Add ASt NeuroPixels dataset

[weiglszonja]

NeuroPixels dataset consists of:

• recording in SpikeGLX format
• sorting in Phy format
• Histology image: shows where the neuropixel probes were inserted

[weiglszonja]

• add convenient methods for running sessions in parallel (similar as before)
• rerun from clean environment
• update stubbed session at private bucket (.../tye_lab_to_nwb/NPX3_ecephys.nwb)

[CodyCBakerPhD]

Overall looks good, my only requested changes involve the massive number of properties added to the Units table

Could you do 2 things there

1. Expand the column names to use full English references; e.g., isi_viol -> isi_violation
2. Add descriptions to many of the columns; you can use this reference to help

[CodyCBakerPhD]

3. Also just noticed; the columns firing_rate and fr are duplicated; you can remove fr altogether in that case

[weiglszonja]

Thank you @CodyCBakerPhD!
For reference the list of items I'll need to ask descriptions for:

3 types of "amplitude" columns that have different values:

Amplitude -> the values for the first 3 units: [2967.  4477.  1957.8]
amp -> the values for the first 3 units: [179.66105769 383.97253788 124.73197993]
amplitude -> the values for the first 3 units: e.g. [61.62191772 88.24165344 71.04180145]

Other columns that I'm not sure about whay they mean:

After looking at spikeinterface metrics these are the items that remain unclear:

sh -> all zeros
spread
velocity_above
velocity_below
depth
peak_channel

Info from Laurel:

sh : this is most likely shank. In our data we have only one shank so this should be all 0s in our data.

spread: we think this is the spread of the unit in depth. It is most likely measured in micrometers.

velocity_above, velocity below: We hypothesize that this relates to the measure of the slope of the waveform (either up or down) or the onset to the peak amplitude.

depth: this is most likely probe depth, measured in micrometers, starting with 0 at the tip of probe.

ch/peak_channel: this is the neuropixel channel. We think peak_channel is calculated from ch then stored in peak_channel.

Amplitude/amplitude/amp: These are very confusing! We see that they seem to have different units but don't appear to correlate in size across the three columns, so they are not direct transformations. Since the information in cluster_info.tsv appears to be an information dump of all variables from Cluster_View, I think it probably represents the 3 amplitude types you can toggle between in that visualization GUI. Their documentation page (https://phy.readthedocs.io/en/latest/visualization/#amplitude-view) lists this for the 3 types:

Different types of amplitudes

You can toggle between different types of amplitudes by pressing a:

template: the template amplitudes (stored in amplitudes.npy)
raw: the raw spike waveform maximum amplitude on the peak channel (at the moment, extracted on the fly from the raw data file, so this is slow).
feature: the spike amplitude on a specific dimension, by default the first PC component on the peak channel. The dimension can be changed from the feature view with control+left click (x axis) and control+right click (y axis).
However, I could not tell you which variable relates to which of these types.

[CodyCBakerPhD]

@weiglszonja Minor conflict here; I'd try to resolve but I don't want to accidentally mess anything up

[weiglszonja]

thank you @CodyCBakerPhD , I updated this.