SpikeInterface-GUI

GUI for the SortingAnalyzer object from SpikeInterface.

This is a cross platform interactive viewer to inspect the final results and quality of any spike sorter supported by spikeinterface.

This interactive GUI offers several views that dynamically refresh other views. This allows us to very quickly check the strengths and weaknesses of any sorter output and to perform manual curation.

This can be used as a replacement of phy.

The documentation is available here.

This viewer has 2 modes:

mode=desktop : based on Qt

mode=web : based on Panel (can be used remotely)

Installation

For beginners users, please see our installation tips where we provide a yaml for Mac/Windows/Linux to help properly install spikeinterface and spikeinterface-gui for you in a dedicated uv environment.

In your environment, if you wish to use the Desktop version of the GUI, you can do:

pip install 'spikeinterface-gui[desktop]'

Note: this installs PySide6. You can use the PyQt5 backend instead by uninstalling PySide6 and then installing PyQt5.

If you wish to use the Web version of the GUI, you can do:

pip install 'spikeinterface-gui[web]'

From source:

git clone https://github.com/SpikeInterface/spikeinterface-gui.git
cd spikeinterface-gui
pip install .

You'll then need to install the appropriate backends yourself (pyqtgraph and PySide6 or PyQt5 for the desktop; panel and bokeh for web).

Main usage

The main idea is to make visible one or several units and visually inspect if they should be merged or removed.

For this visibility:

• ctrl + double click on a unit in probeview
• check the box visible in the unitlist
• double click on one unit in unitlist unit visible alone
• move one of the roi in the probeview

Views can be reorganized by moving docks by clicking in the title bar of a docks. Any dock (view) can be closed. And can be put back with right click in any title bar of any dock.

Every view has a ? button which opens the contextual help. These inplace docs are the most important stuff to be read.

When some units are visible, the related spike list can be refresh. Then selecting spike per spike can also refresh some views. This enables a very quick and convenient spike per spike jump on traces.

Channel visibility can be handled with one of the roi in the probeview.

Shortcuts: many shortcuts are available, please read the ? button in each view.

Curation mode

By default this tool is a viewer only. But you can turn it into a tool for manual curation using the curation=True option. This tool supports the curation format from spikeinterface. This format enables to:

1. remove units
2. merge units
3. split units
4. create manual labels

When this mode is activated a new view is added on top left to maintain the list of removal and merges. The curation format can be exported to json.

Launching the GUI

In order to use this viewer you will need to know a bit of spikeinterface documentation

Step 1: create and compute SortingAnalyzer

You first need to get a SortingAnalyzer object with spikeinterface.

See help here

Note that:

• some extensions are mandatory (unit_location, templates, )
• some extensions are optional
• the more extensions are computed the more views are displayed

Example:

import spikeinterface.full as si
recording = si.read_XXXX('/path/to/my/recording')
recording_filtered = si.bandpass_filter(recording)
sorting = si.run_sorter('YYYYY', recording_filtered)


job_kwargs = dict(n_jobs=-1, progress_bar=True, chunk_duration="1s")

# make the SortingAnalyzer with necessary and some optional extensions
sorting_analyzer = si.create_sorting_analyzer(sorting, recording_filtered,
                                              format="binary_folder", folder="/my_sorting_analyzer",
                                              **job_kwargs)
sorting_analyzer.compute("random_spikes", method="uniform", max_spikes_per_unit=500)
sorting_analyzer.compute("waveforms", **job_kwargs)
sorting_analyzer.compute("templates", **job_kwargs)
sorting_analyzer.compute("noise_levels")
sorting_analyzer.compute("unit_locations", method="monopolar_triangulation")
sorting_analyzer.compute("isi_histograms")
sorting_analyzer.compute("correlograms", window_ms=100, bin_ms=5.)
sorting_analyzer.compute("principal_components", n_components=3, mode='by_channel_global', whiten=True, **job_kwargs)
sorting_analyzer.compute("quality_metrics", metric_names=["snr", "firing_rate"])
sorting_analyzer.compute("template_similarity")
sorting_analyzer.compute("spike_amplitudes", **job_kwargs)

Step 2: open the GUI for one analyzer

With python:

from spikeinterface_gui import run_mainwindow
# reload the SortingAnalyzer
sorting_analyzer = si.load_sorting_analyzer("/my_sorting_analyzer")
# open and run the Qt app
run_mainwindow(sorting_analyzer, mode="desktop", curation=False)
# open and run the Web app with curation activated
run_mainwindow(sorting_analyzer, mode="web", curation=True)

Or from spikeinterface:

import spikeinterface.widgets as sw
sorting_analyzer = load_sorting_analyzer(test_folder / "sorting_analyzer")
sw.plot_sorting_summary(sorting_analyzer, backend="spikeinterface_gui")

With the command line

sigui /path/for/my/sorting_analyzer

The command line supports some options like --no-traces or --curation or --mode

sigui --mode=web --no-traces --curation /path/for/my/sorting_analyzer

Running with curation mode

To open the viewer with curation mode use curation=True.

from spikeinterface_gui import run_mainwindow
run_mainwindow(sorting_analyzer, curation=True)

import spikeinterface.widgets as sw
sw.plot_sorting_summary(sorting_analyzer, curation=True, backend="spikeinterface_gui")

The curation_dict can be saved inside the folder of the analyzer (for "binary_folder" or "zarr" format). Then it is auto-reloaded when the gui is re-opened.

