Reinforcement learning of targeted movement (Chadderdon et al. 2012)

 Download zip file   Auto-launch 
Help downloading and running models
Accession:144538
"Sensorimotor control has traditionally been considered from a control theory perspective, without relation to neurobiology. In contrast, here we utilized a spiking-neuron model of motor cortex and trained it to perform a simple movement task, which consisted of rotating a single-joint “forearm” to a target. Learning was based on a reinforcement mechanism analogous to that of the dopamine system. This provided a global reward or punishment signal in response to decreasing or increasing distance from hand to target, respectively. Output was partially driven by Poisson motor babbling, creating stochastic movements that could then be shaped by learning. The virtual forearm consisted of a single segment rotated around an elbow joint, controlled by flexor and extensor muscles. ..."
Reference:
1 . Chadderdon GL, Neymotin SA, Kerr CC, Lytton WW (2012) Reinforcement learning of targeted movement in a spiking neuronal model of motor cortex. PLoS One 7:e47251 [PubMed]
Citations  Citation Browser
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 fast spiking (FS) interneuron; Neocortex spiking regular (RS) neuron; Neocortex spiking low threshold (LTS) neuron;
Channel(s):
Gap Junctions:
Receptor(s): GabaA; AMPA; NMDA;
Gene(s):
Transmitter(s): Dopamine; Gaba; Glutamate;
Simulation Environment: NEURON;
Model Concept(s): Simplified Models; Synaptic Plasticity; Long-term Synaptic Plasticity; Reinforcement Learning; Reward-modulated STDP;
Implementer(s): Neymotin, Sam [Samuel.Neymotin at nki.rfmh.org]; Chadderdon, George [gchadder3 at gmail.com];
Search NeuronDB for information about:  GabaA; AMPA; NMDA; Dopamine; Gaba; Glutamate;
/
arm1d
README
drspk.mod *
infot.mod *
intf6_.mod *
intfsw.mod *
misc.mod *
nstim.mod *
stats.mod *
updown.mod *
vecst.mod *
arm.hoc
basestdp.hoc
col.hoc *
colors.hoc *
declist.hoc *
decmat.hoc *
decnqs.hoc *
decvec.hoc *
default.hoc *
drline.hoc *
filtutils.hoc *
geom.hoc
grvec.hoc *
hinton.hoc *
infot.hoc *
init.hoc
intfsw.hoc *
labels.hoc *
local.hoc *
misc.h *
mosinit.hoc
network.hoc
nload.hoc
nqs.hoc *
nqsnet.hoc *
nrnoc.hoc *
params.hoc
run.hoc
samutils.hoc *
sense.hoc *
setup.hoc *
sim.hoc
simctrl.hoc *
stats.hoc *
stim.hoc
syncode.hoc *
units.hoc *
xgetargs.hoc *
                            
// $Id: nqsnet.hoc,v 1.73 2011/11/01 01:49:40 samn Exp $
// xopen("nqsnet.hoc")

//* params controlling sim from testrf.hoc

//* mkcp0() pre-id  post-id  pre#  post#   distance weight  syn-id   nc ptr  wt1 (eg AMPA+NMDA)
objref nq[2],sq[CTYPi][CTYPi],cp
obfunc mkcp0 () { localobj lo
  lo = new NQS("PRID","POID","STYP","PIJ","DIV","CONV","NSYN","NPRE")
  lo.useslist("PRID",CTYP) lo.useslist("POID",CTYP) lo.useslist("STYP",STYP)
  return lo
}

// CODE: PRID,POID,INCOL,COL1,COL2
obfunc mksp () { localobj lo
  lo=new NQS("CODE","PR","PO","DEL","WT0","WT1") // CODE==PRID(1),POID(2),COLA(3),COLB(4)
  lo.coddec("CODE")
  // lo.useslist("PRID",CTYP) lo.useslist("POID",CTYP) 
  return lo
}
sp=mksp()

//* Numbers and connectivity params

// layer return layer location with 'sublayer' defined by Inhib (+0.5) or other suffix
// E or I should be 1st letter of name, suffix letter will ideally dichotomize into late
// alphabet or early alphabet
func layer () { local x,in,la
  la=0
  if ($1 >= CTYP.count) return -1
  if (sscanf(CTYP.o($1).s,"%c%d%c",&in,&x,&la)<2) return -1
  if (x==23) x=3 // layer 2/3
  if (in==73) x+=0.5 // ascii 73 is 'I'
  if (la>77) x+=0.2 // <='M'
  return x
}

//** getlayerz -- get average z location of a layer, based on frana and tononi
// returns value in micrometers. 0 is at top (layer 1), max val
// is at layer 6 ... actual #s don't matter, only distances between
// the layers matters...
func getlayerz () {
  if($1 == 2 || $1 == 3){
    return 1740 // average of 1540+1940 from frana
  } else if($1 == 4 ) { // this is just midpoint btwn L2/3 and L5
    return 1435
  } else if($1 == 5 ){
    return 1130
  } else if($1 == 6 ){
    return 488
  } else {
    printf("getz ERRA: invalid layer %d!\n",$1)
    return -1
  }
}

//* routines
//** styp() sets synapse type based on presynaptic cell
func styp () { local pr,po
  pr=$1 po=$2
  if (pr==IN && po==IN) { return GA 
  } else if (pr==IN) { return IX
  } else if (pr==SU || pr==DP) { return EX
  } else if (pr==SM) { return AM
  } else if (strm(CTYP.o[pr].s,"^E")) { return EX
  } else if (strm(CTYP.o[pr].s,"^I")) { return IX
  } else printf("styp ERR %s->%s not classified",CTYP.object(pr).s,CTYP.object(po).s)
}

//** ellfld() place the cells inside an ellipse
// r for an ellipse = a*b/sqrt((a*sin(theta))^2 + (b*cos(theta))^2)
proc ellfld () { local a,b,ii,jj,p,seed localobj xv,yv,xo
  seed=239023229
  a=1 b=2
  p=allocvecs(xv,yv) vrsz(allcells*10,xv,yv)
  xv.setrnd(4,2*a,seed) yv.setrnd(4,2*b) xv.sub(a) yv.sub(b)
  jj=0
  for vtr2(&x,&y,xv,yv,&ii) {
    if (a*x^2+b*y^2<1) { ce.o(jj).xloc=x ce.o(jj).yloc=y jj+=1 }
    if (jj==ce.count) break
  }
  print ii,jj
  if (jj!=ce.count) print "Not filled"
  dealloc(p)
}