SpikeInterface-Training-Milan-Sept25

Material for SpikeInterface Tutorial @ Human Technopole - Milan - September 2025

Schedule

Day 3 - Theoretical overview - Wed 24/09 14:00-15:00

1. Introduction to Spike Sorting (30 min - Sam)
2. Introduction to SpikeInterface (25 min - Alessio)
3. Installation check, get jupyter lab ready to go (5 min - Chris)

Day 3 - Hands on tutorials - Wed 24/09 15:00-18:00

1. Reading + Manipulation (30 min - Chris)

   • Read and visualise data
   • The importance of probes

2. Preprocessing (45 min - Alessio)

   • Apply and visualize preprocessing
   • Detect bad channels
   • Motion correction

3. Spike sort (30 min - Sam)

   • Run a spike sorter
   • Compare spike sorter outputs

4. Analyze Sorting (45 min - Chris)

   • Create a SortingAnalyzer to explore your sorting
   • Compute some extensions
   • Play with the extension outputs

5. Curate (30 min - Alessio)

   • Simple, threshold-based curation
   • Curation using ML models
   • Merging units

Day 4 - GUI and Pipelines - Thu 25/09 09:30-13:00

1. SpikeInterface-GUI (60min - Sam)

10:30-11.00: Coffee break

2. Develop a full pipeline script (90min - Chris)
   • Basics of pipelines in Python
   • Designing a pipeline for your own data

Installation

We recommend that you install uv. This is an unbelievably good new package management tool for Python.

Open your Terminal app and cd (change directory) to where you like to keep github repos. Then clone this repo, and navigate to the notebooks folder of it by entering the following commands into your Terminal:

git clone https://github.com/SpikeInterface/SpikeInterface-Training-Milan-Sept25.git
cd SpikeInterface-Training-Milan-Sept25/notebooks/

Now open jupyter lab (you don't need to install this - uv will take care of it) using uv:

uv run jupyter lab

Or if you prefer to use vscode:

uv run code .

If you like traditional virtual environments, please visit this page to install on your laptop python+spikeinterface.

Computing node with a preinstalled environment will be also provided.

Overview of available datasets

In vivo

M25_D23_2024-11-11_13-11-10_OF1

Neuropixels 2.0 dataset (384 channels) acquired with the Open Ephys GUI in binary format. You can use the spikeinterface.extractor.read_openephys() function to read the data.

1544_2023-04-21_09-55-34_of

Recording with 16 channels organized in 4 tetrodes acquired with the Open Ephys GUI in "openephys" format. You can use the spikeinterface.extractor.read_openephys() function to read the data.

aind_ecephys_session.zarr

Recording with a Neuropixels 2.0 - 4 shank probe stored as a compressed SpikeInterface Zarr folder. You can use the spikeinterface.load() (or spikeinterface.read_zarr()) function to read the data.

cambridgeNT_openephys

Recording with a Cambridge Neurotech ASSY-236-H5 probe and wired to the Open Ephys acquisition system via the Cambridge NeuroTech mini-amp-64. You can use the spikeinterface.extractor.read_openephys() function to read the data. The probe is available from the probeinterface_library (see docs here) and the wiring is available in ProbeInterface.

In vitro

mouse_CTX_14D.brw

Primary mouse cortex for E18 embryos. Plated 80K on CorePlate 1W38/60 3Brain, recorded after 14 days in culture with 3Brain BioCAM system. You can use the spikeinterface.extractor.read_biocam() function to read the data.

human_ngn2_100K_c_55D.brw

Human glutamatergic neurons induced via ngn2 method. Plated 100k on CorePlate 1W38/60 3Brain, recorded after 55 days in culture with 3Brain BioCAM system. You can use the spikeinterface.extractor.read_biocam() function to read the data.

Important

The 3Brain BioCAM system saves the data in unsigned integers. You should convert it to signed integers for a smooth processing. See documentation here. Note that the bit_depth for this system is 12!

Please download these datasets on your laptop from this link:

https://filesender.renater.fr/?s=download&token=420bcfb8-f134-473a-be40-c28235dda5da