Our Portfolio

Axona Data Processing
Columbia University

Role: NWB specialist and developer

portfolio

Collaboration with: Columbia University

The CatalystNeuro team worked with Dr. Hussaini and his lab to utilize a state-of-the-art preprocessing workflow on electrophysiological data collected as part of their research on Alzheimer’s Disease. Data was being collected from the Axona acquisition system and from Intan. They used MountainSort and were interested in comparing these results to other cutting-edge spike sorters and accelerating data processing through curation using spike train metrics, consensus across sorters, and finally manual curation. They wanted the output of this processing system to be capable of writing the results in TINT format. We built this system in SpikeInterface, a package designed to build modular spike sorting pipelines.

DANDI
MIT

Role: NWB specialist and Community Outreach

portfolio

Collaboration with: Massachusetts Institute of Technology

DANDI is a BRAIN-funded project to build an archive for publicly sharing cellular neurophysiology data collected by the BRAIN Initiative. Our role in the project is to build support for NWB files, and to help the DANDI team reach out to NWB users to build a library of data. https://dandiarchive.org/dandiset/000055/draft

Data engineering speech neuroscience collaboration
UCSF

Role: NWB specialist and developer

portfolio

Collaboration with: UC San Francisco

We worked with the labs of Dr. Edward Chang, Dr. Nitin Tandon, and Dr. Mark Richardson to use NWB 2.0 as a foundation for standardizing ECoG data. Through this effort we have restructured internal tools to work with the NWB format, and have publicly released high density ECoG recordings during speech on DANDI here. Through this project, we built:

  • ecogVIS, a python GUI for visualizing and annotation of raw voltage traces and standard Chang Lab ECoG preprocessing
  • ndx-ecog, the NWB extension for storing the cortical surface of a subject.
  • ndx-bipolar-table, an NWB extension for storing bipolar schema.
  • ndx-hierarchical-behavioral-data, an NWB extension for storing hierarchical behavior like transcriptions of natural speech.
  • ndx-survey-data, an NWB extension for survey responses

Data Science Core of Ripple U19
Stanford University

Role: Lead Data Scientist

portfolio

Collaboration with: Stanford University

We are leading the data scientist for a collaboration between Ivan Soltesz (lead PI), Gyorgy Buzsaki, Atilla Losonczy, and Mark Schnitzer to elucidate the cellular mechanisms of hippocampal sharp wave ripples using calcium imaging, electrophysiology and large scale simulation. Our role is to help these groups collaborate, sharing data and analyses through the NWB 2.0 standard, and promote standardization in the broader neuroscience community. We worked with the large scale simulation teams across several groups to establish a standard for simulation output as an NWB extension, and a converter from the SONATA format to NWB 2.0 here. With the help of Konstantinos Nasiotis, we built an NWB:N 2.0 reader in Brainstorm and worked with buzcode to read NWB:N 2.0 files. We have also released data collected by the Buzsaki Lab in NWB 2.0 on DANDI here.

Data Standardization for SCGB
Simons Foundation

Role: NWB specialist and Community Outreach

portfolio

Collaboration with: Simons Foundation

This pilot project with the Simons Foundation is aimed at helping labs adopt NWB while building and publicly releasing tools that will help other labs with the transition. This initial pilot included:

  • hippocampal recordings in NHPs from Beth Buffalo’s lab
  • Neuropixel recordings from Lisa Giocomo’s lab
  • Volumetric calcium imaging of a fruit fly from Richard Axel’s lab.

Through this project we have been working with the CaImAn development team to support reading and writing NWB 2.0 files.

In year two, we are working with:

  • Elizabeth Buffalo’s lab to develop a Django-based database and web UI for experimental records based on IBL’s alyx.
  • IBL to build conversion script from their internal data formats to NWB
  • The Tank and Brody Labs to build a modular analysis and conversion pipeline for Neuropixel data using SpikeInterface.

Interactive data visualization for NWB
Allen Institute and Kitware Inc.