Open the GUI launcher

If you have multiple analyzer folders in the same root folder (or in subfolders), you can use the GUI launcher to select which one to open.

from spikeinterface_gui import run_launcher

run_launcher(root_folder="path-to-my-analyzers", mode="desktop")

You can also use the command line and just specify the --root-folder:

sigui --mode=desktop --root-folder path-to-my-analyzers

sigui --mode=web --root-folder path-to-my-analyzers

Customizing layout and settings

You can create your own custom layout by specifying which views you'd like to see, and where they go. The basic window layout supports eight "zones", which are laid out as follows:

+---------------+--------------+
| zone1   zone2 | zone3  zone4 |
+               +              +
| zone5   zone6 | zone7  zone8 |
+---------------+--------------+

If a zone has free space below it or to the right of it, it will try to use it. Stretching downwards takes precedence over stretching rightwards. E.g. suppose your layout is only non-empty in zones 1, 4, 5, 6 and 7:

+---------------+--------------+
| zone1         |        zone4 |
+               +              +
| zone5   zone6 | zone7        |
+---------------+--------------+

Then zone1 will stretch right-wards to make a two-zone view. Zone4 will stretch downwards to make a long two-zone view.

To specify your own layout, put the specification in a .json file. This should be a list of zones, and which views should appear in which zones. An example:

my_layout.json

{
    "zone1": ["unitlist", "spikelist"], 
    "zone2": ["spikeamplitude"], 
    "zone3": ["waveform", "waveformheatmap"], 
    "zone4": ["similarity"], 
    "zone5": ["spikedepth"], 
    "zone6": [], 
    "zone7": [], 
    "zone8": ["correlogram"]
}

When you open spikeinterface-gui, you can then point to the my_layout.json using the --layout_file flag:

sigui --layout_file=path/to/my_layout.json path/to/sorting_analyzer

Find a list of available views in this file.

You can also edit the initial settings for each view. There are two ways to do this. First, you can open the gui, edit the settings of any or all views, then press the "Save as default settings" button in the mainsettings view. This will save a settings.json file in your spikeinterface-gui config (saved in ~/.config/spikeinterface_gui/). The saved settings will automatically load next time you start the gui.

If you want more direct control, you can pass your own setting json file to the settings-file flag (or by passing a user_settings dict to run_mainwindow). The settings file should contain a settings dict for each view. The view names are the same as used in the layout file, and each setting name can be seen in the settings tab of each view. Below is an example settings file:

my_settings.json

{
    "probe": {
        "show_channel_id": true
    },
    "waveform": {
        "overlap": false, 
        "plot_std": false
    },
    "spikeamplitude": {
        "max_spikes_per_unit": 1000, 
        "alpha": 0.2
    }
}

You can then use this file by running e.g.

sigui --settings_file=path/to/my_settings.json path/to/sorting_analyzer

Deploy the GUI in web mode

The :ref:launching section showed how you can launch the GUI on a local workstation.

To recap, you can launch the GUI in web mode by running:

sigui --mode=web /path/for/my/sorting_analyzer

which will start a server on your localhost:

Launching server at http://localhost:43957

However, the GUI can also be deployed on a remote server, and accessed through a web browser or directly in the cloud, thanks to its web mode.

Deploying on a remote lab server

If you want to deploy the GUI on a remote lab server, the GUI can be used by multiple users with VPN access to the server.

To setup the web GUI on a remote server, follow these steps:

1. Choose a server machine in your lab network (e.g. "my-lab-server") that is accessible through VPN by all lab members.

2. Install spikeinterface-gui and its dependencies on the server machine.

3. Launch the GUI launcher on the server machine with the following command:

sigui --mode=web --address=auto-ip

If all your analyzers will be in the same root folder (or in subfolders), you can also specify the --root-folder option:

sigui --mode=web --address=auto-ip --root-folder=/path/to/my/analyzers

4. The server will start and display the IP address to access the GUI launcher, e.g.:

Launching server at http://SERVER.IP.ADDRESS:43957/launcher

5. Share the displayed IP address with all lab members, so they can connect to the GUI launcher from their local machines.

For developers

Message from dictator (Samuel)

Contrary to the spikeinterface package, for the development of this viewer all good practices of coding are deliberately put aside: no test, no CI, no auto formatting, no doc, ... Feel free to contribute, it is an open wild zone. Code anarchists are very welcome. So in this mess, persona non grata: pre-commit, black, pytest fixture, ...

Credits

Original author: Samuel Garcia, CNRS, Lyon, France

This work is a port of the old tridesclous.gui submodule on top of spikeinterface.

Main authors and maintainers:

• desktop side: Samuel Garcia, CNRS, Lyon, France
• web side: Alessio Paolo Buccino, Allen Institute for Neural Dynamics, Seattle, USA
• maintainer: Chris Halcrow, University of Edinburgh, Scotland, UK