spike time indexing about suite2p HOT 3 CLOSED

I looked through all the traces and discovered that the indexing was a constant value off - this constant corresponded to the length of the variable dat.stat0. So in order to find the spikes for a given ca2+ trace (dat.F.Fcell{dat.cl.iscell}) I subtracted this constant from the iscell values. See below for same traces with corrected indexing (ignore the scaling in these figures):

from suite2p.

Hi,

Sorry for never getting back to this, it slipped under my radar somehow. We have made several big updates to make the whole pipeline more robust, including the spike deconvolution code. You will now find the result in stat(j).c and stat(j).st for each cell. I would be curious if you still get this strange issue.

Best,
Marius

from suite2p.

Oh, I understand the problem now. Indeed, the spikes were only computed for ROIs tentatively identified as possible cells. This excluded exactly the first Nk ROIs, which was the length of stat0.

The pipeline has been reworked to only ever output tentative cells. So the indexing issue should no longer occur.

from suite2p.