These spikes are not part of any cluster

[bapungiri]

I have decent number of threshold crossings (>5 std) for channels in two of the shanks. However, they are not part of any cluster.

Other details:

• I initially ran the .dat file for all shanks. However, when I opened the clusters in Phy, I failed to find any clusters for these two shanks. Not sure what went wrong, I then ran the clustering for only these two shanks specifically. After running for few hours, spyking-circus crashed as it found zero templates.
• Data length > 14 hours
• Attached .params file
  RatK_Day2_2019-08-08_04-00-00.txt
• Attached .prb file
  RatK_Day2_2019-08-08_04-00-00.prb.txt
• spyking-circus version 1.0.0

[yger]

Please upgrade SC (1.0.6), since there might be bug fixes in between.

conda upgrade -c spyking-circus spyking-circus

Then let me know if you still have the issue

[bapungiri]

It is still failing even after upgrading spyking-circus to version 1.0.6

[yger]

From the log, it seems the data are not properly loaded. Could it be possible? What do you see when launching the preview GUI

spyking-circus mydata.dat -p

[bapungiri]

First second of filtered raw data as viewed in neuroscope:

Terminal message when loading the data in preview mode, nothing seems wrong here:

First second in the spyking-circus preview gui (I wish it was possible to preview at > 1s):

[yger]

you can

spyking-circus mydata.dat -p -s 100 (to view between 100 and 101s)

But it looks normal.. Can you send me the .log file, after the clustering?

[bapungiri]

Great, did not know that.
Actually, I ran the clustering again by changing [clustering] max_elts to 8000, which overwrote the .log file in the output folder. I can share the file after the processing is finished.

[yger]

What will also be good to share, in order to debug/understand what is happening, is to set make_plots = png in the [clustering] section. And send me the zipped plot folder in the results

[bapungiri]

Sure will do.

[bapungiri]

Attached .log and .png files
outputfiles.zip