Open-ephys gui - PSTH how to get it working

[Bassam Atallah]

I was hoping someone could tell me what I am doing wrong with the openephys GUI... I wanna see spike\ lfp averages aligned on ttl triggers. I have tried with multiple data flow configurations and all yield the same result: I don't get any PSTH for either spikes for LFPs. There isn't even an empty PSTH axes! What am I doing wrong?

thanks

[Josh Siegle]

Did one of your configurations include a Spike Sorter? The PSTH depends on this module for information about electrode configuration. If it’s not there, then nothing will show up. I just realized this wasn’t in the docs, so I added a note about it.

If you did try it with the Spike Sorter already, send more details about your signal chain and we’ll try to replicate the problem.

Josh

--
You received this message because you are subscribed to the Google Groups "Open Ephys" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-ephys+...@googlegroups.com.
To post to this group, send email to open-...@googlegroups.com.
Visit this group at http://groups.google.com/group/open-ephys.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-ephys/2f8df4c1-f7ef-459c-a976-8a35b547bd0d%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[dmsc...@gmail.com]

Hi Josh,

I am also having problems with the new PSTH functionality. I am trying to create PSTHs triggered off on an external TTL. I'm collecting data with an Intan system. I am sorting a single unit on Ch1 (which is actually just 60Hz noise, but produces many "spikes" per second). A single PSTH window shows up in the PSTH pane, as it should, and the trial counter increments appropriately every time I send a TTL, but the traces remain flatlined at 0 Hz, despite that fact that my sorted unit is "spiking" at 10s of Hz. Any thoughts? Nothing changes with I mark a subset or all of the TTLs, but the flatlines disappear (as I think they should) when I de-select all of the TTLs.

As an aside, I noticed that even if I don't select a TTL for the trigger, my trial counter still increments every time I send a high signal to TTL1.

Thanks,
David

[Bassam Atallah]

 So I tried with the spike sorter.

You are right... now empty figures come up.

However, the figures never fill up with anything, they stay empty. 

I don't see an average LFP or spike rate histograms (despite seeing plenty of spikes in the sorter).

It sounds similar to David's experience.

[Josh Siegle]

I’ll take a look at this next week. It was definitely working at one point, but it sounds like it may have been broken in the meantime.

Josh

To view this discussion on the web visit https://groups.google.com/d/msgid/open-ephys/24eb9cbc-23ca-4660-adab-dda77c3738e0%40googlegroups.com.

[Josh Siegle]

Ok, I made some progress with this today. The main source of problems was that the PSTH node deals with software spike + event times, whereas the rest of the GUI uses hardware spike times. I updated the Spike Sorter so that the timestamps are compatible with the PSTH node (it still keeps track of hardware timestamps for saving purposes).

Now, when I input a TTL (and make sure that appropriate “condition” is highlighted on the right), both the average firing rate and average LFP plots will update. Can someone else confirm that this is the case? I get the same behavior when I input a network event trial start, so I think the behavior is basically OK.

What I *haven’t* done is to input simulated spikes at a known time offset to a TTL event, to make sure the temporal alignment is working. Does anyone want to take a stab at this? It would be pretty simple to set up with a signal generator.

Josh

To view this discussion on the web visit https://groups.google.com/d/msgid/open-ephys/24eb9cbc-23ca-4660-adab-dda77c3738e0%40googlegroups.com.

[Josh Siegle]

We just tested the PSTH node with electrodes in LGN while aligning to brief light flashes. The average spiking response looks exactly as expected. There are still a few bugs (e.g., changing the window size causes a crash), but general alignment to TTL inputs seems to be working again.

Josh

[Jan Zimmermann]

I have tested the PSTH node with both network and TTL and all behavior seems fine.

The only thing that I find odd is the timestamping of network events. They seem to be floored or sampled at a lower frequency than TTLs.

Is that expected?

Jan

[Josh Siegle]

The “floor” effect you’re seeing is due to the fact that we can only estimate event arrival times with the resolution of the software buffer (on the order of ~20 ms). Each network event is tagged with the first hardware timestamp of the nearest buffer, since it’s synced at the level of software. The TTLs, on the other hand, are synced to the hardware clock on the acquisition board, so they have the same resolution as the neural data.

Josh

To view this discussion on the web visit https://groups.google.com/d/msgid/open-ephys/aed831a5-73ff-4565-8f93-db6e8d84c774%40googlegroups.com.