Etienne Ackermann

This is related to within-epoch interpolation (in an attempt to avoid across-epoch interpolation). Not obvious what's going on yet; need clear debugging strategy.

binnedsupport needs to be updated when partitioning

2

currently, partitioning works for EpochArrays, AnalogSignalArrays, and SpikeTrainArrays, but we need to override the method for BinnedSpikeTrainArrays so that it also updates bst.binnedSupport

smooth place fields etc with nans

1

This is a very useful recipe, and should be incorporated in a sensible place: see https://stackoverflow.com/questions/18697532/gaussian-filtering-a-image-with-nan-in-python To smooth in 2D with NANs.

TuningCurve improvements

2

Currently we have 1D and 2D [implicitly continuous] tuning curve objects. However, they explicitly assume a units --> external correlates relationship. When we learn virtual tuning curves with our HMMs,...

implement PHMM.sample()

1

This should either return a BinnedSpikeTrainArray with all the proper unit_ids, labels, etc., or it should return a feature matrix. Requesting multiple sequences may also make the timestamps interesting.

Currently, rastercountplot is almost hard-coded; It creates a figure INSIDE of the function call! Which is just crazy. The correct approach is to add an axis on top of the...

Validate 2D spatial info and sparsity with explicit calculations

1

I extended the 1D calculations to 2D so that dimensions match, but I want to be 100% sure that they are being calculated correctly. Some of the values seem a...