I've been running simulations using Netcon(s) to take a list of spike times and drive synapses onto a single compartment model. So far, so good.
Code: Select all
rf = new File()
for iii = 0,numinputs-1 {
sprint(filename,"dumbfile%d.dat",iii+1)
dumbfile = filename
rf.ropen(dumbfile)
rsp[iii] = new Vector()
n = rsp[iii].scanf(rf)
numspiking = rsp[iii].size()
rf.close
a[iii] = new IC_AMPA(0.5)
a[iii].tau = ampatau
nctemp1 = new NetCon(nil,a[iii],0,1,(ampaconduct/numinputs))
//nctemp1 = new NetCon(nil,a[iii],0,1,(ampaconduct/2))
nclist1.append(nctemp1)
b[iii] = new IC_NMDA(0.5)
nctemp2 = new NetCon(nil,b[iii],0,1,(nmdaconduct/numinputs))
//nctemp2 = new NetCon(nil,b[iii],0,1,(nmdaconduct/2))
nclist2.append(nctemp2)
}
proc loadspikes() { local jj
for jj=0,numinputs-1 {
numss = rsp[jj].size()
for ii=0,numss-1 {
nclist1.o(jj).event((rsp[jj].x[ii] + delay))
nclist2.o(jj).event((rsp[jj].x[ii] + delay))
}}}
fih = new FInitializeHandler("loadspikes()")
Now, I expect that if I use spike times with low ISI - say for instance, 10ms, I would expect that the resulting NMDA current response would overlap, as the next incoming synaptic event would occur before the full decay of the previous evoked NMDA current (Summation of NMDA current).
However, what I get is this:
and if I zoom in on a couple of the EPSCs:
It appears that with each incoming EPSC, the previous one is truncated. Yikes. Any way to fix this?