Skip to content

Latest commit

 

History

History
 
 

Neuropixels

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
high_density_ecephys
high_density_ecephys
 
 
notebooks
notebooks
 
 
test
test
 
 
README.md
README.md
 
 

README.md

Back to Projects List

NWB for Neuropixels Experiments

Key Investigators

  • Josh Siegle (Allen Institute) @jsiegle
  • Xiaoxuan Jia (Allen Institute) @jiaxx
  • Nile Graddis (Allen Institute) @nilegraddis
  • Tom Davidson (UCSF) @tjd2002
  • Dan Millman (Allen Institute) @everythingevolves

Project Description

We will represent high-density exctracellular electrophysiology (ecephys) data in NWB.

Objective

  1. Flexible representation of spike-sorting outputs
  2. Build analogous functionality for Ophys data (DM).

Approach and Plan

  1. Identify and obtain a pool of raw datasets.
  2. Write these datasets to NWB files (and read them back!).
  3. Ensure that common features of these datasets are supported by the core NWB namespace or an extension.
  4. Document the process of writing ecephys data to NWB.

Progress

  • Implemented generic unit metrics group (for storing snr/peak amplitudes) in core. See pr #435
  • Packaged kilosort/phy outputs with pynwb. See notebook

Next Steps

  • We need to associate mean/std waveforms and electrode table regions with UnitTimes in core. See issue #431
  • Create a tutorial for packaging Kilosort/phy outputs with pynwb

Functionality Table

A table of things that we want to write to NWB files. Please fill in/edit entries! Columns are:

  • functionality: short description
  • tools supported: if the functionality is kilosort-specific, for instance
  • PyNWB class: the PyNWB class(es) that currently supports this feature (or the closest match).
  • notes: additional information.
functionality tools supported PyNWB class notes
Sorted units all ecephys.Clustering misc.UnitTimes NB: UnitTimes was recently refactored (pynwb pr #382). Doc fixes outstanding (nwb-schema issue #117)
Event tiimes all ecephys.Clustering misc.UnitTimes
Unitwise metrics (peak amplitude, snr, classifications) all ecephys.Clustering pr #435 Created a new UnitMetrics class to handle unitwise metrics generically
Eventwise metrics (amplitude, PCs) all ecephys.FeatureExtraction (?) These are likely to be lab specific. A solution based on MultiContainerInterface seems sensible
Unit waveforms all ecephys.ClusterWaveforms Currently, it's not clear how to relate waveforms to unit IDs. issue #431
LFP all ecephys.LFP ecephys.ElectricalSeries
Current Source Density all ecephys.ElectricalSeries (?) not sure if the dimensionality works. Should attach to LFP?

Illustrations

Background and References