Another Synchrony Measure

Post Reply
Bill Connelly
Posts: 86
Joined: Thu May 22, 2008 11:54 pm
Location: Australian National University

Another Synchrony Measure

Post by Bill Connelly »

I wasn't sure if here was the best place to ask this, or in the networks section where Bill Lytton might see it; but anyway, The Wang and Buzsaki model in gamma oscillations is pretty popular.

They use a measure of synchrony that I would love to have a look at, but for the life of me, I can't figure out what they are talking about.

Image

I'm with them until "...and that two spike trains are given by X(l))=0..."

What are X(l) and Y(l) supposed to be? Is l essentially the bin number? Where a spike may or may not have happened?

I set up this function to record spike times:

Code: Select all

  for n=0, ncells-1 {
    cellV[n].record(&cl.o(n).soma.v(0.5))
    
    cl.o(n).soma APNC = new NetCon(&v(0.5), nil)
    APNC.threshold = 0
    APNC.record(spikeTime, spikeCell, n)
    cellNC.append(APNC)
  }
So I don't have a nice 2d array of Cell[cellNumber][APnumber]=timeOfAP which would make this calculating this synchrony measure easy. Instead I have two vectors, SpikeTime and SpikeCell, where Action potential X occured in cell spikeCell.x(X) and occured at time spikeTime.x(X)

No; I really can't think of how to code this synchrony measure. Can anyone? Should I try and convert those two arrays into a cell-centric Cell[cellNumber][APnumber]=timeOfAP format. Or even better a Cell[cellNumber][Time] = binaryAPYesNo. Even psuedo code would be appreciated.
wwlytton
Posts: 66
Joined: Wed May 18, 2005 10:37 pm
Contact:

Re: Another Synchrony Measure

Post by wwlytton »

the first thing to do is to get the binning done using either
obj = vdest.hist(vsrc, low, size, width)
OR
newvect = vsrc.histogram(low, high, width)

in order to do this you will need to pick up t vectors on a per-cell basis (as you said)
you can do this by using indvwhere()
vdest.indvwhere(vsource,"==",cellnum)
on your index vectors and then
obj = vdest.index(vsrc, indices)
to grab the correspoing times

I'll look around and see if we already have the code for this particular measure --
though it may be buried in so much additional code that it won't be worth
your while to learn (ie see my NQS stuff on simtooldb)

let me know if this helps
Post Reply