Pierre Yger

From what I can tell, SC exports PCS for all the channels where the templates is defined. I must admit that I do not use the feature view that much,...

What do you mean by pre-detected MUA spikes. The code has a function such that if you are providing your own spike times and labels, it will create the templates...

This is not implemented by default in the algorithm. But I could make it as an option, if you are interested, since it should not be too difficult. I guess...

Indeed, this is strange, because from benchmarks the methods do not behave so differently. The high number of cluster persists even after merging, at the end? Because I agree with...

you were all using more or less independent channels ? Maybe nd-bhatta is less efficient for such a configuration, I'll make some more tests. But yes, otherwise you can safely...

The Bhattacharyya distance is used to distinguish if two distributions are the same or not. While you have a simple case for 1D distribution (bhatta), it can also be extended...

Could be, yes, assuming we have a way to read .mpx file in python. Actually, there seems to be an experimental wrapper in python-neo, I'll have a look, but the...

SC does not use tSNE, mostly because I'm afraid that it might be very slow, and requires fine hand parameter tuning, depending on the dataset. But this is something we...

Never had such an error, and in theory MPI should not care whether you are online or offline. SC can of course run without an internet access (that is not...

No problem, thanks for reporting. But again, sorry to bother, but can you upload a recording/param/probe? Always better to debug. I've no idea why some channels are missing.... Regarding the...