) have a strange mode of AP generation - their membrane potential rises for 120ms after current injection, followed by a spike. They used this and infrared differential microscopy of morphology to locate the projection neurons.
  • stimulated using extracellular current to layer 5 of the cortex or nearby white matter. kept microstim current to a minimum.
  • paired this with AP generation in the SPNs at varying time delays, both at low frequency (0.1Hz)
  • there are a few cholinergic neurons in the striatum, apparently.
  • demonstrated STDP: "synaptic strength is maximally enhanced when cortically evoked EPSPs lead a spike by 10 ms, whereas synaptic strength is maximally depressed when EPSPs follow a spike by 30 ms"
  • also tried eliciting bursts in the SPN, but: "the timing of EPSPs with single APs is as efficient in inducing synaptic plasticity as the timing of EPSPs with AP bursts"
  • the STDP / LTP / LTD was NMDA-R dependent.
  • blocked D1/D5 with SCH-23390, and found that they could not induce LTP / LTD.
  • block of D2 receptor advanced the onset of LTP and delayed the onset of LTD, to a less dramatic degree than the D1/D5 block. Long-term LTP/LTD magnitude was not effected.
  • why did these guys get in J. Neuroscience where as this is in Science? because the Science article studied medium spiny neurons, with GFP labeling the D1/D2 receptors?

____References____

{108}
hide / / print
ref: bookmark-0 tags: STDP hebbian learning dopamine reward robot model ISO date: 0-0-2007 0:0 revision:0 [head]

http://www.berndporr.me.uk/iso3_sab/

  • idea: have a gating signal for the hebbian learning.
    • pure hebbian learning is unsable; it will lead to endless amplification.
  • method: use a bunch of resonators near sub-critically dampled.
  • application: a simple 2-d robot that learns to seek food. not super interesting, but still good.
  • Uses ISO learning - Isotropic sequence order learning.
  • somewhat related: runbot!