Synaptic scaling balances learning in a spiking model of neocortex (Rowan & Neymotin 2013)

 Download zip file   Auto-launch 
Help downloading and running models
Accession:147141
Learning in the brain requires complementary mechanisms: potentiation and activity-dependent homeostatic scaling. We introduce synaptic scaling to a biologically-realistic spiking model of neocortex which can learn changes in oscillatory rhythms using STDP, and show that scaling is necessary to balance both positive and negative changes in input from potentiation and atrophy. We discuss some of the issues that arise when considering synaptic scaling in such a model, and show that scaling regulates activity whilst allowing learning to remain unaltered.
Reference:
1 . Rowan MS, Neymotin SA (2013) Synaptic Scaling Balances Learning in a Spiking Model of Neocortex Adaptive and Natural Computing Algorithms, Tomassini M, Antonioni A, Daolio F, Buesser P, ed. pp.20
Model Information (Click on a link to find other models with that property)
Model Type: Realistic Network;
Brain Region(s)/Organism: Neocortex;
Cell Type(s): Neocortex L5/6 pyramidal GLU cell; Neocortex L2/3 pyramidal GLU cell; Neocortex V1 interneuron basket PV GABA cell; Neocortex fast spiking (FS) interneuron; Neocortex spiny stellate cell; Neocortex spiking regular (RS) neuron; Neocortex spiking low threshold (LTS) neuron; Abstract integrate-and-fire adaptive exponential (AdEx) neuron;
Channel(s):
Gap Junctions:
Receptor(s): GabaA; AMPA; NMDA;
Gene(s):
Transmitter(s): Gaba; Glutamate;
Simulation Environment: NEURON; Python;
Model Concept(s): Synaptic Plasticity; Long-term Synaptic Plasticity; Learning; STDP; Homeostasis;
Implementer(s): Lytton, William [bill.lytton at downstate.edu]; Neymotin, Sam [Samuel.Neymotin at nki.rfmh.org]; Rowan, Mark [m.s.rowan at cs.bham.ac.uk];
Search NeuronDB for information about:  Neocortex L5/6 pyramidal GLU cell; Neocortex L2/3 pyramidal GLU cell; Neocortex V1 interneuron basket PV GABA cell; GabaA; AMPA; NMDA; Gaba; Glutamate;
/
stdpscalingpaper
batchscripts
mod
README
alz.hoc
autotune.hoc *
basestdp.hoc *
batch.hoc *
batch2.hoc *
batchcommon
checkirreg.hoc *
clusterrun.sh
col.dot *
col.hoc *
comppowspec.hoc *
condisconcellfig.hoc *
condisconpowfig.hoc *
declist.hoc *
decmat.hoc *
decnqs.hoc *
decvec.hoc *
default.hoc *
drline.hoc *
e2hubsdisconpow.hoc *
e2incconpow.hoc *
filtutils.hoc *
geom.hoc *
graphplug.hoc *
grvec.hoc *
init.hoc *
labels.hoc *
load.hoc *
local.hoc *
makepopspikenq.hoc *
matfftpowplug.hoc *
matpmtmplug.hoc *
matpmtmsubpopplug.hoc *
matspecplug.hoc *
network.hoc *
nload.hoc *
nqpplug.hoc *
nqs.hoc *
nqsnet.hoc *
nrnoc.hoc *
params.hoc
plot.py
plotbatch.sh
plotbatchcluster.sh
powchgtest.hoc *
python.hoc *
pywrap.hoc *
redE2.hoc *
run.hoc
runsim.sh
setup.hoc *
shufmua.hoc *
sim.hoc
simctrl.hoc *
spkts.hoc *
stats.hoc *
stdpscaling.hoc
syncode.hoc *
vsampenplug.hoc *
writedata.hoc
xgetargs.hoc *
                            
// $Id: matspecplug.hoc,v 1.3 2010/09/27 17:07:56 samn Exp $ 


// "plugin" (for batch.hoc) to do analysis on sim data

binsz = 5 // bin size in ms
sampr = 1e3 / binsz // sampling rate
initAllMyNQs() // initialize counts per time, by type, column, etc.

