Phy only shows average templates after spikeinterface.export_to_phy; raw per-spike waveforms cannot be displayed

[Sicheng-package]

Phy only shows average templates after spikeinterface.export_to_phy; raw per-spike waveforms cannot be displayed.What can i do to solve this issue？

[alejoe91]

Hi, do you see the raw traces? The per-spike waveforms are extracted from the traces on the fly, so if Phy cannot load traces, then it won't show single waveforms either!

[Sicheng-package]

Hi, thanks for the quick reply.
I also can’t see raw traces in Phy. TraceView fails to open, so single-spike waveforms don’t show either.

What I see in Phy
(phy2) E:\testephys\phy_qcpass_clusters2> phy template-gui params.py
22:00:32.261 [W] model:1415 File None does not exist.
22:00:34.752 [W] model:667 Skipping spike waveforms that do not exist, they will be extracted on the fly from the raw data as needed.
22:00:34.754 [W] traces:462 File E:\testephys\phy_qcpass_clusters2\None does not exist.
22:00:34.754 [W] traces:462 File E:\testephys\phy_qcpass_clusters2\None does not exist.
QMimeDatabase: Error loading internal MIME data
An error has been encountered at line 1 of : Premature end of document.:
22:17:41.848 [W] gui:671 Could not create view TraceView.
22:17:49.060 [W] gui:671 Could not create view TraceView.
22:18:04.658 [W] gui:671 Could not create view TraceView.

The key part seems to be File None does not exist. and the TraceView creation failure.

How I exported from SpikeInterface

export_to_phy

an_qc = an.select_units(unit_ids=qc_unit_ids)
sorting_dedup.copy_metadata(an_qc.sorting, ids=an_qc.sorting.unit_ids)
normalize_str_props(an_qc.sorting, keys=("group","KSLabel","quality"))
props_for_export = [p for p in ("group","KSLabel","quality") if p in an_qc.sorting.get_property_keys()]

from pathlib import Path
PHY_DIR = Path(r"E:/testephys/phy_qcpass_clusters2")

export_to_phy(
sorting_analyzer=an_qc,
output_folder=PHY_DIR,
remove_if_exists=True,
copy_binary=False, # <-- I set this to False
additional_properties=props_for_export,
)
print("Exported to Phy ->", PHY_DIR)

Should I re-export with copy_binary=True so the exporter writes a proper .bin and params.py?

[alejoe91]

Exactly, that's the reason. You need copy_binary=True

[Sicheng-package]

Thanks — that was exactly it. Setting copy_binary=True makes Phy load the traces, so single-spike waveforms show up. 🙏

Quick question: does SpikeInterface have a built-in way to handle events/TTL (e.g., extracting edges from NIDQ and aligning spikes to events) so I can keep everything inside SI? I know I can load the NIDQ stream with read_spikeglx(..., stream_name="nidq"), and I see there’s a BaseEvent API for events, but is there an official helper/example for turning NIDQ TTLs into BaseEvent objects and doing basic alignment/PETH? If so, could you share a pointer? Thanks a lot!

[alejoe91]

Good question. Note really currently but @h-mayorquin can comment on it.

Tagging also @guptadivyansh and @manimoh since they looked into this:
NeuralEnsemble/python-neo#1657

[h-mayorquin]

There is a SpikeGLXEventExtractor in the extractor clasess that should give you the timing of the events. It would be good to have it tested.

As far as I am aware we don't the machinery for doing the interpolation but that should not be too difficult with numpy.

[Sicheng-package]

Thanks a lot for the pointer to SpikeGLXEventExtractor—that’s exactly what I needed for getting event timings. I’ll test it on my SpikeGLX datasets.

[h-mayorquin]

Great. Let us know if it works as expected and, if you can, give us some stars on the repo as that is good for us to get funding for doing more development : )

[Sicheng-package]

Everything is working great on my side—thanks for the pointers!
I extracted the events and performed interpolation to align AP and NIDAQ to the same timebase.
Also, I just ⭐️ the repo—hope it helps with funding!