Role: NWB specialist and developer

portfolio

Collaboration with: Allen Institute and Kitware Inc.

We worked with the cutting edge Neuropixel dataset being released by the Allen Institute, as well as state-of-the-art visualization tools from Kitware to provide users with interactive visualization to explore neural data. Our visualization focused on different visual techniques to capture hierarchical structure in the design of the experiment. These visualizations have been added to the NWBWidgets library for general use by the NWB community.

Neurodata Without Borders Dissemination
Lawrence Berkley National Laboratory

portfolio

The Lawrence Berkeley National Laboratory (LBNL) has been recently awarded a Brain Research Through Advancing Innovative Neurotechnologies (BRAIN) Initiative U24 grant from National Institute of Health for dissemination of the Neurodata Without Borders (NWB) standard for neurophysiology data to the neuroscience community over the next five (5) years. 

CatalystNeuro’s role is to work on plans for managing NWB expansion and modification, partner selection and prioritization, and sustainability, including community-driven processes for integration of new neuroscience technologies and data types with NWB, a user selection and prioritization plan, and a sustainability roadmap. The CatalystNeuro team also conducts optimization and standardization activities, including user support, tutorial coverage, and test coverage; conducts outreach activities, including tutorials and workshops; and overall dissemination of NWB.

NWB:N Community Liaison
Lawrence Berkeley National Laboratory sponsored by Kavli Foundation

Role: NWB specialist and Community Outreach

portfolio

Collaboration with: Lawrence Berkeley National Laboratory

We help the systems neuroscience community adopt NWB:N 2.0 through triaging issues, core development, and community engagement like organizing hackathons and writing the newsletter. This contract funds consulting for labs to help with NWB:N 2.0 adoption. Until February 1, 2021 I am able to provide the following services free of charge:

  • Discuss how NWB adoption could accelerate scientific discovery in your lab and help determine if NWB is appropriate for you.
  • Help to create a project plan for NWB adoption into your labs ecosystem.  I can help determine what analysis and conversion tools and extensions would be relevant for you.
  • Review pre-released NWB files for compliance with the NWB schema and best practices.

Spike Interface
Simons Foundation

portfolio

CatalystNeuro provides software development services is creating a unified and standardized platform for spike sorting, based on SpikeInterface, that integrates available open-source projects to facilitate the availability and usage of the latest technologies to neuroscience end-users. We provide a universal layer enabling comparison and easy exchanging of individual components of the extracellular electrophysiology data processing pipeline. This project aims at improving the SpikeInterface framework to better meet the needs of the neuroscience community and, in particular, Neuropixels users.

University of Washington

portfolio

NWB Integration – Brunton Lab

CatalystNeuro converted whole-day ECoG and behavioral recordings associated with the manuscript “Behavioral and Neural Variability of Naturalistic Arm Movements” (https://www.biorxiv.org/content/10.1101/2020.04.17.047357v2) to Neurodata Without Borders. We assisted in the publication of this data on DANDI, and publicly released the conversion code and some example code for accessing the data from NWB. We also provided interactive visualizations that allow data users to explore the data in the published NWB files.

 

Software Development – Buffalo Lab

CatalystNeuro built infrastructure around the existing lab records database to make this data more easily integrated into the lab’s workflow. CatalystNeuro, collaborating with Monadical, together provided expertise in neural data management and software engineering and delivering the following services: support and maintenance of existing database; migration from excel files into the database; a query “cookbook” in MATLAB that connects directly to the database and illustrates queries for a number of different practical queries; and a dashboard for visualization of the records data.

 

Availability for

helping other labs

We are always interested in opportunities to work with labs on data infrastructure. We have a team of computational neuroscientists who are experts in building conversion scripts, porting tools to NWB, and helping to package and publicly release data. We also have experience building custom databases and user interfaces for lab experimental records. If you are interested in working with us

Please contact us at inquiry@catalystneuro.com

Our work is entirely open source. See our GitHub organization for code of past and ongoing projects.