3D olfactory bulb: operators (Migliore et al, 2015)

 Download zip file   Auto-launch 
Help downloading and running models
Accession:168591
"... Using a 3D model of mitral and granule cell interactions supported by experimental findings, combined with a matrix-based representation of glomerular operations, we identify the mechanisms for forming one or more glomerular units in response to a given odor, how and to what extent the glomerular units interfere or interact with each other during learning, their computational role within the olfactory bulb microcircuit, and how their actions can be formalized into a theoretical framework in which the olfactory bulb can be considered to contain "odor operators" unique to each individual. ..."
Reference:
1 . Migliore M, Cavarretta F, Marasco A, Tulumello E, Hines ML, Shepherd GM (2015) Synaptic clusters function as odor operators in the olfactory bulb. Proc Natl Acad Sci U S A 112:8499-504 [PubMed]
Model Information (Click on a link to find other models with that property)
Model Type: Realistic Network;
Brain Region(s)/Organism:
Cell Type(s): Olfactory bulb main mitral GLU cell; Olfactory bulb main interneuron granule MC GABA cell;
Channel(s): I Na,t; I A; I K;
Gap Junctions:
Receptor(s): AMPA; NMDA; Gaba;
Gene(s):
Transmitter(s): Gaba; Glutamate;
Simulation Environment: NEURON; Python;
Model Concept(s): Activity Patterns; Dendritic Action Potentials; Active Dendrites; Synaptic Plasticity; Action Potentials; Synaptic Integration; Unsupervised Learning; Sensory processing; Olfaction;
Implementer(s): Migliore, Michele [Michele.Migliore at Yale.edu]; Cavarretta, Francesco [francescocavarretta at hotmail.it];
Search NeuronDB for information about:  Olfactory bulb main mitral GLU cell; Olfactory bulb main interneuron granule MC GABA cell; AMPA; NMDA; Gaba; I Na,t; I A; I K; Gaba; Glutamate;
/
figure1eBulb3D
readme.html
ampanmda.mod *
distrt.mod *
fi.mod *
fi_stdp.mod *
kamt.mod *
kdrmt.mod *
naxn.mod *
ThreshDetect.mod *
.hg_archival.txt
all2all.py *
balance.py *
bindict.py
binsave.py
binspikes.py
BulbSurf.py
catfiles.sh
colors.py *
common.py
complexity.py *
custom_params.py *
customsim.py
destroy_model.py *
determine_connections.py
distribute.py *
falsegloms.txt
fixnseg.hoc *
g37e1i002.py
gidfunc.py *
Glom.py *
granule.hoc *
granules.py
grow.py
input-odors.txt *
loadbalutil.py *
lpt.py *
m2g_connections.py
mayasyn.py
mgrs.py
misc.py
mitral.hoc *
mkdict.py
mkmitral.py
modeldata.py *
multisplit_distrib.py *
net_mitral_centric.py
odordisp.py *
odors.py *
odorstim.py
params.py
parrun.py
realgloms.txt *
realSoma.py *
runsim.py
spike2file.hoc *
split.py *
util.py *
vrecord.py
weightsave.py *
                            
from common import *
h.load_file("spike2file.hoc")


idvec = h.Vector()
idvec.buffer_size(5000000)
spikevec = h.Vector()
spikevec.buffer_size(5000000)

n_spkout_files = max(nhost/64, 1) # each file contains spikes from 64 ranks
n_spkout_sort = min(n_spkout_files*8, nhost) #each file serializes from 8 ranks
# so each sorting rank gathers spikes from nhost/n_spkout_sort ranks
checkpoint_interval = 50000.

def prun(tstop):
  cvode = h.CVode()
  cvode.cache_efficient(1)
  #pc.spike_compress(0,0,1)
  pc.setup_transfer()
  mindelay = pc.set_maxstep(10)
  if rank == 0: print 'mindelay = %g'%mindelay
  runtime = h.startsw()
  exchtime = pc.wait_time()

  inittime = h.startsw()
  h.stdinit()
  inittime = h.startsw() - inittime
  if rank == 0: print 'init time = %g'%inittime
  
  while h.t < tstop:
    told = h.t
    tnext = h.t + checkpoint_interval
    if tnext > tstop:
      tnext = tstop
    pc.psolve(tnext)
    if h.t == told:
      if rank == 0:
        print "psolve did not advance time from t=%.20g to tnext=%.20g\n"%(h.t, tnext)
      break
  
    # save spikes and dictionary in a binary format to
    # make them more comprimibles
    import binsave
    binsave.save(params.filename, spikevec, idvec)
    #@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@

    
    h.spike2file(params.filename, spikevec, idvec, n_spkout_sort, n_spkout_files)
  
  runtime = h.startsw() - runtime
  comptime = pc.step_time()
  splittime = pc.vtransfer_time(1)
  gaptime = pc.vtransfer_time()
  exchtime = pc.wait_time() - exchtime
  if rank == 0: print 'runtime = %g'% runtime
  printperf([comptime, exchtime, splittime, gaptime])

def printperf(p):
  avgp = []
  maxp = []
  header = ['comp','spk','split','gap']
  for i in p:
    avgp.append(pc.allreduce(i, 1)/nhost)
    maxp.append(pc.allreduce(i, 2))
  if rank > 0:
    return
  b = avgp[0]/maxp[0]
  print 'Load Balance = %g'% b
  print '\n     ',
  for i in header: print '%12s'%i,
  print '\n avg ',
  for i in avgp: print '%12.2f'%i,
  print '\n max ',
  for i in maxp: print '%12.2f'%i,
  print ''
 
if __name__ == '__main__':
  import common
  import util
  common.nmitral = 1
  common.ncell = 2
  import net_mitral_centric as nmc
  nmc.build_net_roundrobin(getmodel())
  pc.spike_record(-1, spikevec, idvec)
  from odorstim import OdorStim
  from odors import odors
  ods = OdorStim(odors['Apple'])
  ods.setup(nmc.mitrals, 10., 20., 100.)
  prun(200.)
  util.finish()

Loading data, please wait...