objref nqs,nqtmp
objref vintraty[numcols][CTYPi] // each type, within column
objref vintraE[numcols]         // all Es within column
objref vintraI[numcols]         // all Is within column
objref vinterty[CTYPi]          // each type, across columns
objref vinterE,vinterI          // all Es,Is across columns
objref vty

{vty=new Vector() vty.append(E2,I2)}

sz=nqCTY[0].v[E2].size

proc myrsz () { // util func to call matspecgram and add results to nqs
  {vec.resize(0) vec.copy($o1) vec.sub(vec.mean)}
  nqtmp=matspecgram(vec,sampr,sampr/2,0,0)
  nqs.append($s2,nqtmp)
  nqsdel(nqtmp)
}

obfunc mymuafcor () { local sz,i,j,a localobj v0,v1,mc
  {a=allocvecs(v0,v1) sz=$o1.v.size mc=new Matrix(sz,sz)}
  for i=0,sz-1 {
    {v0.resize(0) v0.copy($o1.get("pow",i).o)}
    for j=0,sz-1 {
      {v1.resize(0) v1.copy($o2.get("pow",j).o)}
      mc.x(i,j)=v0.pcorrel(v1)
    }
  }
  {dealloc(a) return mc}
}

{nqs=new NQS("name","nqspec") nqs.strdec("name") nqs.odec("nqspec")}

{vinterE=new Vector(sz) vinterI=new Vector(sz)}
for i=0,CTYPi-1 if(col.numc[i]) vinterty[i]=new Vector(sz)
for i = 0 , numcols - 1 { // setup all the vectors that will have matfftpow run on them
  {vintraE[i]=new Vector(sz)  vintraI[i]=new Vector(sz)}
  for j = 0, CTYPi - 1 {
    if(nqCTY[i].v[j].size>0) {
      
      vintraty[i][j]=new Vector(sz)
      vintraty[i][j].copy(nqCTY[i].v[j])
      
      vinterty[j].add(nqCTY[i].v[j])

      if(ice(j)) {
        vintraI[i].add(nqCTY[i].v[j])
        vinterI.add(nqCTY[i].v[j])
      } else {
        vintraE[i].add(nqCTY[i].v[j])
        vinterE.add(nqCTY[i].v[j])
      }
    }
  }
}

myrsz(vinterE,"interE")
myrsz(vinterI,"interI")

for i=0,CTYPi-1 if(vinterty[i]!=nil && vty.contains(i)) {
  sprint(tstr,"inter%s",CTYP.o(i).s) 
  myrsz(vinterty[i],tstr)
}

for i=0,numcols-1 {
  {sprint(tstr,"C%dintraE",i) myrsz(vintraE[i],tstr)}
  {sprint(tstr,"C%dintraI",i) myrsz(vintraI[i],tstr)}
  for j=0,CTYPi-1 if(vintraty[i][j]!=nil && vty.contains(j)) {
    if(vintraty[i][j].size>0) {sprint(tstr,"C%dintra%s",i,CTYP.o(j).s) myrsz(vintraty[i][j],tstr)}
  }
}

sprint(tstr,"/u/samn/intfcol/data/%s_nqspecg.nqs",strv)
nqs.sv(tstr)

objref nqmc,nqs1,nqs2,mc,vtmp
strdef sc1,sc2
{nqmc=new NQS("name1","name2","vpfc","sz","x01","x10") nqmc.strdec("name1") nqmc.strdec("name2") nqmc.odec("vpfc")}
for(i=0;i<nqs.v.size;i+=2) {
  {nqs1=nqs.get("nqspec",i).o  nqs2=nqs.get("nqspec",i+1).o}
  {mc=mymuafcor(nqs1,nqs1)  vtmp=mat2vec(mc) sc1=nqs.get("name",i).s}
  nqmc.append(sc1,sc1,vtmp,mc.nrow,mc.x(0,1),mc.x(1,0))
  {mc=mymuafcor(nqs1,nqs2)  vtmp=mat2vec(mc) sc2=nqs.get("name",i+1).s}
  nqmc.append(sc1,sc2,vtmp,mc.nrow,mc.x(0,1),mc.x(1,0))  
}

sprint(tstr,"/u/samn/intfcol/data/%s_nqspec_pfcor.nqs",strv)
nqmc.sv(tstr)
nqsdel(nqmc)
nqsdel(nqs)

Loading data, please wait...